landscape(s) in transition: an environmental history of a village in north-east botswana

This article was downloaded by: [University of Western Ontario]On: 17 November 2014, At: 21:22Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number:1072954 Registered office: Mortimer House, 37-41 Mortimer Street,London W1T 3JH, UK

Journal of Southern AfricanStudiesPublication details, including instructions forauthors and subscription information:http://www.tandfonline.com/loi/cjss20

Landscape(s) in Transition:An Environmental Historyof a Village in North-eastBotswanaAnnika C. DahlbergPublished online: 04 Aug 2010.

To cite this article: Annika C. Dahlberg (2000) Landscape(s) in Transition: AnEnvironmental History of a Village in North-east Botswana, Journal of SouthernAfrican Studies, 26:4, 759-782, DOI: 10.1080/713683601

To link to this article: http://dx.doi.org/10.1080/713683601

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of allthe information (the “Content”) contained in the publications on ourplatform. However, Taylor & Francis, our agents, and our licensorsmake no representations or warranties whatsoever as to the accuracy,completeness, or suitability for any purpose of the Content. Any opinionsand views expressed in this publication are the opinions and views ofthe authors, and are not the views of or endorsed by Taylor & Francis.The accuracy of the Content should not be relied upon and should beindependently verified with primary sources of information. Taylor andFrancis shall not be liable for any losses, actions, claims, proceedings,demands, costs, expenses, damages, and other liabilities whatsoeveror howsoever caused arising directly or indirectly in connection with, inrelation to or arising out of the use of the Content.

This article may be used for research, teaching, and private studypurposes. Any substantial or systematic reproduction, redistribution,reselling, loan, sub-licensing, systematic supply, or distribution in any

http://www.tandfonline.com/loi/cjss20

http://www.tandfonline.com/action/showCitFormats?doi=10.1080/713683601

http://dx.doi.org/10.1080/713683601

form to anyone is expressly forbidden. Terms & Conditions of accessand use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

Journal of Southern African Studies, Volume 26, Number 4, December 2000

Landscape(s) in Transition: anEnvironmental History of a Village inNorth-east Botswana*

ANNIKA C. DAHLBERG

(Department of Physical Geography, Stockholm University)

Environmental change in Africa is often associated with processes of land exploitation anddegradation caused by human agents. However, recent studies have demonstrated howearlier assumptions about the causes and effects of environmental change in some caseshave been misleading and simplistic. This paper explores changes during the last centuryin the abundance and distribution of such natural resources as grazing, browse, � rewoodand edible fruits and berries in a village in north-eastern Botswana. A variety of datasources were used, primarily aerial photographs from different years, vegetation sampling,colonial documentation and recent of� cial records, as well as the oral histories of villagers.Initially, descriptions of the environment appeared confusing and contradictory. By apply-ing a geographical and historical approach, seemingly con� icting data can be broughttogether, thereby contributing to an understanding of actual and perceived change. Byfocusing on how the availability of speci� c natural resources has varied over time, differentand transitory landscapes, as experienced by different observers, are revealed.

Introduction

In Botswana, communal lands are mainly inhabited by subsistence farmers who cultivatesmall � elds with a variety of grain and vegetables, who may own herds of cattle, goats andsheep, and who utilise the natural produce of the surrounding lands for grazing, foraging,� rewood, building material, foodstuffs and medicine. However, households seldom relysolely on the land and most livelihood strategies include some dependence on outside wageincome. Thus, rural households are as sensitive to changes in the outside community as theyare to variations in the physical world, and to changes at household level. All these typesof changes in� uence a household’s decisions about how to manage the land, decisions thatmay have a strong impact on the local environment.

The debate about the environmental condition of the communal lands of Africa, andwhether they are sustainably managed, has been raging since early in the twentieth century,1

* The Of� ce of the President, Republic of Botswana, kindly authorised the research, and the � eldwork was � nancedby grants from the Swedish Agency for Research Cooperation with Developing Countries (SAREC), theScandinavian Institute of African Studies, the Swedish Society for Anthropology and Geography (SSAG), andthe C.F. Liljevalch Jr. research foundation. Much help was provided by the villagers of Kalakamate, and by staffat the District Agricultural Of� ce in Masunga, the Regional Range Ecology Of� ce in Francistown and theDepartment of Environmental Science, University of Botswana. I am grateful to Mats Leine, Karin Weilow andHans Drake for assistance with GIS and maps. Several colleagues, as well as an anonymous reviewer, madevaluable comments on earlier drafts. I am especially grateful to Piers Blaikie and Erling Dahlberg for givinggenerously of their time and interest.

1 M. Mainguet, Deserti� cation: Natural Background and Human Mismanagement (Berlin, 1991), p. 306; R.H.Behnke Jr., I. Scoones and C. Kerven (eds), Range Ecology at Disequilibrium: New Models of Natural Variability

ISSN 0305-7070 print; 1465-3893 online/00/040759-24 Ó 2000 Journal of Southern African StudiesDOI: 10.1080/0305707002000826 0

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14

760 Journal of Southern African Studies

and will not be re-hashed here. Suf� ce to state that government of� cials and scientists haveat times been too hasty in describing scenarios of environmental devastation, whether in theform of soil erosion, bush encroachment or a loss of bio-diversity. This observed orassumed land degradation was in most cases attributed to increased human and livestockpopulations in combination with traditional farming practices. Over the last decade andmore, a number of studies have investigated the interaction of environmental and socialdynamics. This has resulted in some important re-appraisals of previously establishedenvironmental histories for different parts of Africa.2 In semi-arid environments, areas ofbare ground and areas covered by dense stands of certain species of shrubs are oftenassumed to indicate overuse by people and livestock. A decrease in grass cover and certainchanges in species composition are other features used as indicators of land degradation.For the North East District, Botswana, numerous reports have, throughout the twentiethcentury, described how over-exploitation by people and livestock has caused degradation ofthe communal lands, resulting in the decrease of many valuable natural resources.3

Against this background, research was carried out between 1991 and 1996 in thecommunal area of Kalakamate, a village in North East District, Botswana.4 The overallobjective was to unravel the pattern of the environmental and social history of the areaduring the twentieth century. The bio-physical, social and historical setting of the study areais given in more detail elsewhere,5 as is an examination of the interface betweensocio-economic and environmental change, including a discussion of some ensuing theoreti-cal implications.6 The present paper presents the environmental outcomes of the bio-physical and socio-economic processes that have had an impact on the landscape over thelast century. More speci� cally, this paper describes the temporal and spatial changes of thewoody vegetation, important for browse, edible fruits and berries, � rewood and buildingmaterial, and of the grass cover, used for grazing and for roof thatching material. Italso offers an analysis of different and sometimes con� icting environmental histories,and presents one interpretation of the causative linkages behind speci� c environmentaloutcomes.

Footnote 1 continuedand Pastoral Adaptation in African Savannas (London, 1993), p. 248; A. [C.] Dahlberg, Contesting Views andChanging Paradigms: the Land Degradation Debate in Southern Africa, Discussion Paper No 6, ScandinavianInstitute of African Studies (Uppsala, 1994), p. 59; D.S.G. Thomas and N.J. Middleton, Deserti� cation: Explodingthe Myth (Chichester, 1994), p. 194; and J. Fairhead and M. Leach, Reframing Deforestation. Global Analysisand Local Realities: Studies in West Africa (London, 1998), p. 238.

2 U. Hellden, ‘Deserti� cation – time for an assessment’, Ambio, 20, 8 (1991), pp. 372–383; C.M. Shackleton, ‘Arethe communal grazing lands in need of saving?’, Development Southern Africa, 10, 1 (1993), pp. 65–78; M. Tiffen,M. Mortimore and F. Gichuki, More People, Less Erosion. Environmental Recovery in Kenya (Chichester, 1994),p. 311; J. Fairhead and M. Leach, ‘Reading forest history backwards: the interaction of policy and local land usein Guinea’s forest–savanna mosaic, 1893–1993’, Environment and History, 1 (1995), pp. 55–91; M.T. Hoffman,W.J. Bond and W.D. Stock, ‘Deserti� cation of the Eastern Karoo, South Africa: con� icting paleoecological ,historical and soil isotopic evidence’, Environmental Monitoring and Assessment, 37 (1995), pp. 159–177; andY.A. Harrison and C.M. Shackleton, ‘Resilience of South African communal grazing lands after removal of highgrazing pressure’, Land Degradation and Development, 10 (1999), pp. 225–239.

3 Reviewed in L. Fortmann, ‘Peasant and of� cial views of rangeland use in Botswana: � fty years of devastation?’,Land Use Policy, 6 (1989), pp. 197–202; in A.C. Dahlberg, ‘Interpretations of environmental change and diversity:a study from North East District, Botswana’ (PhD thesis, Dissertation Series No 7, Department of PhysicalGeography, Stockholm University, 1996), and in P. Kinlund, ‘Does land degradation matter? Perspectives onenvironmenta l change in North–Eastern Botswana’ (PhD thesis, Stockholm Studies in Human Geography, No 7,Stockholm University (Stockholm, 1996), p. 246.

4 Dahlberg, ‘Interpretations of environmenta l change and diversity’.5 A.C. Dahlberg, ‘On interpreting environmenta l change: time, space and perception in the case of Kalakamate,

North East District, Botswana’, in F. Ganry and B. Campbell (eds), Sustainable Land Management in AfricanSemi–arid and Subhumid Regions (Montpellier,1995), pp. 257–271.

6 A.C. Dahlberg and P.M. Blaikie, ‘Changes in landscape or in interpretation? Re� ections based on theenvironmenta l and socio–economic history of a village in NE Botswana’, Environment and History, 5, 2 (1999),pp. 127–174.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14

Landscape(s) in Transition 761

In the study area, several of the environmental changes described or assumed for thedistrict as a whole, as well as for the country’s communal lands in general, had in manycases not even occurred. Where they were observed, they were not as permanent as usuallypresumed, and had in many cases been triggered by events and circumstances quite differentfrom those often assumed as the main causative agents. For example, the total area of bareground had not increased, grass cover had not suffered an overall decline, � rewood was notbecoming scarce and, although bush encroachment had occurred, it was not caused byovergrazing and was naturally reversible. However, in certain areas, grass cover had beenreplaced by woody species, bare ground had become more widespread round the centralsettlement and preferred thatching grass species seemed to have decreased.

The Study Area

A brief description of the study area is presented here; more details are given elsewhere.7

North East District is one of the smallest and most densely populated districts in Botswana(Figure 1). Settled by the British in the late nineteenth century, most of the district was soonturned into private farms for Europeans, while the local population was compelled tocongregate in designated Native Reserves. The area in the north-west part of the district haslong been settled by agro-pastoral Kalanga-speaking people and, since the turn of the last

Figure 1. Botswana, North East District, and the study area, the village of Kalakamate.

7 Dahlberg, ‘On interpreting environmenta l change: time, space and perception’; Dahlberg, ‘Interpretations ofenvironmenta l change and diversity’; Dahlberg and Blaikie, ‘Changes in landscape or in interpretation?’; and A.C.Dahlberg, ‘Vegetation diversity and change in relation to land use, soil and rainfall: a case study from North EastDistrict, Botswana’, Journal of Arid Environments, 44(2000) , pp. 19–40.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


century, also by different Setswana-speaking groups.8 The village9 of Kalakamate todaymeasures some 70 square kilometres, whereas previously it extended both east and north.Vukwi River constituted the eastern border until the freehold farm west of the river wasestablished. The farmland continued to be used by villagers until it was fenced in the 1950s.In the late 1980s a veterinary cordon fence was erected to the north and north-east, furthercurtailing the movement of people and cattle.

The landscape is � at to gently undulating, with granite kopjes and ridges of differentbasic rock scattered throughout , and crossed by a number of seasonal rivers and streams.Soils vary in depth and quality over short distances but are mainly shallow with a highcontent of gravel. They are moderately to well drained and vary from loamy sand to sandyclay loam. The most common soils are regosols, luvisols and leptosols, with alluvium alongthe watercourses and a large expanse of vertisol in the south.10 Annual rainfall is generallyquite low and extremely variable, with an average of 458 mm/year, a median of 426 mm/year and a coef� cient of variation of 39 percent.11 Vegetation is characterised by moderatelydense to dense woodland and bush-savanna, interspersed with grassy patches and areas witha low vegetative cover. In many areas Colophospermum mopane dominates among thewoody species, with other common genera being Acacia, Combretum, Commiphora,Terminalia and Grewia. Abandoned � elds and homesteads are covered with dense bush,either a combination of the thorny species Acacia tortilis and Dichrostachys cinerea, or bypure stands of C. mopane. Riverine forests are characterised by tall trees of many differentspecies. The grass sward is dominated by the genera Aristida, Eragrostis, Digitaria, Tragusand Urochloa, and cover and species composition varies between areas, as well as betweenseasons and years.

Methodology

Data were collected from archival documents and other historical texts, from of� cial recordsof rainfall, population, livestock and land use, from aerial photographs from 1964, 1971,1981 and 1988, � eld observations, soil and vegetation sampling, and from interviews.12

Interviews were conducted with 42 villagers (6 percent of the population) and werefocused on the old people. The median age of respondents was 63 and about 35 percent ofall villagers aged 50 or more were interviewed. The majority of the men had been born inthe village, although many had been away for varying lengths of time as migratory workers,while over half of the women had moved to Kalakamate at marriage. Most respondents werepresently occupied with farming and a few combined this with other work. In selectingrespondents, a roughly equal number of men and women, as well as an even spatialdistribution of homesteads and � elds throughout the village area were aimed for. Theinterviews were conducted during two growing seasons, in 1991/92 (a dry year) and

8 R.P. Werbner, ‘Land, movement and status among Kalanga of Botswana’, in M. Fortes and S. Patterson (eds),Studies in African Social Anthropology (London, 1975), pp. 95–120.

9 In this paper ‘village’ refers to the whole area under the authority of the headman of Kalakamate, while the term‘settlement’ is used for the parts of this area occupied by homesteads.

10 Botswana, ‘Ramokgwebana sheet (1:250 000). Soil Map’ (Soil Mapping and Advisory Service Project Bot/80/003,FAO/UNDP/Government of Botswana, 1984), and D.J. Radcliffe (ed), The Soils of North Eastern Botswana, SoilMapping and Advisory Services, Field Document 17, FAO/UNDP/Government of Botswana (Gaborone, 1990),p. 169.

11 The rainfall data for 1922/23 to 1993/94 were obtained from the Department of Meteorological Services, Ministryof Works and Communication , Government of Botswana and are for Francistown, 66 km south-east ofKalakamate.

12 Aerial photograph interpretation, GIS analysis and details about interview procedure and respondent s are describedin Dahlberg ‘Interpretations of environmenta l change and diversity’, and in Dahlberg and Blaikie, ‘Changes inlandscape or in interpretation?’

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


1993/94 (a year of average rainfall). Interviews were semi-structured and often conductedat the � elds or when walking through the grazing land. Different topics were highlightedwith different respondents. Thus, the data presented do not show full coverage for allquestions.

Environmental history, a subject label more often used by social scientists than naturalscientists, can be approached in many different ways. In this study, physical and social datahave been given equal weight and explanations have been sought through the examinationof the micro-history13 of one village. It is recognised that the use of oral histories posesproblems of reliability and objectivity14 as does the use of many traditionally more acceptedsources of data. However, for many regions oral histories are one of the few sources ofinformation about the past. When used in conjunction with other data, oral histories can besubjected to critical analysis, just as other data – such as of� cial records and aerialphotograph interpretations – can be checked against information from interviews.15

Processes Resulting in Environmental Change

Preliminary analyses determined the � nal analytical approach, which was to view recordedenvironmental change as environmental outcomes of different types of change that hadoccurred in the bio-physical and social sphere.16 Changes within the social sphere that couldbe expected to have had an impact on the environment were investigated. An increase inhuman and livestock population is often assumed for communal areas, since this hasoccurred at a national and regional scale. In the study area, records of human populationonly go back to 1964, and between then and the census of 1991 these show an increase of1.5 percent. In addition, when existing data are computed as a least squares exponential � tfor the whole period considered (ie the twentieth century), the estimated average annualgrowth rate is 1.5 percent, which is less than the averages for the country, the district andthe local chiefdom. For the livestock population, national � gures show an eightfold increasefrom the beginning of the twentieth century up until the early 1990s,17 and for FrancistownVeterinary Region (within which the study area is situated) there is also a clear trend ofincrease over this period. However, for the district, the chiefdom and the study area, notrends of increase occurred for the time periods for which records existed.18 Instead, as alsostated by villagers, records showed how livestock numbers had varied over time, dependingon rainfall, diseases and access to grazing areas.

Agricultural practices had changed very little during the last century, while the spatialpattern of cultivation and grazing had changed more, especially in the last three decades.As already described, the land east of the village was fenced off in the 1950s and aveterinary cordon fence erected to the north and north-east of the village in the 1980s,which severely curtailed villagers’ access to farmland and grazing land. After independencein 1966, the Government invested in water availability, healthcare, education and communi-

13 G. Levi, ‘On microhistory’, in P. Burke (ed), New Perspectives on Historical Writing (Cambridge, 1991),pp. 93–113.

14 P. Thompson, The Voice of the Past: Oral History (Second edition, Oxford, 1988) p. 314; G. Prins, ‘Oral history’,in P. Burke (ed), New Perspectives on Historical Writing, pp. 114–119.

15 See, for example, discussions in Prins, ‘Oral history’, and in several contributions in C.K. Adenaike and J. Vansina(eds), In Pursuit of History: Fieldwork in Africa (Portsmouth, 1996), p. 155.

16 Dahlberg and Blaikie, ‘Changes in landscape or in interpretation?’17 N.O.J. Abel, ‘What’s in a number? The carrying capacity controversy on the communal rangelands of Southern

Africa’ (PhD thesis, University of East Anglia, 1992).18 These time periods differ, with district records going back to 1905, records for the chiefdom to 1930 and for the

study area to 1976. It should be noted that for none of these areas, not even for the veterinary region, do recordsexist for all years.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


cation. In Kalakamate this has meant a new bore-hole, a dam, a clinic, schools and a tarredroad. The Kalanga people of the area had previously lived in scattered homesteads, withtheir � elds nearby, which resulted in a patchwork of homesteads, � elds and grazing landthroughout the whole communal area. Since independence, most people have built newhomes close to the improved infrastructure, thus creating a nucleated settlement in the westand north-west of Kalakamate – along parts of the road and especially around the schools,clinic and tribal court. Fields situated far from the new homes have been abandoned andnew ones cleared where land was available nearby. This has, in some cases, resulted inpeople now farming soils less suitable for crops and/or farming the same land for longerperiods.

In semi-arid regions, rainfall has a strong impact on the environment and in most areaswater is a limiting factor for plant growth. Rainfall is extremely variable within andbetween years and this is mirrored in the temporal variation of vegetation type and cover.Rainfall records for North East District show a consistently variable pattern but no overalltrend can be discerned.19 Many villagers claimed that rainfall had decreased since they wereyoung but this is not re� ected in the records (Figure 2). It is likely that the broad patternsof rainfall variation over the last few decades have had an in� uence on this perception. The1970s was characterised by relatively high rainfall, which was followed by a long period(the 1980s and early 1990s) of recurring agricultural droughts. The perceived decline inrainfall was often used by villagers to explain observed long-term environmental changesand also underlay assumptions about changes presumed to have occurred.

Figure 2. Annual rainfall (July–June) at Francistown for the period 1922/23 to 1993/94.

19 See Dahlberg and Blaikie, ‘Changes in landscape or in interpretation?’, pp. 147–165, for a detailed analysis ofrainfall patterns and variation in the study area, including how these are linked to perceived and measuredenvironmenta l change.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


Environmental Outcomes – Results and Discussion

A variety of data sources were used to uncover environmental changes that had occurredin the study area. Available information mainly covers those environmental variables thatare important for the economy of the household, community or district. The environmentaloutcomes caused by changes in the bio-physical and social sphere discussed here deal withthe cover, density and species composition of the woody vegetation and the grass sward,and with the speci� c abundance and availability of � rewood, fruits, berries and thatchinggrass.

Trees and Shrubs

When comparing the present cover and density of trees and shrubs with that rememberedfrom their youth, most respondents described an overall increase throughout the area. Thiswas described visually: ‘Before, the trees grew scattered, with grass in between, but nowthere is less grass and the trees stand closer’; ‘Now the country looks forested’; and as aphysical experience: ‘Now there are places where it is dif� cult to walk’; or as a feeling:‘Now you fear to walk because you cannot see ahead’. Sometimes this increase referred tospeci� c areas and/or speci� c species: ‘Now there are only thorn bushes where there usedto grow nyevi’ (an edible wild plant found in cultivated � elds and recently abandoned � eldsand homesteads) or: ‘Now the shrubs have closed over the path, but it is not a generalincrease’. For certain species a decline had been observed, mostly with reference to speci� careas: ‘There are less trees around the houses now, because people use them to build kraals(enclosures for livestock), but between the houses and the � elds trees have increased’. Thetwo thorny species Acacia tortilis and Dichrostachys cinerea, as well as Colophospermummopane , were mentioned as the main contributors to the increase (see Table 1). Acaciatortilis and D. cinerea were said to have increased ‘all over’ but especially in abandoned� elds and homesteads: ‘Where these species grow in abundance, you know there has beena � eld or a ruin’.

In the range ecology literature, A. tortilis and D. cinerea, common throughoutBotswana, are described as invader species (ie bush-encroachment species) and thickets ofthese species are considered indicators of disturbed land, usually associated with over-

Table 1. Reported change in some common woody species (listed in typicalorder of preference as browse species)

Species Increase No change Decrease

Colophospermum mopane 10 2 3Grewia bicolor/Grewia spp. 2 4 3Combretum apiculatum 4 1 2Dichrostachys cinerea (pods) 11 6 3Acacia tortilis (pods) 14 3 2Combretum hereroense 3 1 –Commiphora spp. – 1 2Kirkia acuminata – – 1Lonchocarpus nelsii 2 – –Grewia � avescens 1 1 –Combretum imberbe 1 1 –

Numbers refer to how many respondents mentioned a particular change.Several of these species provide � rewood or edible berries, and are thereforelisted also in Tables 3 and/or 4.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


grazing.20 They are drought resistant, and D. cinerea is fairly resistant to bush � res. Seedgermination is enhanced by passage through a browsing animal, they grow slowly, the podsare nutritious and the wood provides good fuel. Along the banks of rivers and streams A.tortilis is commonly found growing quite tall.21 Colophospermum mopane is very commonin the northern and north-eastern parts of the country, often forming single species stands.The leaves and pods provide important fodder and the seeds are dispersed mainly byadhering to the hooves of passing animals. Seeds germinate easily but seedlings growslowly. The wood is resistant to termites and thus popular for poles. It is also used forcarvings and provides excellent fuel.22

Aerial photographs of Kalakamate show that between 1964 and 1988 medium and densewoodland had increased in certain parts of the village, while in other parts the cover ofthese vegetation types had decreased (see Figures 3(a) and (b)).23 Area calculations for thesevegetation classes showed that the total cover of dense woodland had remained virtually thesame (about 21 km2), while the total area of medium cover had increased from approxi-mately 24 km2 to 28 km2. The total areas of the classes ‘very sparse’ and ‘sparse woodycover’, which include temporarily and permanently abandoned � elds, had both decreased.The increase of woody vegetation was particularly noticeable in the middle and easternparts of the village – those areas previously occupied by scattered homesteads and � elds,many of which had been abandoned as people moved to the main settlements in the westand north-west.

Riverine forests can be seen in Figure 3 along the Shashe River (the western border ofthe village) and as narrow strips of dense woody vegetation throughout the area. For thisvegetation category respondents claimed there had been a general increase of all species,and Combretum hereroense, Combretum imberbe and A. tortilis were mentionedspeci� cally. Classi� cation of the aerial photographs showed that the total area of denseriverine forest was almost the same for the two years, while the area of riverine forestclassi� ed as less dense had decreased. For the following discussion, the change in thespatial pattern of cover of different density classes of woody vegetation is of as much, ormore, interest as the total areas of the different classes.

Species are valued for the different resources they provide and ascribed values will varybetween people depending on their speci� c needs. Table 2 shows the types of produce thecommon woody species were valued for and compares the information given in Kalakamatewith that found in other studies.24 A number of species were valued for a variety of productsalthough browse and � rewood were the most important. These and other uses are importantto local communities throughout Africa.25

20 J. Timberlake, Handbook of Botswana Acacias, Botswana, Division of Land Utilisation, Ministry of Agriculture(Gaborone, 1980), pp. 98–100; and J.A. van Vegten, ‘Man-made vegetation changes: an example from Botswana’sSavanna’ National Institute of Development and Cultural Research (NIR), Working Paper No 40, (Gaborone.1981), p. 81.

21 Timberlake, Handbook of Botswana Acacias; and K. Coates Palgrave, Trees of Southern Africa (Cape Town,1983), pp. 252–269.

22 Coates Palgrave, Trees of Southern Africa; J. Woollard, A Vegetative Key to the Woody Plants of SouthernBotswana (Gaborone, 1984), p. 95; and P. van Wyk, Field Guide to the Trees of the Kruger National Park (CapeTown, 1984), pp. 76–83.

23 Figures 3(a) and (b) are based on interpretation of aerial photographs , and are recti� ed against the topographicmap.

24 See, for example, Timberlake, Handbook of Botswana Acacias; J.R. Timberlake, ‘Colophospermum mopane.Annotated Bibliography and Review’, Zimbabwe Bulletin of Forestry Research No 11, Division of Research andDevelopment , Forestry Commission (Harare, 1995), p. 49; Coates Palgrave, Trees of Southern Africa; Woollard,A Vegetative Key to the Woody Plants of Southern Botswana; and van Wyk, Field Guide to the Trees of the KrugerNational Park.

25 See, for example, I.M. Grundy, B.M. Campbell, S. Balebereho, R. Cunliffe, C. Tafangenyasha , R. Fergusson andD. Parry, ‘Availability and use of trees in Mutanda Resettlement Area, Zimbabwe’, Forest Ecology andManagement , 56 (1993), pp. 243–266.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


Table 2. Ascribed user-value of selected woody species

Browse Firewood Building Edible Produce

Colophospermum mopane X 1 X 1 X 1 X 1Acacia tortilis X 1 X 1 X 1Grewia spp. X 1 X X 1Dichrostachys cinerea X 1 X 1 1Combretum apiculatum X 1 X 1Acacia nigrescens 1 X 1Commiphora mollis X 1Securinega virosa X 1Kirkia acuminata X

X stands for user value ascribed by villagers, and 1 stands for user value described in theliterature. Edible produce means not only fruit and berries but also honey and certain larvae.Table 2 only shows what a particular species may be used for, not the relative value ascribedto different species. In Tables 1, 3 and 4, on the other hand, species are listed in order ofpreference based on the average ranking by the respondents .

It was noticeable that women were more concerned and knowledgeable than men aboutchanges in the abundance of fruits, berries and good � rewood (these are discussed in moredetail later in this paper). On the other hand, men knew more about the trees and shrubsthat provide good browse for livestock. Most men claimed that access to nutritious browseis necessary only for part of the dry season and in drought years but some described how,‘Cattle mix grass and browse even when there is plenty of grass’, and how cattle activelyseek out preferred woody species. When discussing the availability of browse there wereno complaints about a decline. Instead, several respondents described an increase, althoughnot necessarily of the most preferred species. Villagers offered a number of alternativeexplanations for this, some of which were con� icting or speculative. Some respondentsattributed the observed change to the will of God or simply had no opinion. Many othersmentioned exceptional rainfall events as the cause of the increase; for example: ‘Theincrease started during the good rains in the 1940s, and since then there have been otherperiods with enough rain for germination’. The death of individual trees or stands was oftenattributed to speci� c drought events: ‘Trees are dying this year [1992] because of thedrought’. Thus C. mopane was said to have declined in some areas due to drought orbecause of excessive cutting of poles. However, drought was also claimed sometimes to bethe cause of increase: ‘Mosu (A. tortilis), moselesele (D. cinerea) and mophane (C. mopane)can germinate also in drought years – they even produce more seeds then’, which wasthought to give these species a competitive advantage. The germination of these, and a fewother species, was also said to be enhanced by livestock, who ‘spread these seeds, so thetrees and shrubs grow where the animals move’.

As one of several laws intended by the government to help protect the country’srangelands, the Herbages Preservation Act (also termed the Prevention of Fires Act) wasproclaimed in 1977.26 This and related laws are supposed to be enforced by variousinstitutions at District level, and although in most areas such control over land management

26 This law and its effects are discussed, for example, in J.A. van Vegten, ‘Some aspects of African ecology withspecial reference to Botswana’, National Institute of Development and Cultural Research (NIR), Working PaperNo 26, (Gaborone, 1979), p. 26; and in J.W. Arntzen, ‘Agricultural developmen t and land use in eastern communalBotswana: the Case of Kgatleng District’, National Institute of Development Research and Documentation (NIR),Working Paper No 50 (Gaborone, 1985), p. 84.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


Figure 3a. Cover of woody vegetation in Kalakamate 1964. Figures 3a and 3b are based on interpretation of aerialphotographs , and are recti� ed against the topographi c map.

does not exist,27 villagers claim that most people nowadays refrain from burning the grazingland. The law against the burning of communal land was mentioned by several villagers asan important cause of the increase of shrubs, which represent much of the ‘medium cover’in Figure 3. It was stated that, before this law, � res had been common. Some respondentsclaimed that only particular areas were burnt, while according to others, ‘Every year wewould burn half of the grazing land and leave half for the cattle to graze’. Land beingcleared for cultivation was also regularly burnt. All agreed that burning should only takeplace in late winter or early spring: ‘You should burn in September or October, just

27 Arntzen, ‘Agricultural developmen t and land use in eastern communal Botswana’; and Gulbrandsen, ‘Access toagricultural land and communal land management in eastern Botswana’, Government of Botswana, Ministry ofLocal Government s and Lands, Applied Research Unit (Gaborone, 1984), p. 117.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


Figure 3b. Cover of woody vegetation in Kalakamate 1988.

before you expect the rains to come, then when the rain falls new grass will germinate’. Themain reasons for burning, apart from clearing land for � elds, were to get rid of ticks, todestroy the old grass and encourage new growth, the fear of wild animals and to get rid ofshrubs: ‘These � res were the only thing that could stop the increase of mosu (A. tortilis)and moselesele (D. cinerea)’. In areas where � elds had been cleared on land dominated byC. mopane, an increase of this species was attributed to stumps not being burnt off, causingexcessive sprouting as soon as the � eld was abandoned. Villagers had heard that the reasonfor regulating � res was that, ‘If the land is burnt and then no rain follows there would beno grazing at all, whereas without the burning at least there would be some grass’. Someagreed with this, especially under conditions of perceived rainfall decline, but others said,‘This might be a good policy for areas where there are not much shrubs, while in a placelike Kalakamate the burning would provide better grazing’.

Villagers in the eastern part of the district claim that trees in and around their villages

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


had declined in numbers over the previous ten years.28 A survey in the village of Moroka,in the north-east part of the district (Figure 1), revealed a gradual, but still drastic, declinein woody vegetation.29 Here many species bearing fruit or berries were said to havedeclined, as well as Ziziphus mucronata and Commiphora spp., which in Moroka weredescribed as valuable browse species. In preparation of the National Conservation Strat-egy30 village meetings were held throughout the district. From these it was reported that insome areas C. mopane and a few other species were becoming scarce, while A. tortilis andD. cinerea were increasing. The reasons for the decrease in trees were listed as: the cuttingof live trees for fuel, excessive cutting of � rewood by commercial dealers and, in villagesclose to Francistown, the continuous allocation of new farmland because of increasingpopulation pressure.31

For Kalakamate, the combination of local ecological knowledge with data provided bythe aerial photographs and � eld-checks results in a picture of change where the main featureis the nucleation of settlements and � elds. As people have moved, their old � elds andhomesteads have been abandoned and have become overgrown, clearings and paths are nolonger maintained and the pattern of human and cattle movement through the landscape haschanged. In addition, around the main settlement, some � elds have been abandoned, eithertemporarily due to shortage of labour and/or draught power, or more permanently due toinfestation of weeds. Further discussions often revealed that the initial generalised statementabout an increase of trees and shrubs more speci� cally referred to the colonisation ofabandoned � elds by the species A. tortilis, D. cinerea and, in some areas, C. mopane.

On most soils A. tortilis and D. cinerea are the � rst to colonise a disturbed site and, ina few years, they grow into dense stands two to three metres in height. Apart from age, thedensity seems to depend on soil type and, in Figure 3, this vegetation type is representedin both the classes ‘medium’ and ‘dense’. These thorn bush thickets remain for varyinglengths of time, depending mainly on soil type, but are gradually replaced by a morespecies-rich, and less dense, mix of tree and shrub species after a period of a few decades.In an area where cultivation was abandoned in the 1930s and 1940s a total of 37 woodyspecies were encountered and, among these, A. tortilis and D. cinerea occurred only asscattered individuals.32 In Kalakamate, as elsewhere in eastern Botswana, farming ofteninvolves a cyclic process of land use: land is cleared, destumped or burnt or both, cultivated(often for many years), and then gradually abandoned and allowed to become overgrownwith bush or forest before someone decades later clears the land anew. This process has astrong impact on changes in the density and species composition of woody vegetation.33

The environmental history of land used for cultivation was investigated and the process ofconstant variation in the vegetative cover is exempli� ed in the two quotations below:

We started to clear our new � eld last year [1993]. Part of this land has never been used as a� eld before, and the trees here are tall with plenty of grass underneath. The trees on themotlhabana [sandy] soil are mainly mhoto [Lannea stuhlmannii], morula [Sclerocarya birrea],modumela [Kirkia acuminata], mokoba [Acacia nigrescens], mohudiri [Combretum apiculatum]

28 L. Fortmann, K.E. Gobotswang, U. Edzani, T. Woto, A. Magama and L. Motswogole, ‘Local institutions, villagedevelopment , and resource management : case studies from North East District, Botswana’, Government ofBotswana, Ministry of Local Government and Lands, Applied Research Unit, with the Land Tenure Center,University of Wisconsin (Gaborone, 1983), p. 232.

29 Kinlund, Does Land Degradation Matter?30 ‘Botswana National ConservationStrategy. National Policy on Natural Resources Conservation and Development’

(Gaborone, Government Paper No 1 of 1990).31 G. Asselman, ‘National Conservation Strategy Consultation Report for North East District’ (Francistown,

unpublished report, District Of� ce Lands, 1986), p. 10.32 Dahlberg, ‘Vegetation diversity and change in relation to land use, soil and rainfall’.33 Dahlberg, ‘On interpreting environmenta l change: time, space and perception’.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


and mokhomoto [Commiphora spp.], while on the mokwakwa [more clay-rich] soil mophane[C. mopane] dominates. Part of the land allocated to us was previously ploughed by myhusband’s father. It was abandoned in 1971 when he died, and here mosu [A. tortilis] andmoselesele [D. cinerea] dominate.

We cleared our present � eld when we returned to Kalakamate after 1966. The soil ismotlhabana [sandy] mixed with mokwakwa [more clay-rich] soil and the land was covered bytall bush, mainly mosu [A. tortilis] and moselesele [D. cinerea]. This land had previously beenused by my parents and abandoned by them when I was young, in the 1940s or early 1950s.They abandoned it because of problems with motlhwa [Cynodon dactylon].34

Farmers describe the management cycle of clearing land, then ploughing this area formany years before abandoning it due to different factors affecting yields (such as infestationof weeds or pests or simply a decline in soil fertility), as an ongoing practice. New � eldscleared within reasonable walking distance from the growing settlement areas in thenorth-west part of the village were still considered quite new and able to provide a goodyield. However, as houses now are built of more permanent materials, farmers are makinguse of favourable government loans to buy fencing materials for their � elds, and asuncultivated land becomes more scarce around the main settlement, it is likely that the oldpractice of shifting the place of a family’s homestead and � elds will become moreuncommon. This, in turn, can be expected to have an impact on the environment.

An explanation of the observed changes in the woody vegetation in Kalakamate wassought in the different trends and events that have affected local livelihood strategies andland management decisions. An increase in human population would be expected toincrease the off-take of wood for fuel and building, but in Kalakamate the populationincrease has been comparatively modest, especially when effects of migration are taken intoaccount. For Kalakamate, no statistics on past migration exist but among those interviewedfor this study 70 percent of the men had worked elsewhere for periods of between 7 and46 years, and 40 percent had been employed far from their homes for more than 20 years.Many had worked in the mines in South Africa, which was especially common during the� rst half of the last century,35 while others had found employment in other parts of theProtectorate. These migration patterns are likely to have affected the environment in manydifferent ways. The absence of husbands and sons will have had an impact on where andwhen new � elds were cleared, on ploughing, and on herd-size and grazing practices.However, in Kalakamate as elsewhere, the effect on land use, and thus on the environment,will have varied between households and over time, depending on such factors as length ofabsence and the age and sex of remaining family members.36

However, more important than the overall increase and/or variation in populationnumbers, is the change in the density of the population in different parts of the village. Theincrease of woody vegetation has occurred predominantly in areas almost completelyabandoned for purposes of habitation and cultivation since independence, and which arenow less exploited than before. The decrease in woody vegetation around the main

34 The grass Cynodon dactylon is considered a valuable fodder species when found in grazing areas, but in � eldsit is described as a weed. It spreads quickly and, due to its deep underground stolons, is very dif� cult to get ridof, and where it occurs will soon out-compete crops. See, for example, M. Phillips, A Guide to the Weeds ofBotswana, Government of Botswana, Ministry of Agriculture, Department of Agricultural Research,(Gaborone,1991) , p. 159.

35 See, for example, I. Schapera, Migrant Labour and Tribal Life: a Study of Conditions in the BechuanalandProtectorate (London, 1947); and N. Tumkaya, ‘Botswana’s population trends: past and future’, Botswana Notesand Records, 19 (1987), pp. 113–128.

36 J. Hesselberg and G. Wikan, ‘The impact of absenteeism on crop production and standard of living in two villagesin Botswana’, Botswana Notes and Records, 14 (1982), pp. 69–73; and W.G. Morapedi, ‘Migrant labour and thepeasantry in the Bechuanaland Protectorate, 1930–1965’, Journal of Southern African Studies, 25, 2 (1999),pp. 197–214.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


settlement areas, revealed by the aerial photographs, has been caused by the gradualincrease in population density in these parts of the village. The described decrease of woodyvegetation in certain areas further removed from settlement areas, through the death of aspeci� c species, was generally blamed on drought. Such drought-related die-offs have alsobeen described for other parts of the country.37

In Kalakamate, much of the increase of woody vegetation is represented by a fewspecies of thorny shrubs. In reports from other areas in Botswana, an increase of thornyshrubs is attributed almost exclusively to high grazing pressure.38 Livestock statistics for thecountry show that by the early 1990s there had been an eightfold increase in livestocknumbers since the beginning of the last century.39 For Francistown Agricultural Region(formerly Veterinary Region), within which Kalakamate falls, the increase has been evenhigher.40 However, much of this increase has occurred in certain areas given over almostexclusively to commercial cattle ranches, which produce beef for export. It cannot beassumed that such increases in cattle numbers have also occurred in the communal areasand any such generalisations should be avoided. For the communal areas, regular countswere not conducted. Thus � gures only exist (and/or can be calculated based on varioussecondary sources) for odd years and the data are often of very poor quality. Nevertheless,according to the livestock records that are available for Kalakamate and the surroundingcommunal areas, there seems to be no evidence of a trend of increasing numbers.41

Kalakamate was part of the colonial administrative unit Tati Native Reserve and, for thisarea, the � rst recorded cattle count was done in the early part of the twentieth century.Subsequent counts reveal that the highest numbers occurred in the early 1950s. For theChiefdom of Ramokate, to which Kalakamate belongs, records show cattle numbers to havebeen highest in the late 1940s. In Kalakamate itself, villagers’ statements indicate that thenumber of cattle has decreased over the last few decades. The records for Kalakamate onlygo back to 1976 – since then cattle numbers were highest (over 16,000) in the early 1980sand have since � uctuated at much lower levels.

In Kalakamate, high livestock pressure, indicated by an increase in the area of bareground, has only been observed close to the main settlement area and around wateringpoints. In spite of this, an increase in the thorny shrubs A. tortilis and D. cinerea is reportedand today dense thickets of these species are common. However, although grazing is likelyto enhance the germination of these species, bush encroachment in this area is primarilyfound on sites disturbed by cultivation or settlement, and then abandoned. This causativelink is recognised by villagers and corroborated by aerial photographs and � eld-checks. Therecurrent abandonment of � elds and homesteads has long been characteristic of the area, butthe nucleation of settlements and � elds caused a drastic change in the spatial pattern of landuse, which in turn affected the pattern of bush encroachment. A previous patch-likeoccurrence of small areas of thorny shrubs caused by the gradual abandonment of parts of� elds, a whole � eld or an isolated homestead, was replaced by a situation where, in certainareas, all � elds and homesteads were abandoned.

Environmental outcomes due to reduced burning of grazing land are more dif� cult toassess. In spite of much research and the compilation of a number of excellent and thorough

37 D. Parry, ‘Some effects of the 1981–87 drought on woody plants in Botswana’, Botswana Notes and Records,20 (1989), pp. 155–159.

38 van Vegten, ‘Man-made vegetation changes: an example from Botswana’s Savanna’; and J.W Arntzen and E.MVeenendaal, A Pro� le of Environment and Development in Botswana, Netherlands, Free University and Universityof Botswana, National Institute of Development Research and Documentation (NIR) (Gaborone, 1986), p. 171.

39 Abel, ‘What’s in a number?’40 See various statistics compiled in Dahlberg and Blaikie, ‘Changes in landscape or in interpretation?’41 Dahlberg and Blaikie, ‘Changes in landscape or in interpretation?’

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


reviews on the ecological effects of � re and of � re as a management tool, it is stillimpossible to be sure of the outcome of different � re regimes, including the absence of� re.42 The direct effects of � re vary due to inherent features such as the frequency andintensity of the � re, and at what time of the year it occurs. In addition, the ecological effectsof � re are extremely dif� cult to assess (and predict), especially in such � re-adaptedenvironments as the savanna, because of the complexity of the interactions between � re anda number of other factors, such as herbivory and rainfall.43 These interactions operate overtime, so that the short- and long-term effects of any speci� c � re regime will depend on thepast history of the environment. In a similar manner, the effects will vary depending onevents subsequent to the � re, especially in relation to rainfall and the intensity of grazingand browsing. Thus, lacking detailed data on past grazing and browsing pressure as well ason the � re history of Kalakamate, it is impossible to state with certainty to what extent theincrease in the woody vegetation in large parts of the village is due to a reduction in thefrequency of burning.

In savanna environments, � re is generally regarded as a tool to help maintain the grasscover and reduce the spread of shrubs and trees. In grazed moist savannas (average annualrainfall exceeding 600 mm) � re alone can control bush encroachment. This is not the casein drier environments such as the study area. Here, rainfall is too low and erratic to producea fuel load that could support � res occurring with enough frequency and/or intensity to, ontheir own, prevent the regeneration of bush from coppice and seedling growth. However,even less frequent � res, if combined with browsing animals such as goats (which arecommon in Kalakamate), have been shown to be successful in preventing bush encroach-ment.44 Research from similar environments in Southern Africa is therefore in agreementwith the conclusions drawn by villagers: that the prohibition on burning of grazing land is(at least partly) the cause of the increase in woody vegetation.

The increase in woody vegetation in the study area since 1964 or longer seems torepresent an unusual trend for communal areas in this part of the country (although it hasbeen described from communal areas in other parts of Botswana). The causes for thisincrease are to be found in a combination of events originating at the national and districtlevel. These have in� uenced the changes in local livelihood and land management strategieswhich, in combination with the speci� c natural environment and history of the study area(which differ in some respects from those of neighbouring areas), lie behind the observedvegetation changes. It was stated earlier in this paper that environmental and social changesare closely inter-linked, dynamic and variable in time and space, and therefore must beexamined at the local level. A study on the effects of land resettlement in Zimbabwe onwoodland resources draws similar conclusions.45 Observed patterns of woodland change in

42 See, for example, P.J. Crutzen and J.G. Goldammer (eds), Fire in the Environment: the Ecological, Atmosphericand Climatic Importance of Vegetation Fires (Chichester, 1993), p. 400; and R.J. Whelan, The Ecology of Fire(Cambridge, 1995), p. 346.

43 See examples and discussions in P. de V. Booysen and N.M. Tainton (eds), Ecological Effects of Fire in SouthAfrican Ecosystems (Berlin, 1984), p. 426, especially W.S.W. Trollope, ‘Fire in Savanna’, pp. 149–175. See alsoa number of contributions in J.G. Goldammer (ed), Fire in the Tropical Biota: Ecosystem Processes and GlobalChallenges (Berlin, 1990), p. 497, especially B.W van Wilgen, C.S. Everson and W.S.W. Trollope, ‘Firemanagement in southern Africa: some examples of current objectives, practices and problems’, pp. 179–215.Similar arguments are also put forward in W.S.W. Trollope, ‘Ecological effects of � re in South African savannas’,in B.J. Huntley and B.H. Walker (eds), Ecology of Tropical Savannas (Berlin, 1982), pp. 292–306; in P.J. Joss,P.W. Lynch and O.B. Williams (eds), Rangelands: a Resource Under Siege, Proceedings of the SecondInternational Rangeland Congress (Cambridge, 1986), p. 634; and in P. Stott, ‘Recent trends in the ecology andmanagement of the world’s savanna formations’, Progress in Physical Geography, 15, 1 (1991), pp. 18–28.

44 See especially B.W van Wilgen, C.S. Everson and W.S.W. Trollope, ‘Fire management in southern Africa’.45 J.A. Elliott, ‘Resource implications of land resettlement in Zimbabwe – insights from woodland changes’. Paper

presented at Conference on ‘Land Reform in Zimbabwe: the Way Forward’, held at SOAS, London, March 1998,p. 9.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


different resettlement areas were found to be highly place-speci� c, both when comparingoverall change in different communities and when looking more speci� cally at change indifferent areas within a community.

Firewood

Interview surveys from other villages in the district conclude that, at least in some areas,� rewood was becoming scarce, especially close to the settlements.46 Few species werementioned speci� cally, but a decline of Colophospermum mopane was noted for someareas. In Kalakamate, most respondents initially stated that it is harder to � nd � rewood nowthan when they were young. Some blamed this on the increase of the village population orstated that, ‘people come from outside settlements to collect here’. During the interviewsa direct question about the � rewood supply in general would often be answered withcomplaints of increasing scarcity. However, when the same respondent was asked aboutspeci� c species, a contradictory picture usually emerged. As can be seen in Table 3, whileopinions varied, most respondents felt that the abundance of several of the most preferred� rewood species had increased or remained unchanged. Only a few of the interviewedvillagers described a decrease.

Firewood is vital for food preparation and heating, especially in rural areas. In the 1981and 1991 censuses, 95 percent of the households in Kalakamate stated that they used woodfor cooking (while wood for lighting had decreased from 8 percent in 1981 to only 2percent in 1991).47 In general, women and children are responsible for � rewood collection,even though today some households transport large quantities by truck, and a few buy fromcommercial dealers. The species reported to be used most were C. mopane, Combretumapiculatum and Combretum hereroense, but many other species were also said to providegood � rewood or charcoal.

Table 3. Firewood species in order of preference: reported change


Colophospermum mopane 10 2 3Combretum apiculatum 4 1 2Combretum hereroense 3 1 –Acacia tortilis 14 3 2Dichrostachys cinerea 11 6 3Acacia nigrescens – 2 –Grewia � avescens 1 1 –Combretum imberbe 1 1 –Peltophorum africanum – 1 –Bolusanthus speciosus – 1 –

Numbers refer to how many respondents mentioned a particular change.Several of these species also provide browse or berries, and are thereforelisted also in Tables 1 and/or 4.

46 NEDCC, ‘Feedback on National Conservation Strategy Consultation Process’ (unpublished report, Francistown,Regional Agricultural Of� ce, North East District Conservation Committee, undated) , p. 3; Asselman, ‘NationalConservation Strategy Consultation Report for North East District’; and Kinlund, Does Land Degradation Matter?

47 CSO, ‘Summary statistics on small areas. 1981 Population and Housing Census’, Government of Botswana,Ministry of Finance and Development Planning, Central Statistical Of� ce (Gaborone, 1982), and Central StatisticalOf� ce, ‘Data from 1991 population census: computer printouts, transcripts’, Government of Botswana, Ministryof Finance and Development Planning, Central Statistics Of� ce (Gaborone, 1994).

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


The various and contradictory perceptions of change in abundance of speci� c � rewoodspecies are, to a large extent, due to the heterogeneous distribution of species through thelandscape. For example, on certain soils C. mopane dominates completely, while in otherparts of the village it grows scattered among other species. Stands of C. apiculatum areparticularly dense on soils derived from the basic rocks found in certain parts of the village.Therefore, the location of a homestead or � eld will result in different experiences of whatis common or not, and also of speci� c changes. Although woody vegetation in general hasincreased in the area, the complaint of having to walk further in search of � rewood nowmay well hold true. Before the nucleation of settlement, most homesteads were surroundedby woodland and bush, while today many homesteads are surrounded by other houses andby cultivated or recently abandoned � elds. The nucleation of settlements also means thatmore people now share the same spatial range for � rewood collection. Finally, old agenaturally makes the task of collecting � rewood harder and, compared with only onegeneration before, daughters and grandchildren have less time to help because of outsidework and school. Since the interviews targeted old people it is only natural that manyrespondents now experience � rewood collection as a more arduous task than that remem-bered from their youth. These factors combine to provide a spatially concentrated percep-tion of environmental change depending on speci� c patterns of movement (because of, forexample, age and gender) which is then generalised to the whole village.

This example demonstrates how important it is to analyse statements about change ina broad framework, which acknowledges overall spatial patterns as well as temporalchanges at the level of the village and the individual. Descriptions of change often tell amuch more complex story than that anticipated by the researcher. Differences betweenhouseholds in terms of access to � rewood have always existed but these are likely to haveincreased. Households with access to motorised transport bene� t from the increase inwoody vegetation in more remote parts of the village and households with enough outsideincome may buy � rewood from commercial dealers. For the poorer, and especially thosewithout the help of children or other close relatives, � rewood collection has become moredif� cult. These differences naturally cause contradictory perceptions of the scarcity ofnatural resources and, by extension, of the nature of environmental change.

Fruits and berries

About half the respondents who were asked stated that, in general, shrubs and trees thatcarry edible fruit or berries have become scarcer, but others had not observed any change.When asked about the abundance of fruits and berries on individual trees, many respondentsinitially complained of a decrease. However, further discussion revealed a general consen-sus that the abundance of fruit per tree varies with rainfall and that: ‘In a good rainfall yearnow you will � nd as much fruit as in a good rainfall year when I was young’. Manyvillagers were of the opinion that, although rainfall varies from year to year, there are, onaverage, fewer years now with good rainfall and therefore fewer years with much fruit: ‘Inolden times there was rain, so then you could collect these fruits in most years’. Someexplained how ‘the long period of low rainfall’ had caused the death of trees and a gradualdecline of fruit: ‘The trees have died due to lack of rain and now people don’t have theseeds of these trees, so they could not come back even in good rainfall years’.

The fruits and berries most commonly collected were from the species morula(Sclerocarya birrea), mogwana (Grewia bicolor and other Grewia spp.), mmilo (Vangueriainfausta) and morojwa (Azanza garckeana). Concerning the species morula and morojwa,most respondents claimed the amount of fruit, size of fruit and/or numbers of trees had

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


Table 4. Species with edible fruit or berries in order of preference:reported change


Sclerocarya birrea – 2 17Grewia bicolor 2 4 3Vangueria infausta 2 3 4Azanza garckeana – 2 7Ximenia americana/X. caffra – 1 3Grewia � avescens 1 1 –Strychnos madagascariensi s – 1 2Cissus quadrangulari s – – 3Lannea stuhlmannii – 1 1Securinega virosa – 1 1Grewia retinervis – 1 1

Numbers refer to how many respondents mentioned a particular change.Some of these species also provide browse or � rewood, and are thereforelisted also in Tables 1 and/or 3.

decreased. For mogwana and mmilo opinions were divided between a described increase,decrease and a situation of no change (see Table 4).

Other surveys from the North-East District contain little information about fruits andberries. In a report summarising several village meetings, respondents stated that morula (S.birrea), one of the most popular fruit species, had declined.48 However, it is not clearwhether this refers to a decline in numbers of trees or harvest per tree. Villagers in Morokasaid that morula (S. birrea), mogwana (Grewia spp.), mmilo (V. infausta), morojwa (A.garckeana), mogwagwa (Strychnos pungens), mogorogorwane (Strychnos cocculoides) andmoretologa (Ximenia spp.), all with edible fruits or berries, had declined in numbers.49

The results from Kalakamate are not as straightforward as those reported from the othervillages and it is suggested that the perceived change in the availability of fruits, berries andnumber of trees and shrubs must be analysed in the context of changing livelihoodstrategies and ‘individual landscapes’. Throughout the community the way of life haschanged dramatically, and it is common for changes within the social and physical sphereto be confused. As one respondent said, ‘There are not as many fruit-trees now as beforethe 1970s. See, now my granddaughter does not know these fruits’. Yet children used tobe responsible for much of the collection of fruits and berries, and today they have muchless time for such tasks. A common complaint was that children ‘no longer know aboutedible fruits and where they can be found’. In addition, access to cash income and a changein lifestyle has reduced the incentive for using wild fruits and berries: ‘People have replacedthese dishes with European food’. Thus, several respondents said that while they still collectsome fruits when they are ripe, they no longer bother to preserve and store them.

Where households still take care of fruits and berries this is the domain of women, evenwhen children do some of the actual collection. Women collect fruit and berries mainlyfrom trees in their gardens and � elds and from trees they encounter on the way to their� elds. As in the case of perceived changes in � rewood abundance, changes in the landscapewithin which women move in� uence their perceptions about the abundance of fruit and

48 NEDCC, ‘Feedback on National Conservation Strategy Consultation Process’; and Asselman, ‘NationalConservation Strategy Consultation Report for North East District’.

49 Kinlund, Does Land Degradation Matter?

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


berries. First, the nucleation of settlements and � elds means that more women now movein a spatially more restricted area, as compared to when the homesteads and � elds weremore scattered. Thus, competition for fruits and berries is higher than before. Second, forthe same individual, access (in this case geographical access) changes for individuals overtime, and the decline in fruits perceived by several of the really old women is, at leastpartly, due to their being less mobile now: ‘There is less fruit now. I could collect morewhen I was a young girl – but then, it was a long time since I collected any’. Finally, aswith descriptions of change in grass cover, the abundance of fruits and berries is consideredto mirror the amount of rainfall. Thus, if rainfall is perceived to have declined, then so mustfruits. This argument can also be turned around, so that if the abundance of fruits andberries is taken to depend on rainfall, and this produce is perceived to have decreased, thenthat means that rainfall has declined.

Grass cover

Local knowledge about grasses was much less detailed than the knowledge of trees andshrubs. The names of species used for thatching were known, as were the names of commonweeds, but only a few respondents, all men, were aware of any local names of foddergrasses. A few good fodder grasses were mentioned or pointed out, for example Urochloamosambicensis and Schmidtia pappophoroide s (both called tshwang), Urochloa trichopusand Chloris virgata. The discussion below refers mainly to grass cover in general, with afew references to speci� c species. The estimates refer to the availability of grazing. Grassspecies used for roof thatching are discussed separately.

Respondents were mainly divided between those who claimed that grass cover hadgradually declined due to a continuous trend of declining rainfall and those who stated thatgrass cover had always � uctuated with variations in rainfall and that: ‘A good rainfall yearnow would provide as much grass as a good rainfall year in the 1940s’. A few describedan increase, ascribing this either to good rainfall or to a decline in cattle numbers. Thosewho described a gradual decline stated that this had started well before the droughts of the1980s, but disagreed about whether the change had begun in the 1940s, 1950s, 1960s or1970s. This perceived long-term decline in rainfall is not mirrored in the of� cial rainfallrecords.

The aerial photographs (interpreted in Figure 3) were taken between March andSeptember (i.e. mainly in the dry season). In semi-arid areas grass cover is highly seasonaland in most years many areas covered by grass during the wet season will, towards the endof the dry season, be left with a very denuded cover. In black and white aerial photograph,grass is identi� ed by a grey colour and a slight difference in texture as compared to bareground, which looks whiter. Thus, for some environmental features, such as grass cover, dryseason photographs are not fully representative. In addition, even the grass still remainingin the dry season is often dif� cult to discern in such photographs because it is sparser andthus easier to mistake for bare ground. The perceptions about changes in grass cover areshown in Table 5. Again some interesting divisions emerged. For example, far more womenthan men stated that grass cover had drastically declined. It was more common for men todescribe a situation of continuous � uctuation. One woman said she had no opinion, adding,‘I do not know this because I am a woman, ask the men, they will know’. The oldestrespondents were more likely to describe a situation of � uctuating grass cover than thoseborn after 1940. However, most respondents stated that over the last decade grass cover hadbeen very scarce, ‘Cattle have had to rely much more on browse, such as mopane leaves,than usual’.

Only one respondent said that an increase of cattle could cause, or contribute to, a

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


Table 5. Change in grass cover since respondent was young

Opinions about changes in grass cover Women Men

There is more grass now – because there are less cattle – 3There is more grass now – because there is more rain 1 –There is less grass now – because there is less rain 10 5There is less grass now – because there are more cattle 1 –There is less grass now – because there is no burning – 1Grass cover always � uctuates with rainfall 7 11Does not know if grass has changed 1 –

The respondent s were asked to compare the period of their youth with presenttimes.

decline in grass cover. Even when asked directly, most maintained that a decrease of grasscover was not related to an increase in cattle numbers: ‘Grass is less in times of drought,and at those times there are few cattle anyway’. Many related the perceived decrease ingrass cover to an increase in woody vegetation: ‘Before, one could see far, there would bescattered trees with plenty of grass in between. Now there are more trees and shrubs andless grass’. In discussions about grazing land, respondents talked about the density andheight of grass, or of the cover of grass in relation to the cover of shrubs and trees. Onlywhen asked speci� cally about the effects of the fencing of the private farm and the erectionof the veterinary cordon fence would some say that this (of course) had caused a decreasein the area available for grazing.

The very strong in� uence of rainfall on grass cover was mentioned by all, while effectsdue to changes in human population, livestock numbers, the cover of woody species and thereduced frequency of burning were all considered to be less important, as illustrated by thefollowing comment:

On the land around our house there is less grass this year than before. For example, in the1960s and 1970s the grass was more dense as well as taller. This change is caused by lack ofrain, neither people nor cattle can change the growth of grass or trees. In a bad rainfall yearthe grasses will be less also in the hills where there are no cattle – so this change must be dueto lack of rain. When the rain falls grass will germinate from the ground, and its growth willnot be changed by grazing. Grass that has been grazed will still germinate in the same placenext year – so cattle cannot change the composition of species. In addition, there are plenty ofgrass seeds in the ground; soil dug up from a deep pit will still have grasses growing from it.

In response to both indirect and direct questions, none of the respondents thought there hadbeen any change in the inherent capacity of the land to maintain good grass cover. Mostwould spontaneously explain that when (or if) ‘good rainfall returned’ so would the grass,while others gave the same response when speci� cally probed on this issue. Severalrespondents gave detailed descriptions of the relationship between grass growth and rainfallunder conditions of varying dryness. For example:

The shortage of grass this year [1992] is caused by lack of rain. Also in the year immediatelyafter a drought the grass will not look the same as before [the drought], even if rainfall is good.This is because during a drought the grasses will produce less seeds, which means less grassthe following year. With other kinds of grass things are different. These will grow as tall anddense as before even in the year after a drought, since these grasses germinate from the sameroot. For the grass cover in general, it will have recovered after two years, i.e. if rainfall is goodfollowing a year of drought.

Most other interview surveys from the district give few details on perceptions of how grasscover has changed. One report simply states that, according to villagers, grazing had

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


declined over the previous decade50 and, in the district consultations for the NationalConservation Strategy, villagers similarly stated that grass had declined, and even disap-peared, and that this change was caused by droughts and the existence of freehold farms.51

In Moroka, villagers said that palatable grasses were becoming scarce, forcing livestock torely more on browse. A few species described as good fodder grasses were named and saidto have become dif� cult to � nd (for example Cynodon dactylon and Digitaria milanjiana),while species of lower fodder value (mainly Aristida congesta) had become more com-mon.52 The species C. dactylon, listed as a good fodder grass in Moroka, is in Kalakamatedescribed mainly as a troublesome weed in � elds.

Compared with the cover of woody species it is much more dif� cult to draw anyconclusions about changes in grass cover. Aerial photographs are of little help and relianceon interviews is complicated by the severe droughts of the 1980s, which have colouredpeople’s perception of change, and so make it more dif� cult to separate long-term changefrom the effects of the recent droughts. Respondents say that grass cover is primarily linkedto rainfall, an explanation in consonance with recent thinking in range ecology.53 For theimmediate past, this linkage would de� nitely have caused a decline. After exceptionallygood rains throughout the 1970s, there followed a period of almost 15 years with belowaverage rainfall. There are also spatially speci� c processes which may explain the relativelylow grass cover in certain places. Close to the main settlement this is due to increasedpressure by people and livestock. In the surrounding areas there are now large expanses ofabandoned � elds, covered by dense thickets of thorny shrubs effectively excluding grasses,whereas before such areas had a more patch-like occurrence. Many of these areas are closeto where villagers, especially women, often walk, thus giving them a possibly biasedperception of the study area as a whole. It is mainly the men who walk in the grazing areasfar from the main settlement and in areas where few � elds have been cleared in theirlifetime, and it was usually the men who stated that there had been no long-term changein overall grass cover.

Thatching grass

When discussing different fodder grasses, the men were more knowledgeable than thewomen and expressed stronger opinions about change. However, for thatching grasses (usedfor roofs) it was the women who knew where these species used to grow and where theymay still be found. Of the different local species used, magurwani (Heteropogon contortus),mutsikiri (Eragrostis pallens) and ‘rondavel grass’ (e.g. Hyparrhenia spp.) are preferred byall. These species are tufted, tall perennials common in Botswana.54 There was a generalconsensus that these species had declined dramatically and some women claimed they haddisappeared altogether (Table 6). Now people are forced to use inferior species such as voko(Eragrostis rigidior) or buy bundles of grass cut in Nata or even further north (egHyperthelia dissoluta and Sorghum versicolor). The descriptions given of the way thechanges had occurred were, however, contradictory. About half the respondents claimedthat the decrease had occurred gradually over a long time, ‘as rainfall has declined’, while

50 Fortmann et al., ‘Local institutions, village development , and resource management ’.51 Asselman, ‘National Conservation Strategy Consultation Report for North East District’.52 Kinlund, Does Land Degradation Matter?53 J.E. Ellis, M.B Coughenour and D.M. Swift, ‘Climate variability, ecosystem stability and the implications for range

and livestock development ’, in R.H. Behnke Jr., I. Scoones and C. Kerven (eds), Range Ecology at Disequilibrium(London, 1993), pp. 31–41; and B.H. Walker, ‘Rangeland ecology: understanding and managing change’, Ambio,22, 2–3 (1993), pp. 80–87.

54 D.I. Field, A Handbook of Common Grasses in Botswana, Government of Botswana, Ministry of Agriculture(Gaborone, 1976), p. 163.

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


Table 6. Observed change of thatching grass and perceived causes

Observations of change Women Men

Good species have disappeared 4 2Good species have decreased 5 1All species � uctuate between years 6 1Now people collect these grasses in other places 2 1Now people use other, inferior, grass species 3 –Still collects these species near house and � eld 1 –

Reason for changeLack of rainfall – a � uctuation of rainfall 9 1Overuse by people 1 –

the rest stated that the drastic change had occurred over the past decade (ie when rainfallhas been extremely low). Irrespective of this, all agreed that if years with good rainfall wereto become more common again, then so would the good thatching species. Only onerespondent mentioned a possible cause of decrease other than rainfall, and that wasindiscriminate collection by people ‘before the grasses have distributed their seeds’.

Interview surveys from other parts of the district report very similar perceptions ofchange. On communal land, a general decline of thatching grasses was described for theten-year period before the survey.55 Villagers said that these grasses could now only befound in the freehold farms. This situation was said to be caused by inadequate rainfall,increased demand due to population growth and the fact that people cut the grass before theseeds have dispersed.56 People complained that they now have to buy grass at high pricesfrom northern Botswana or from Zimbabwe.57

Conclusions

The results presented above show that observations, perceptions and explanations ofenvironmental change in the communal lands of North East District vary. The differentenvironmental histories recorded for Kalakamate, including the aerial photograph interpret-ation and � eld measurements presented here, also contain several contradictions. Thedifferent observations, perceptions and explanations of change have been examined sepa-rately, and an attempt has then been made to achieve a more complete picture of what hashappened, as well as why, by piecing different histories together as a coherent whole. Tothis end, speci� c environmental changes have been viewed as the direct and indirectoutcomes of events, cycles and trends that have occurred in the physical and social sphere,and these were related to the existing range of socio-environmental relationships found.

Several changes have occurred in Kalakamate. However, these were not always asexpected and, in many cases, they differ from those presented as typical for the communallands of Botswana. The common view about the communal lands is that human andlivestock populations have increased drastically. In Kalakamate, however, the populationincrease has been quite low and livestock numbers have � uctuated but show no increasingtrend. With respect to rainfall, of� cial records show no clear long-term trends ofincrease or of decline. The low population increase is not likely to have had any severe

55 Fortmann et al, ‘Local institutions, village development , and resource management ’.56 Asselman, ‘National Conservation Strategy Consultation Report for North East District’.57 Kinlund, Does Land Degradation Matter?

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


environmental impact, while short-term � uctuations in rainfall and in livestock numbers arenatural components of the system.

Since environmental change has occurred, the causes must be sought elsewhere. Theoverall cover and density of woody vegetation have increased somewhat, while grass coverhas decreased in some areas. Around the main settlement the area of bare ground hasincreased, although in other areas it has decreased. In summary, the major change does notlie in any overall change of the vegetation, but rather in the spatial distribution of differentvegetation types. This has primarily been caused by changes in the spatial pattern ofsettlement and cultivation, a shift initiated by government policies after independence. Therecent nucleation of settlements and � elds has meant that, today, few villagers utilise theoutlying communal areas, while competition for resources near the main settlement hasincreased. This has not only caused actual environmental changes, but has also changedpatterns of access and thus peoples’ use and perceptions of different natural resources. Theperception of the abundance of a particular resource varies between individuals dependingon how they move through the landscape. Independence can thus be seen as an event thatset in motion a gradual change in land use. Other events, such as the erection of fences thatcut off access to water and grazing land, and the law against burning which is likely to haveresulted in a shift from grass cover to woody vegetation, were also caused by outsideintervention. These represent single events, which have caused abrupt changes in land use,resulting in long-term environmental changes.

Other changes have been more gradual, such as the increased access to incomeopportunities outside the village (especially for women), which has reduced the dependenceon local natural resources. Similarly, an increase in education opportunities has alsochanged peoples’ perceptions as well as restricting the time they spend on other activities.Today, fruits and berries are not as important for the household economy, knowledgeabout these resources is declining among the young, and older women are hinderedfrom collecting as much as they used to by age, illness and distance factors. Today,some households would rather buy � rewood and building material than spend timecollecting it, while for others the distance to stands of pure woodland has increased. Thecontradictory statements, concerning whether the abundance of these resources haschanged, are more likely to re� ect changes in access and dependency, than actual changesdue to overuse. Grass cover around the main settlement has decreased because of the in� uxof people and cattle from other parts of the village and because more � elds have beencleared nearby. Women claim good thatching grass has declined, while most men have notnoticed a corresponding decline in good fodder grasses. These different observations maybe the result of actual changes in abundance, but may also represent changes in spatialdistribution and be caused by differences in how men and women move through thelandscape.

Thus, perceptions and interpretations of change will differ between people and fromplace to place. People, including scientists, perceive the same reality in different waysand different experiences and interests. Local and scienti� c accounts of environmentalchange are subjective in different ways. Villagers experience their surroundings in theimmediate context of securing a livelihood, while scientists collect data in order to explore,explain and increase predictability. Local environmental histories are of interest as theystand, and they provide scientists with important complementary and in-depth social andenvironmental data. A key issue in this respect is to understand why different ‘histories’differ, and how they can complement each other – rather than to prove one more accuratethan the other.

Data from Kalakamate demonstrate how people adjust to events and trends at the

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14


national level and how these shifts in livelihood opportunities result in changes in landmanagement, with both immediate and gradual effects on the environment. This is in linewith results from several recent studies from semi-arid areas of Africa, which show thatsocial and natural systems are extremely dynamic and complex, especially in areas wherethese systems are subjected to frequent interaction. In Kalakamate, overall environmentalchange has not been as dramatic or negative as is often assumed for communal areas. Thespatial distribution and abundance of resources have � uctuated over time and available dataindicate, contrary to entrenched of� cial beliefs, that most changes are transitory andreversible.

ANNIKA C. DAHLBERG

Department of Physical Geography, Stockholm University, S-10691 Stockholm, Sweden

Dow

nloa

ded

by [

Uni

vers

ity o

f W

este

rn O

ntar

io]

at 2

1:22

17

Nov

embe

r 20

14

landscape(s) in transition: an environmental history of a village in north-east botswana

Documents