We will not sit back and watch as the wealth builds up in the cities, while on the edges of these cities, in the small towns and in the countryside, we continue to starve.

This was the challenge issued by 700 representatives of 357 rural and landless communities at the National Land Conference in Bloemfontein in February 1994. It is a sharp reminder to the new government of national unity, preoccupied as it will be by the overwhelming need for jobs and housing in the urban areas, that the link between access to land and political freedom in South Africa is a fundamental one.

South Africa's countryside is highly disparate. First, it was divided into expansive tracts of white-owned farmland, where Africans' direct access to land was reduced to a bare minimum and whose proprietors could take advantage of generous state assistance, and small and increasingly congested African reserves where it became impossible for any but a very few to make a living off the land. Second, ecological conditions vary greatly between different regions, from the lush vineyards of the western Cape to the and sheep pastures of the Karoo; from the maize heartlands of the highveld, subject to the vicissitudes of drought and shifting interest rates, to the subtropical luxuriance of the eastern Transvaal lowveld. Third, rural people pursue a diversity of strategies to secure their own and their children's livelihoods, and combine these in different ways within each locality and region. Fourth, and above all, the countryside has been a place of struggles: those of farm workers and labour tenants and their families; those of dispossessed communities; those of landless people, confined to reserves, and increasingly dependent on access to urban wages. Often such struggles have been obscure, overshadowed by the 'struggle for the city'. Their central theme is access to land and use of land.

The Rural Heritage
There are many land questions in the new South Africa. What sorts of demands are being made for land? What land, of what kind, may be available for redistribution? How will it be acquired? Who will have access to such land? What criteria will be applied? How will they farm? What will they produce? Who will decide how they farm? How will decisions be made on each of these questions? What are the implications of the changing strategies of white and black farmers? What is the position of farm workers? How will they benefit from changes in land and agricultural policies? Or might they lead to new sources of dispossession? What lessons can be learned from experience both in South Africa itself and in other countries in the region, notably Zimbabwe? This issue contains diverse articles on different aspects of these themes and Briefings from front-line NG0s. It also contains an extensive Briefing by Morris Szeftel on the South African election of April 1994, and an Debates piece on the contradictions of multi-partyism and structural adjustment in Mozambique by Graham Harrison.

What do rural people want? Many different things, according to the list of demands presented by community representatives at the National Land Conference: restoration of land lost; no more evictions of labour tenants and of farm workers; basic services such as clean water and accessible health care; that rights to own land and inherit property under customary law should be extended to women; community representation in decisions on local development. The conference was organised by the National Land Committee, an umbrella NGO which embraces many different regional affiliates and which has devoted its energies to assisting and publicising campaigns over land restitution and greater security for farm workers and labour tenants. A report on the conference appears in the Briefings section, together with items produced by some of its affiliates.

Rural peoples' demands can only be understood in the context of the history of conquest and dispossession, of territorial segregation and political exclusion, of generations of state control of the movement of people and of the socioeconomic conditions of their lives. The claims of people to land have always been closely linked to the demand for political representation. The Act of Union in 1910 excluded black people from the franchise, except for a limited number of qualified voters in the Cape. The South African Native National Congress, later the African National Congress (ANC), was formed in 1912 to unify African opinion in opposition to their exclusion from the franchise and impending loss of rights to land. In 1913, the white parliament passed the Natives Land Act, which built on earlier republican and colonial legislation. It prohibited sharecropping contracts between white landowners and black peasant farmers. It also required the designation of 'scheduled' areas outside which Africans could not buy or rent land and inside which non-Africans could not acquire rights to land. At that time, the land 'scheduled' in this way represented about 7 per cent of the land area of South Africa. Most of the rest of the country was reserved for white ownership. Limited provision was then made, in principle, for the identification of further land to be 'released' for African occupation in due course, but legislation to facilitate this was not passed for more than twenty years. The Native Trust and Land Act of 1936 established the South African Native Trust (later the Development Trust) as a state agency with wide-ranging powers to acquire and administer such land. In practice, acquisition was partial and ponderous. Even once the nominal quota of land was 'released' in this way, the African reserves or 'homelands' would amount to barely 13 per cent of the land area of the country. The corollary of this limited provision of 'additional' land was the destruction of the qualified franchise for Africans in the Cape Province, despite a vigorous campaign of opposition from the 1935 All-African Convention.

The Trust regime in the African reserves was highly authoritarian, and Africans did their best to subvert it. Its regulations were inspired by the view that Africans did not know how to farm and that conservation measures had to be strictly administered by the central state. The 'betterment' and rehabilitation programmes were initiated in the 1930s and extended across South Africa, and the British colonies of southern and east Africa. They involved the separation of residential, arable, grazing and woodland and the resettlement of people in planned villages. In South Africa, as everywhere else, they provoked enormous resentment, and in many regions bitter revolt, over the culling of livestock, the cutting of arable lands, the enforced removal and concentration of settlement, and the imposition of Native/Bantu Authorities through which the state sought to enforce its policies. They were designed to improve production and conserve the soil in the African 'reserves' without altering the distribution of land.

These programmes of land rehabilitation were incorporated into the National Party Government's grand strategy of territorial segregation, population resettlement and political exclusion, in terms of which all Africans, including those settled for generations in 'white' South Africa, were to be politically associated with one or other Bantustan. They could exercise political rights, pursue business enterprises and hold land rights only in the bits and pieces of 'homeland' associated with their 'own' ethnonational group. This was a vast experiment in 'ethnic cleansing' whose consequences were viciously destructive. 'Black spots' - fragments of black-owned land surrounded by 'white' countryside - were eliminated, and whole communities were dispossessed and forcibly removed. Hundreds of thousands of people were dumped in remote and barren settlements in the Bantustans. Conflicts whose origins lay in acute competition for scarce material resources - residential sites, schools, basic services, permission to seek employment - developed along ethnic lines, because Bantustan citizenship became the criterion for distinguishing 'insiders' and 'outsiders'.

Agricultural production in the Bantustans was constrained by lack of access to land, transport and markets. 'Betterment' only made things worse. For most rural families in the Bantustans, agricultural production declined over the last 40 years and came to provide a minute share of family incomes. Most households in the Bantustans survived only through the incomes of migrant members employed in the mines or in manufacturing industry or through the meagre pensions paid to elderly people. Government schemes increased overall agricultural output, but at costs in excess of returns. In some areas, farmers with access to cattle and the ability to acquire tractors have been able to farm successfully and to expand their access to land by sharecropping with others. Some of these have been beneficiaries of political patronage; others have not - these include former labour tenants resettled in Bantustans. Cane growers farming small plots in KwaZulu and larger areas in KaNgwane have increased production, generally as contract farmers for the sugar mills. Like the country as a whole, the former Bantustans are diverse. Access to resources is differentiated and determined, to a significant extent, by access to incomes from the urban economy, trading activities or the local state.

In the 'white' countryside itself, legislation, accentuated by private and public coercion, sought to secure farmers an adequate supply of cheap labour which would be subordinate to their authority. For more than a century, official policy has sought to turn African sharecropping peasants first into labour tenants and then into wage labourers. Africans responded in various ways: migration with their stock, evasion of controls - often in collusion with landowners, and resistance to eviction and their loss of rights to land and grazing. Farm workers were tied to the farms by extremely low wages, tight legal restrictions on their freedom of movement and their dependence on the farmer's goodwill for a home for themselves and their families. Labour tenancy survived its legal abolition, itself removed in 1986. For most Africans on the farms, their decline in access to land to grow crops and keep stock was not matched by improvements in wages. The adoption of combine harvesters and chemical technologies from the 1960s led to a decline in the demand for labour and evictions of many people from their jobs and homes. The owner-managed farm, using the most up-to-date technology and employing wage labour, has been the ideal for 'progressive' farmers since the beginning of the century; it has often produced insecurity for and eviction of farm workers.

Chris de Wet discusses the experiences of people who were relocated from land and homes both within the Bantustans and the 'white' farms and small towns and the consequent disruption to the lives of individuals and of communities which followed from various forms of resettlement and 'betterment planning'. He draws attention to the issues which must be considered in any future plans to relocate people again as part of a strategy of restoring or securing their access to land. White farmers had access to a generous infrastructure of state assistance for commercial agriculture, above all, cheap credit through the Land Bank, and an increasingly elaborate framework of subsidies, guaranteed markets and price controls. Networks of political patronage closely linked the National Party and the white farmers with three of the four provincial Farmers' Unions and the supply and marketing co-operatives. Land reform is not simply a question of the redistribution of land. Its outcome will also depend on re-structuring complex chains of forward and backward linkages in the interests of all farmers and of consumers rather than of large-scale farmers, input and wholesale monopolies, and the supermarket giants.

Land Reform: Redistribution and Restitution
As the new South Africa emerges, many difficult questions arise. The ANC's Reconstruction and Development Programme (RDP) identifies a 'national land reform programme' as 'the central and driving force of a programme of rural development'. Its two aspects are redistribution of 'residential and productive land to the poorest section of the rural population and to aspirant farmers' and restitution for those 'dispossessed by discriminatory legislation since 1913'. The priority is restitution. The ANC has committed itself to a Land Claims Court to arbitrate conflicting claims and deal with urgent claims for restitution. President de Klerk appointed an (Advisory) Commission on Land Allocation (ACLA, later CLA) in 1991, which resolved very few of the claims placed before it, not least because many claims involved land which the state had transferred to whites. The new government's Department of Land Affairs is already trying to deal with a number of claims through negotiation. Where agreement is not reached, people will have to wait for the Land Claims Court. This procedure will be expensive, cumbersome and protracted; but it is essential if there is to be a start to restoring to people the rights which were overridden in the past, and it is vital to the political credibility of the new government. The constitutional provisions to protect property against expropriation mean that it will be expensive to compensate those who have since acquired the land.

John Sharp's article on the Komaggas reserve, in Namaqualand in the northern Cape, demonstrates that some land claims fall outside the criteria for restitution that are likely to be applied by a Land Claims Court. He describes how a community's claims to land have been bound up with conceptions of their identity and involved recourse to different criteria in a changing political and legal climate. Recently, they have raised the issue of the original rights of their Nama foremothers rather than appeals to Victorian treaties. Either way, their claims to land now held by the state and mining and power companies predate 1913, the likely cut-off date for claims under relevant legislation.

The restitution of land leaves open the question of how land and other rights are allocated among members of dispossessed communities and how they may use the land. The official paper explaining the 1993 Provision of Certain Land for Settlement Act recalled the 'three key land policy objectives' which informed the 1991 Abolition of Racially Based Land Measures Act; namely, 'the broadening of access to land for the entire population, the upgrading of the quality and security of title on land and the judicious utilisation of land as a national asset.' These are outlined in a Briefing on 1993 land legislation. It continues to insist on limiting settlement of both 'families and small and/or large livestock units' to the 'carrying capacity of the land'. As Harald Winkler's Briefing, 'New Methods of Control' shows, the outgoing administration in the Transvaal sought to restrict the numbers of people who could take up land and dictate how they used it. Jocelyn Alexander demonstrates how the government in independent Zimbabwe has tried to impose on both resettlement and communal areas the forms of residential and land use planning embodied in the Rhodesian Native Land Husbandry Act of 1951 and in South African 'betterment' schemes. On the point of leaving the office of President, F. W. de Klerk signed a trusteeship of land in the former KwaZulu to King Goodwill Zwelithini which may remove it from the direct control of the new regional government in KwaZulu/Natal. It is likely to consolidate the existing political structures of 'tribal authorities' and their gendered criteria for allocating land. It is unlikely to facilitate women's ability to claim rights to land and a recognised voice in the affairs of their communities.

Cherryl Walker here analyses the conflict between the accommodation of 'tradition' the formal structures and the popular discourse of rural patriarchy - and the ANC's commitment to promoting gender equality. She argues that the two political objectives are fundamentally incompatible and that the language of the Reconstruction and Development Programme (RDP) betrays a rhetorical indulgence of the struggle for gender equality rather than a strategic integration of it.

For the vast majority of people who have no viable claims to recover land, there are proposals for redistribution of land. Several lines of argument have been advanced against land redistribution. The first assumes that large-scale farms are able to adopt more advanced mechanical, chemical and biochemical technologies which will realise economies of scale and that small-scale farms are incapable of producing sufficient foods to meet South Africa's domestic needs and supply its export markets. The second argues that black people want jobs, not land, and that the debate on land redistribution is a diversion. These two views are linked by common assumptions about the necessary directions of historical progress.

State policies in the past subsidised the acquisition of machinery and extension of grain production into ecologically fragile areas. They encouraged the consolidation of land into ever larger farms and prohibited its sub-division below a legal minimum. Rising costs and falling prices have partly reversed the trend to highly-mechanised grain production. However, farmers have tended to shift towards stock farming, reducing employment opportunities for farm workers. In some regions, farms have been converted to game farming and even 'conservation', reproducing images of a landed aristocracy and further reducing opportunities for access to jobs or land. Both employment and residence of blacks on white-owned farms have continued to fall for the last two decades. The 1991 census suggests that 200,000 people have been displaced from white farms in the Orange Free State alone since 1985 - and evictions have continued since 1991, not least to pre-empt the extension of rights to farm workers and labour tenants.

The RDP talks of 'a dramatic land reform programme to transfer land from the inefficient, debt-ridden, ecologically-damaging and white-dominated large farm sector to all those who wish to produce incomes through farming in a more sustainable agricultural system'. On the other hand, it recognises that the 'present commercial agricultural sector will remain an important provider of food and fibre, jobs and foreign exchange'. Large-scale farmers have benefited from extensive subsidies and adjusted their strategies to the tax and price regime. This does not prove that they are 'inefficient' in using resources - though the new government cannot afford to sustain this largesse. Within the system of large-scale farming, there are 'economies of scale' in the use of machinery and the farmers' own managerial capacities. It does not follow, however, that smallholder farming would be ,uneconomic'. There is historical and contemporary evidence, both from South Africa and elsewhere in Africa, that smallholders make effective use of the resources at their disposal and respond actively to market opportunities. In Kenya and Zimbabwe after independence, smallholders dramatically increased their production as high-value crops (in Kenya) and favourable prices (in Zimbabwe) were made available to them. Redistribution of white-owned land did not lead to a collapse in farm output; in Kenya, production increased markedly. What has proved expensive is the cost of state strategies to resettle farmers, to provide them with inputs and direct their priorities, as the successive settlement and farmers' support programmes funded by the Development Bank of South Africa, among examples from many other countries, indicate. The question is how existing and prospective smallholders can get access to resources and opportunities.

Black families want to have access to land as well as to jobs. They are well aware of the vulnerability of farmers to drought and debts and of the need for income from outside farming to secure household needs and the costs of farming. Rural workers confront declining employment opportunities in farming, and in mining, and many have had to earn what income they can find in the urban settlements which have expanded in the erstwhile Bantustans. Farm workers, and former farm workers, have relevant experience of farming, but lack capital resources to invest in farming. 'Yeomen politicians', as Merle and Michael Lipton (IDS Bulletin 25, 1, 1994) have called them, might be better placed than farm workers to take advantage of the opening of land to black people. Further, people want land, in rural and in urban areas, for a variety of purposes. They want somewhere to live; they want land to grow food to eat and to sell; they want a place to keep chicken and goats, to provide grazing for sheep and cattle, they want a place to return to, and land they can rent, sell and pass on to their heirs. People will change the ways they use land as circumstances change, and in different ways from one another. They will not all give priority to production of crops or stock for the market. They do not generally like to be told what to use it for or how to use it.

There is no single 'moment' of transition. Changing farming strategies have evolved over time in response to changing economic conditions. White farmers felt economically betrayed by the Nationalist government long before they felt politically betrayed. The life-chances of both farmers and farm workers are more deeply affected by the forces of drought and economic recession, of positive real interest and adverse terms of trade, than they are by the forces of political transition. William Beinart explores the shifting strategies of white farmers in a zone of arable and pastoral farming in the southern Orange Free State. He found that farming strategies, and their implications for employment, tended to vary, in part, by size of farming enterprise, and that smaller farms were quite intensively worked. Strategies to bring about change need to start from an understanding of how things are changing and what the actual consequences may be, whether or not they are intended.

Andries du Toit points out that discussions about land reform and rural restructuring in the former 'white' farming areas tend to ignore most of the people who live and work there. Farm workers are exploited and insecure. They have been locked into a repressive paternalism which subordinates them to die boer se wet - 'the farmer's law' - beyond the formal legal constraints to which they have been subject. The struggles of farm workers throughout South African history have been about access to land and security, to respect and to rights in law, to run their own lives and make their own decisions. They take place 'within and against' the terms of paternalist discourse, practices and institutions. The 1994 Agricultural Labour Relations Act, which is outlined in a Briefing, does not of itself change social relations on the farm but it does create a framework in which relations of power can be contested and changed. The problems and struggles of farm workers have to be recognised in their own right, not subsumed under either the struggles of urban workers or the broad rubric of land reform and the agrarian question. Du Toit observes that, in the process of political transition in South Africa, policy debates have been cast not in political terms but as development problems, opening the way to 'a technocratic discourse' transcending past political and ideological divisions. These make it possible for the late government, the ANC, the World Bank, academics and consultants, NG0s and research institutes to join together in formulating policies. This process is one both of strategic opportunity and of tortuous political difficulty.

Setting the Agenda
In 1992 and 1993, the World Bank entered a series of dialogues with policy-makers concerned with housing and urban issues, education, health, land and agriculture and macro-economic strategy for the 'new South Africa'. Through the Land and Agricultural Policy Centre (LAPC), it funded a series of reports, mainly by South Africans, whose findings were incorporated selectively into the World Bank's 'Options for Land Reform and Rural Restructuring'. Many activists in affiliates of the National Land Committee were uneasy about this process, fearing that instead of developing local research capacity it would limit the scope for South Africans to think through the issues and define their own policy agendas.

'Options' clearly states its 'guiding principle ... political and economic liberalization. At the heart of such a process would be a new agricultural pricing and marketing policy and a program for land reform'. It wishes to extend recent policies of abolishing subsidies, removing current regulations and liberalising markets which will, it argues, reduce the unfair advantages which state policies currently confer on large-scale producers. The RDP similarly argues for 'removing unnecessary controls and levies as well as unsustainable subsidies' to the large-farm sector.

The World Bank argues that its models 'indicate a substantial increase in rural employment and income as a result of land redistribution'. 'Options' envisages and costs at market prices a substantial transfer of perhaps 30 per cent of medium- to high-quality land from large-scale white to small-scale black producers. The RDP commits the ANC to these targets and to have a land reform programme in place in 1995. 'Options' distinguishes between those who might receive a 'basic grant' which could ,pay for a major share of a rural housing site' and 'individuals or groups who will use land in a productive manner', who could receive a grant for the cost of half the land and a loan for, say, 30 per cent; and who could commit 20 per cent of the costs themselves. State land, and private land which may be acquired for redistribution, is not usually empty of people; farm workers, labour tenants and squatters may all make a prior claim to land and not welcome its allocation to incoming tenants. They may find it difficult themselves to raise the capital costs of entry into the scheme. The cost of such a redistribution is claimed by the World Bank to be 'surprisingly small'. 'Options' estimates the public costs of settling over 600,000 smallholdings on 24 million hectares of agricultural land in four fertile regions at R17.5 billion - R3.5 billion (c. US$1 billion) per annum over five years. This would be funded from past, current or future taxes. It would compete for resources needed, for example, to provide safe water, roads and schools in the former Bantustans as well as with the more vocal demands from the cities. The estimates are based, not on evidence, but on indicative models. The outcomes may change markedly if the assumptions are varied. The key elements are land prices, yields, crop prices, target incomes, interest rates and administrative costs; these determine the costs of settling families and the numbers which can be settled. If farmers are to repay loans at more realistic interest rates than the models adopt, farmers' incomes would fall by between about 10 and 30 per cent for three of the four regional models. In that case, more support would be needed to bail out project beneficiaries and less land would be available for others. The annual sum envisaged is more than the R2.5 billion the new government has allocated to the entire Reconstruction and Development Programme for 1994-95, and more than 40 per cent of the R40 billion planned for the RDP over five years.

The World Bank criticises Zimbabwean resettlement schemes for imposing cultivation rules on settlers and expecting them to give up urban employment. It insists that its models are illustrative and are not intended to guide the way beneficiaries use their land or as targets 'driving the planning process'. 'Options', unlike the prior terms of reference for the research it drew on, avoids the term 'resettlement'. It is, nevertheless, difficult to see how a programme to move 600,000 families over five years could be anything but 'a vast resettlement project'. 'Options' envisages a common framework within which broadly similar processes of land redistribution would be managed. It is hard to imagine how strategies appropriate to one of South Africa's diverse regions, let alone the different localities within them, would be applicable in another. Nor can one envisage such a programme responding to the diverse claims of very different groups of people - relocated communities, farm workers and labour tenants, migrant workers or ex-workers, widows and single mothers, small-scale cane growers, aspirant farmers - to acquire land for diverse purposes - as somewhere to live, to plant crops, to graze stock, to return to, or even to rent out at a profit. The World Bank's 'Options' seeks to square a number of circles: redistributing land and maintaining agricultural production; providing for the poor while settling people who have the resources to take up land and cultivate it commercially; setting up a national programme yet implementing it at the local level - through the newly created regional governments. These dilemmas are not resolved. The ANC, for its part, needed a plausible land reform policy and 'Options' provided the basis for it. The ANC could not afford the rhetorical luxury of the Pan Africanist Congress's demand to restore the land to the people. But how will it work out in practice?

The Danish agency, DANIDA, is funding the LAPC to carry out a second stage to the programme. This was intended to 'field test' land reform proposals in different regions. Organisations were invited to 'tender' to undertake research within a structure of national and regional managers. Field research needed to draw on the skills and political credibility of the local affiliates of the NLC. This gave the regional research groups some scope to undermine the technocratic style of the new programme and to define their own research objectives. However, these are at variance with the commitment of the funders and the new government to initiate pilot projects. The research may not inform policy; it does make it possible to claim that policies are based on the findings of science. In any case, land reforms are not like new seed varieties or fertiliser combinations. They cannot just be tested in the field. This research process itself illustrates the tension between the need for policy-making at the national level and the need for sound empirical research at the regional level on recent trends in land ownership, indebtedness, land use, and land claims. Experience in different regions through 1994 demonstrates the complexities of resolving this tension between differing strategic priorities, and also the conflicts that arise between different organisations, with different histories and priorities, striving to cooperate to implement a regional research programme.

One regional research programme that has recently come to fruition is the MacArthur Foundation-sponsored project on 'Community Perspectives on Land and Agrarian Reform in South Africa' (CPLAR), Its programme, led by Dan Wiener and Richard Levin who discussed 'The Agrarian Question and Politics in the "New" South Africa' in ROAPE 57, was explicitly committed to developing a participatory research method by which communities were actively represented in the shaping of the research process, and the conduct of the research was integrated into the local procedures for the management of conflict. Community representatives took part in presentation and discussion of the findings.

Land in the Cities
Land questions are not restricted to rural areas. The apartheid state sought to segregate urban land as it did rural land ownership. The decline of the apartheid system, and of the state's capacity to impose its policies, led to renewed and often violent struggles to gain access to and control over urban land. lain Edwards examines the complex and contentious history of Cato Manor in Durban, the struggles of its residents to make a living and find a place to live and the conflicts to which this gave rise in the 1940s and 1950s. The state cleared the areas of its African and most of its Indian residents but, as with District Six in Cape Town, was unable to develop it for white residents. Protagonists in current political struggles over access to and control of land and housing in Cato Manor lay claim to alternative interpretations of this history. Edwards' historical account sheds light on one aspect of the political divisions between Africans and Indians in Durban, which have arisen out of the ways in which the whites who controlled local, provincial and national governments excluded both from the areas and resources which whites sought to reserve for themselves.

People seeking land for housing in urban areas come from the overcrowded houses and backyards of the townships more often than from rural areas. Election victory for the ANC has encouraged homeless people in the major cities to stake out new claims to their own land. This has opened up new opportunities for political entrepreneurship and confronted local, provincial and national governments with harsh dilemmas. Evictions ordered in the middle of the coldest night of a highveld winter by the Johannesburg City Council, still controlled by the National and Democratic Parties, attracted strong criticisms. Joe Slovo, the national Housing Minister, warned that some land invasions were orchestrated 'by outsiders ... for their own personal and political gain'. The ANC regional government has stopped further evictions. Housing is a priority for the RDP. The new government faces the problem of mobilising the resources to build the houses and to decide through what mechanisms and by what criteria to allocate houses and who shall get what land on which to make their own homes.

In Conclusion
Some lessons are clear from historical studies of South Africa and other countries in Africa. Issues concerning land will continue to be central in the lives of rural people. State policies will continue to shape the extent to which people get access to land and the ways they can use it, but will not necessarily do so in the ways in which planners and policy-makers intend. In considering policies we need to try to understand, in inconvenient detail, the complex and protracted changes that have been taking place in different parts of the country which operate with time-scales that transcend the period of 'political transition' or the horizons of planners.

Questions about land are also questions about much else - gender relations, generational differences, labour and employment, access to markets. They are rural and urban and crucially involve the connections between town and country. They are about class formation, class privilege and class power; these are interlinked with the process of the reformation of the state, at all levels, state allocation of resources and state power. Bureaucrats and businessmen will be week-end farmers. Like their predecessors, with whom they will share power, they are likely to take advantage of their privileged access to state patronage. They may allocate land publicly with one hand and receive land privately with the other. Regional governments, which have little capacity to raise their own taxes but will be responsible for spending a large share of the taxes raised through the centre, are likely to become major dispensers of local patronage. They will incorporate politicians, officials and businessmen from the former Bantustans who have already demonstrated their ability to acquire land and other resources from the local state. For those without these advantages, the capital resources needed to cultivate land or to invest in livestock are likely to be derived, in large part, from the earnings of urban migrants. In recent decades, numerous people have been shut out from such sources of income. Opportunities in the countryside will, in large measure, be differentiated by people's capacity to acquire resources in the towns. The variety of links between the rural and urban economies must be well understood both by policy-makers and by their critics.

In the light of the dismal history of 'top-down' state intervention in rural development, strong institutions of civil society - trade unions, women's organisations, farmers' associations, religious institutions, local associations of one kind or another, and service organisations - are all necessary, if by no means sufficient, if rural transformation is to benefit poor people. They need to be able to maintain their autonomy from, yet gain access to, the local, regional and national state and a capacity to have their voices listened to. During the apartheid years, society was uncivil in the extreme. In nearly all regions of the country rural people, above all, voted for the ANC and look to it to bring changes to their lives. After years of struggle against a brutal white supremacist state, people cannot now invest their faith uncritically in the promises of a new state, even one which has declared its commitment to undo the injustices of apartheid. Apartheid's legacies will live on beyond its demise. Nor can the new state assume political support for further 'top-down' measures of reform. The challenge is how to resolve constructively the inevitable tensions between national policy-making and the diverse needs and conflicting demands of the people in the different regions and localities which make up South Africa. Coherent state structures are required both to carry through ambitious programmes of support for disadvantaged people and to adjudicate conflicting demands.

Against those who question the relevance of land reform in modern South Africa, we emphasise the diverse claims of people in different localities and circumstances to get access to land for a plurality of purposes. Over the coming decades, we expect that substantial areas of land will be transferred from white to black ownership and occupation, even if not on the scale and at the pace outlined in the World Bank's 'Options' and by the RDP. The critical problem is to find ways of enabling as many people as possible to meet their needs for security and for productive activities, lest the majority of people continue to find themselves excluded by state policies and market criteria from being able to provide for their needs.

Colin Murray, Gavin Williams