FORGING THE LINKS BETWEEN HISTORICAL RESEARCH AND THE POLICY PROCESS 18- 19 SEPTEMBER 1999 UNIVERSITY OF THE WITWATERSRAND Land Tenure in South Africa’s Communal Areas: A Case Study of the Arabie-Olifants Scheme Edward Lahiff A symposium hosled by tlic History Workshop, the University of the Witwatersrand LAND TENURE IN SOUTH AFRICA’S COMMUNAL AREAS: A CASE STUDY OF THE ARABIE-OLIFANTS SCHEME Paper presented to the History Workshop at the University of the Witwatersrand, Johannesburg 18-19*" September, 1999 Dr Edward Lahiff Nkuzi Development Association, Pietersburg 1. Introduction This paper looks a( the evolution of the land tenure system, and the prospects for tenure reform, in one of South Africa’s black communal areas. It is based on a case study of the Arabie-Olifants irrigation scheme, located approximately twenty kilometers south of Lebowakgomo in the former ‘homeland’ of Lebowa, now part of Northern Province. The study was funded by the International Water Management Institute (IWMI) as part of its ‘Water for People’ programme In South Africa. Research activities included a review of the literature on land tenure in South Africa, empirical analysis of official records from the Arabie-Olifants scheme and the Deeds Office, Pretoria, and interviews with chiefs, members of Tribal Authorities and Transitional Local Councils (TLCs), government officials, farmers’ representatives and individual plot-holders on the Arabie-Olifants scheme. Fieldwork was conducted during the period February to May, 1999. 1. The Tenure Debate in South Africa The debate around tenure reform in South Africa is not a new one. Indeed, it has long been a central feature of the wider debates around access to land by African people under colonialism, segregation, apartheid and, latterly, democracy. This debate has pitied social and economic reformers of various hues against successive governments, with political expediency generally gaining the upper hand. Today, as in tlie past, the debate centers around individual versus collective control of land, and western versus African concepts of property. As early as 1905, the South African Native Affairs Commission had this to say: "Recognising the attachment o f the Natives to the present advantages o f their own communal or tribal system o f land tenure, the Commission does not advise any general compulsory measure of sub-division and individual holding o f the lands now set apart fo r their occupation; but recommetuls that movement in that direction be encouraged, and that, where the Natives exhibit in sufficient numbers a desire to secure and a capacity to hold and enjoy individual rights to arable plots and residential sites on such land, provision should be nutde accordingly under well-defined conditions. . . ". (quoted in Davenport and Hunt 1974: 40) In a similar vein, die Native Economic Commission of 1932 (Union of South Africa, 1932: 23) observed that: "The granting o f a title deed and the beaconing off o f plots is not enough to secure economic progress . . . i f production is to be increased, holdings must be consolidated rather than further sub-divided, and unrestricted grazing rights on communal land must go ". By mid-century, the Tomlinson Commission (Union of South Africa, 1955: 151) reached the by now familiar conclusion: “A revision o f the systems o f land tenure is regarded as one o f the prerequisites to the stabalisation o f the land in the Bantu Areas and the full economic development o f their potential ...The Commission recommends that in areas where the Bantu desire that their land should be granted to them under title deed, this should be done and the existing forms o f tenure should he superseded by such grants ... The principle o f 'one-man-one-lot'... reduces every Bantu to a low level o f uniformity with no prospects o f expanding his activities nor o f exercising his initiative. It is essential to make opportunities fo r the creation o f a class o f contented full-time Bantu farmers with holdings o f sufficient size to enable them to farm profitably and to exercise their initiative and to develop according to their individual ability and resources". Offensive language aside, such sentiments will be familiar to anyone engaged in current debates around land reform in South Africa (see Cross and Haines 1988; Levin and Weiner 1997). While the general thrust of this debate is towards the need for more individually- based forms of land-holding (e.g. freehold). Cousins (1996: 173) and the Department of Land Affairs (1997), amongst others, stress the benefits of the communal system, especially communal grazing, in terms of social equity and environmental management. 3, Land Tenure In The Homelands The pattern and forms of land-holding and land use in the former homelands have been directly influenced by the policies and actions of the South African state (in its various forms) in pursuit of racial segregation and the promotion of an oppressive migrant labour system. State policy on land in the reserves/homelands since 1948 has been based on a number of key elements, described by Hendricks (1990: 162) as the 'three rural pillars of Apartheid' - namely the soxalled communal form of tenure, the system of tribal administration ('the chieftaincy'), and various forms of rural planning and development, generally referred to as betterment. To these may be added a fourth important element - the forced removal of millions of black people from 'white' farms and towns to the reserves/homelands, which began in earnest in the Free State with the Natives Land Act of 1913, and accelerated dramatically throughout the country in the 1960s and 1970s. The roots of the forms of tenure found in the homelands of the Northern Province can be traced back at least as far as the mid-nineteenth cenmry. The settler government of the Transvaal as early as 1855 (Resolution 159) precluded anybody who was not a 'burger' from owning land in the Transvaal while simultaneously precluding 'natives' from burger rights. Following the first Anglo-Boer War, the Pretoria Convention of 1881 (Article XllI) laid down that 'Natives will be allowed to acquire land, but the grant or transfer of such land will in every case be made to, and registered in the name of, the Native Commission’ (quoted in Davenport and Hunt 1974: 40). For only a brief period, between 1905 and 1913, Africans in the Transvaal were allowed to acquire land in their own name. Since that date, African people in the Transvaal (and elsewhere in South Africa), continue, with the exception of certain urban areas, to live under a variety of tenure systems that deny them full rights of land ownership. The great majority of land in the former homelands is held under some form of communal tenure. Other tenurial forms include freehold land held by individuals and groups, including church missions, and state land, but these account for relatively small areas. Communal land tenure in South Africa is a hybrid form, specific to the homelands, which combines elements of individual and collective property rights. Although having some basis in African customary law, communal tenure has been greatly modified by successive governments over the course of the twentieth century, while alternative forms of land-holding were effectively denied to black people by law. Audiors such as Lacey (1981), Haines and Cross (1988) and Hendricks (1990) argue that communal tenure was an essential component of the migrant labour system, facilitating the concentration of the maximum possible number of Africans in the reserves/homelands, preventing the emergence of a stratum of rich peasants or capitalist farmers and providing the basis for a high degree of social control through the tribal leaders who controlled access to land. Legally, most communal land (with the exception of 'bought farms') is nominally owned by the state, but is generally held in trust for specific tribal communities and allocated by the chiefs to people living under their jurisdiction on a usufructuary basis (Budlender and Latsky 1991: 121). In popular perceptions, virtually all categories of land in the homelands are believed to belong to the community, or the chief (whether in a moral or a legal sense), despite the fact that formal title (in the form of deeds) are, in most cases, held by the state. Land administered under the communal system can be divided into three broad categories, although the differences (at the level of law and popular perceptions) are often blurred. At the heart of most of the homelands lies land which is generally referred to as 'tribal land' (or tribal farms). This is land which was occupied by tribes prior to the passing of the 1936 Native Trust and Land Act and, in many cases, without interruption since pre-colonial times. Nominal ownership to most of this land was appropriated by the state at various times prior to 1936, but this was not generally accompanied by any change in occupation or land use. This land makes up most of the land 'scheduled' for black occupation under the 1913 Land Act, Nominal ownership of this land passed to the South African Native Trust in 1936, but again diis brought little or no immediate change to die inhabitants. The 'reforms' introduced by the Trust after 1936 were mainly focused on newly-acquired land ('Trust farms'), with the result that the older tribal farms have relatively greater continuity of occupation and social structure. From 1936 onwards, the Native Trust (later, and somewhat euphemistically, the South African Development Trust, or SADT) set about purchasing thousands of farms for addition to the then native reserves. This land was generally allocated for the use of specific 'tribal communities' and, according to the 1936 Act, was held in trust by the State President (or officials appointed by the President). The Trust also set about buying-up much of the privately-owned land within the enlarged reserves, including mission land and land belonging to whites. In addition, the Trust acquired nominal ownership to state land ear marked for inclusion in the homelands (released areas) and all tribal farms (scheduled areas) not in private ownership. The third category of land that constimtes today's black areas is privately-owned land, typically land that was bought in undivided shares by groups of named black farmers. Such purchases were made in the scheduled areas prior to 1913, outside the scheduled areas until 1936, and from the Trust after 1936 (Vink 1986; 33). While some groups were successful in having title deeds issued in their own name, others were obliged by the racial laws of the day to register the land in the name of a tribe or state otficial, to be held in trust for the named purchasers. Over time, the sense of private ownership would appear to have faded (if indeed it ever existed) and today, most such land is used and administered in a way that is indistinguishable from other communal land. As suggested above, popular perceptions of land ownership do not tend to correspond fully with the otficial legal position. On tribal farms, the sense of community ownership is probably the strongest, based on uninterrupted occupation, strong historical claims and a relative lack of state interference over the years. These are the bastions of chiefly power and there is little or no awareness of ( and certainly no sympathy for) the position of the state as nominal owner of the land. On Trust farms perceptions are somewhat different, as there is generally a greater awareness of the state as nominal owner even though many people believe that the permission given to the community by the slate to occupy the land is tantamount to a transfer of ownership. The small number of farms bought outright by tribal groups or others constitute a separate category of full (individual or collective) private ownership, with no state involvement, but in practice, popular perceptions do not differ greatly between the three categories. In a study of the communal land system in the Transkei, Solinjani (1986) found that "most informants believed that the land belonged to the tribe and that the chief ... had authority over all the people and all the land", a good example of the conflation of property rights and socio-political jurisdiction that characterises much of the debate around land and traditional leaders in South Africa. The minority who differed from this view, according to Solinjani, believed that the land belonged to the government and was under the control of the Magistrates, as it was they who issued Permission to Occupy (PTO) certificates. 4. Communal Tenure The communal tenure system found in South Africa is 'communal' in the sense that an individual's entitlement to land flows from membership of a socio-political community (a village or tribe), rather than from private ownership (Bennett 1995: 168). Communal tenure, in the South African context, does not imply communal (or collective) agricultural production, even on shared resources such as communal grazing land. Nor does it imply that all decisions regarding the allocation of land are made by the entire community. Examples of collective production may be found, such as exchanges of labour or ploughing cattle, or joint production of vegetables or poultry on small projects, but these are the exception rather that the rule and are largely independent of the land tenure system. Land for arable and residential purposes is usually obtained through the tribal chief or, more commonly, the village headman acting on behalf of the chief, who may allocate plots from whatever land is currently available. Under customary law, the right to land usually applies only to male 'household heads', but in practice is sometimes extended to women (Bennett 1995: 170). For Solinjani, referring to the Transkei, "Anyone who was married and was a permanent resident of any of the areas under (the chief's) jurisdiction was qualified to apply for land. In addition, all unmarried females who had children could also apply for land if they were permanently resident". This combination of requirements - membership of the community and head of a household - together with some discrimination between men and women - recurs throughout the homeland areas, with minor local variations. Once allocated, residential and arable plots are reserved for the exclusive use of the occupying household. Unallocated lands are generally available to community members as a common pool resource (commonage), providing pasture for livestock and other natural resources such as timber, thatching grass, edible fruits and plants and materials for use in traditional medicine (Cousins 1996:168). Those who obtain land receive a right to the use and benefits of that land, but have no right to sell it and can only transfer it to another family member with permission of the tribal leaders. Chiefs and Tribal Authorities have, in principle, the power to repossess land if it is abandoned, if it is needed for another purpose such as a road or a public building, if it is deemed surplus to the needs of the holders or in order to punish a landholder for some offence. Examples of such repossession are rare, however, and the communal system is generally seen as a reasonably secure form of tenure (Bromberger 1988: 208). In addition to the Land Acts, the principal legal instruments governing access and use of South African Development Trust Land were (using their original title.s): • the Control of Irrigation Schemes in Bantu Areas (Prockimation R 5 o f 1963) which consolidated regulations controlling irrigation schemes in Black Areas; the Trust Forest Regulations {Proclamation R 191 o f 1967) which regulated the use of forest land in Black Areas; and * the Bantu Areas Land Regulations (Proclamation 188 o f 1969) which regulated the tenure in all South African Development Trust areas. All of these were proclaimed under the authority of the Native Administration Act, 1927 (Act No. 38 of 1927) and the Native Trust and Land Act, 1936 (Act No, 18 of 1936). Prior to the legislative reforms and the virtual collapse of the homeland administrations in the early 1990s, occupants of communal land could register their residential and arable holdings with the local Tribal Authority and Magistrate's offices, where they would officially be granted 'Permission to Occupy' (PTO), either verbally or in writing. This system was not observed in all cases and, since the abolition of all 'racially based' land laws in 1991, the legal situation is far from clear (Cross and Putsch 1995: 23; Westaway 1995:11). 5. The Arabie-Olifants Scheme The Arabie-Olifants irrigation scheme is located on the right (southern) bank of the Olifants River, at approximately 24° 30' S and 29“ 30' E. The overall scheme is loosely structured, in that it does not share a common infrastructure or central management. Water is drawn directly from the Olifants River, from the Arabie Dam (upstream of the scheme) and from tributaries of the Olifants River. The irrigated land, which comprise approximately 2,223 hectares in total, is located on 14 farms'. There are an estimated 1,650 plots, ranging in size from less than one hectare to 5 hectares. Flood irrigation, furrow irrigation, sprinklers and central pivot irrigation are all in use on the scheme. Irrigation management and farmer support services are provided by state and parastatal agencies but, since 1996, have been in severe decline, leading many plot-holders to suspend or abandon agriculniral activities.^ The land that comprises the Arabie-Olifants scheme is highly suited to irrigated agriculture, and has been used for such since at least the early twentieth century. Official land registry records show that this land was first surveyed and allocated to private owners by the government of the Transvaal during the period 1871 to 1873. This was a time of steady encroachment by white settlers onto the land of die Pedi people under their chief, Sekhukhune, prior to their defeat and subjugation by British-led forces in 1879. The earliest deeds (referred to as an Original Deed of Grant) issued on the farms in question were to the farms Gataan and Vogelstruiskopje, issued on the 4' ̂March, 1871, and the last was to the fann De Paarl, issued on 22'“’ December, 1873. An examination of the original documents in the Pretoria Deeds Office revealed that all the farms were granted to separate individuals (as opposed to companies), all with Afrikaner or English surnames. This is in line widi the pattern of conquest and expropriation throughout southern Africa during the eighteen and nineteenth centuries, whereby the settler-colonialists surveyed The term 'farm', in the South African context, has a meaning which goes beyond the normal usage in the agricultural literature. It refers to an original land grant, which has been surveyed and recorded in the deeds registry and topo-cadastral maps. Thus, in the case of the Arabie-Olifants scheme, it indicates a property (or administrative) unit rather than an agricultural enterprise. ^See Mpahlele, Malakalaka and Hedden-Dunkhorst (1999) for detailed treatment of these topics. and granted property rights to white people without reference to the pre-existing rights (and occupancy) of African peoples. Although no firm evidence could be found as part of this study, the likelihood is that these farms were acquired by their new white owners complete with resident populations of Africans, who would have been coerced into providing labour to the emerging land-owning class (see Delius, 1983: 129). Over the sixty or seventy years that followed the original deed of grant, all of these farms changed hands at least once, and some as much as five limes, as a result of inheritance, sale and bankruptcy. The farm Strydkraal, for example, was originally granted to a J.B. Brown in 1872, who sold it to the Oriental (Transvaal) Land & Exploration Company, Ltd, in 1885. It was subsequently acquired, by order of court, by United African Lands Ltd, in 1910. This company then separated the mineral rights, which it retained for itself, from the surface rights, which it sold to O.T. van Niewkerk in 1920. This pattern, whereby farms were acquired by mineral speculation companies, who separated the mineral rights from the surface rights, was repeated on twelve out of the fourteen farms that make up the Arabie-Olifants scheme. By the mid-1930s, all the farms (minus the mineral rights) were back in private hands. Entries on the title deeds from this period indicate that the land-owners, particularly those on the upper reaches, were paying water levies to the Middleburg Irrigation Board, which suggests that irrigation was already established on these farms. In 1936, after more than twenty years of deliberation, the South African government passed the Native Trust and Land Act, which made provision for the purchase of land for the purposes of extending the so-called Native Reserves. Over the twenty five years that followed, the South African Native Trust purchased the fourteen farms that today make up the Arabie-Olifants scheme, starting with the farm Mooiplaats, acquired in 1938, and ending with the farm Goedverwacht in 1963. No information could be found regarding the existence of African residents on these farms at the time of purchase by the Trust, but it is likely that at least the former farm workers and/or labour tenants were still there. By the 1950s and 1960s, however, when the current residents took up occupation, many of these farms were empty, which suggests a pattern of forced evictions by the Trust. This hypothesis is supported by a land claim which has been lodged with the Commission for Restiniiion of Land Rights (under the 1994 Restitution of I-and Rights Act) by the Phetwane Community on the farm Hindostan. These people, part of the Ga Phahia community and currently residing in the Leeuwfontein township adjacent to Marble Hall, claim that they were forcibly removed by the Native Trust in 1958 for reasons of ‘ethnicity’, as the farm was intended for the Matlala people. Three of the communities currently inhabiting the Arabie-Olifants farms have been relocated from outside the immediate area. This was part of the processes of ‘homeland consolidation’ and social engineering that were integral to the policy of Grand Apartheid. The Masha community was moved from Kalkfontein, in what is now Mpumalanga, in the 1950s. The Mampana community was previously scattered on various white farms and parts of Sekekhuneland and was brought together under its own chief on the farm Krokodilheuvel in 1962. The Matlala community, which occupies the farms Hindostan and Coetzeesdraai was moved from Jane Purse, where it shared land with other tribes, in the 1960s. The Masemola community originated in the areas of the present-day scheme and, although its members were for long scattered under various white farmers, they were eventually consolidated onto the nine farms at the centre of the scheme. When the communities now occupying the irrigation scheme were first granted occupation, the land had the status of ‘Trust Farms’, in that it was owned by the South African Native Trust (a state body). Farms were allocated for use purposes to four specific communities (’Tribes’), each under a chief. Within a chieftaincy, the population was divided into villages, and each farm was for the sole use of one, or occasionally two, villages. Matlala community was granted two farms on the scheme (as well as other land), Mampana was granted one farm on the scheme, Masemola was granted nine farms on the scheme. D and Masha community was granted two (and no odier land). The irrigation system itself took many years to develop and not all of the farms were irrigated when these communities first arrived. The legal status of these farms has changed somewhat since the Trust first acquired them. From about 1986, South African government policy was to transfer the ownership of land in the homelands to the homeland governments and, in some cases, to the occupying communities themselves. Thus, we find tliat the farms Hindostan, Coetzeesdraai, Krokodilheuvel, De Paarl, Gataan, Goedverwacht, Vlakplaats, Strydkraal and Mooiplaats were transferred to the Government of Lebowa in 1993. The farms Veeplaats and Nooitgezien remain registered in the name of the SADT and two farms, Haakdorinsdraai and Vogelstruiskopje, were acquired by the Masemola Tribe in 1994. It is unlikely that these changes were intended to affect the way in which land was held by individuals (or households) within those communities, as land-holders continued to be subject to the rules of the wider community as before. With the exception of farms transferred to particular tribes, all the land on the scheme can be classified as state land which has been allocated for use purposes to specific communities, with formal title retained by the Minister of Land Affairs. Current government policy recognises and protects the rights of such long-established communities occupying state land, and the Department of Land Affairs was, until recently, in the process of preparing legislation that would allow for the transfer of ownership to democratically constituted community groups, although this process would now appear to be on hold. 6. Official Land Allocation and Tenure System at Arabie-Olifants Plot-holders and officials on the Arabie-Olifants scheme were found to be in broad agreement as to how the plots on the irrigation scheme were originally allocated. The design of the scheme - and hence decisions about the size and number of irrigation plots - would appear to have been carried out by officials of the SANT without consultation with the would-be beneficiaries. Most of the farms were intended to be used for both agriculmral and residential purposes and so were divided into distinct zones - irrigated lands (sub-divided into individual plots), residential areas (also subdivided into stands) and communal grazing for the use of those with livestock. Given the limited areas allocated for grazing, it would appear likely that restrictions were placed on the numbers of livestock that could be brought onto the farms. Residential stands are generally described on the PTO certificates as being one-quarter hectare in extent. The irrigated plots demarcated during the 1950s and 1960s were generally 1.28 hectares in extent, equivalent to 1.5 morgen, a standard size for plots on SANT irrigation schemes throughout the country at that time. In later years, there was a trend towards larger plot sizes. The precise sizes of plots recorded on the PTO certificates, however, show some variation. On the farms De Paarl and Gataan,, for example, where plot sizes were all described by officials as being 1.28 hectares, the sizes shown on the PTOs ranged from 0.83 ha to 1,28, while on Haakdorinsdraai tliey varied from 0,75 ha to 1.25 ha. At Krokodilheuvel plot-holders reported that, as they arrived on the farm, they were allocated housing sites and irrigable plots by the Native Commissioner, assisted by the chief Every household was allocated one residential site and one irrigated plot each, and it would appear that no household was let) without a plot. This neat match of households to plots would suggest some degree of pre-selection of beneficiaries at the time the community was preparing for relocation, but no information could be obtained on this question. It was suggested on this and otlier farms that the numbered plots were linked to tile numbered residential sites - in other words, the plot a particular household was granted depended on where they were to live, and no deviation from this pattern was allowed. All respondents emphasised that plots were, in the first instance, allocated, to ‘households’ and registered in the name of the household head, usually a man. It is not clear whether any female-headed households were amongst tliose which received land at this time. It could not be established what form the original registration of plots took, but from 1969 onwards, all lands were held on the basis of Permission to Occupy (PTO) certificates, issued under the Bantu Areas Land Regulations (Proclamation RI88 o f 1969), which drew its legal authority from the 1936 Native Trust and I,and Act. This Proclamation does not appear to have greatly changed the system of land administration or tenure, but codified existing practice. Proclamation R188 gave the Bantu Commissioner authority to grant PTOs on arable and residential land in the areas falling under his authority. The Commissioner could, if land were available and after consultation with the tribal or community authority, allocate land to the following categories of people: * A black person who was the lawful but unregistered occupier of any Trust land for arable or residential purposes; ■ The male head of a black household, or the widowed female head of a black household, who would not become the holder of more than one piece of land; * Any church or missionary society for occupation by a black minister, preacher or evangelist in its employ. Land allocations other than for arable and residential purposes had to be approved by the Secretary for Bantu Administration and Development. The extent of land which could be allotted under this system could not exceed half a morgen for a residential plot and four morgen for an arable plot. The holder of such a right was required to pay a nominal rent to the Bantu Authorities Trust Fund. Permission to vacate a plot for a period exceeding one year, but not exceeding two, had to be obtained in writing from the Commissioner. Upon the death of a registered PTO holder, the residential or arable allotment reverted back to the eommonage for re-allocation. Any attempt by the registered holder to leave a testament passing his holding to a family member had no force in law and any such testament was deemed void. The wife of a deceased occupier could remain on the land until her death, but the land remained registered in the name of the former male head of the household. Proclamation R188 further gave the Minister of Bantu Administration and Development the power to terminate any or all rights in any arable or residential allotment “whenever in his discretion he deem it expedient in the interest of the State or the Trust or in the general interest to do so” [Section 58). The Commissioner had to give notice to the holder of the PTO whose rights were being terminated and there was no system of appeal from such an order. The person whose rights were cancelled under this section had, however, a right “to be allotted other land in an area set aside for residential purposes or elsewhere in a Bantu area ...and in addition be paid such compensation in money out of Trust funds as the Bantu Affairs Commissioner shall assess.” The PTO certificates held by plot-holders at the Arabie-Olifants scheme combine an allocation of irrigated land with a residential stand and, in a minority of cases, permission to graze a set number of cattle (typically five or six). Only the measure of land to be allocated, not the precise location, is recorded, along with the farm name. The residential stand number and the corresponding irrigated plot number are, however, written on the top right-hand corner of the certificates. These PTO certificates are today the main visible claim to the land, even though they are technically obsolete since the repeal of the 1936 Act in 1991 (Abolition of Racially Based Land Measures Act, 1991). The question of payment for land could not be resolved with any certainty. A number of informants stated that annual fees must be paid by land-holders, either to the Tribal Authority or the Department of Agriculture. Amounts of R5 per hectare and R12 per plot were mentioned. In addition, offieials of the Department of Agriculture reported that plot-holders at Veeplaats and Goedverwacht pay an annual fee of R50 to the Department. It was not possible to determine how effectively these fees are collected or the consequences, if any, of non-payment. Thus, the formal tenurial arrangements on the Arabie-Olifants scheme at the time the current residents took up occupation can be summarised as follows: * Land was state-owned and allocated to Tribal Communities, divided into a number of villages; * Household heads (men and their widows) were allocated residential and irrigated land, plus a share of communal grazing; * Allocation was made by government officials, who communicated with the plot-holders through their chiefs; * All land was granted under permit (PTO), in usufruct, and could, in theory, be withdrawn by the state for any reason, although in practice such allocations were for life; * Inheritance of land within families was not sanctioned by the law, but was tolerated in practice; ‘ Land could not be sold, leased-out or otherwise alienated. 0 7. Popular Perceptions of Land Allocation and Tenure at Arabie-Olifants According to popular perception of the land allocation process, people in need of land, for residential or arable purposes, must first approach their village headman who may, at his discretion, refer the application to the Tribal Authority (also referred to as the Tribal Council) which may approve, amend or refuse the application. The Tribal Authority is under the jurisdiction of a tribal chief, who is not usually involved in its day-lo-day proceedings. According to one chief at the Arabie-Olifants scheme, anyone applying for land may appeal a decision of the Tribal Authority to the chief, and a chief has the power to overturn any decision of the Tribal Autliority. Other chiefs suggested a more consensual model, whereby the chief and the Tribal Authority reach decisions Jointly, in practice, it is extremely rare for a Tribal Authority and a chief to be in open disagreement. Once an application has been approved by the chief and Tribal Authority, and a specific piece of land has been allocated, tlie applicant is given a letter from the chief to lake to the local Magistrate’s office, where an official Permission to Occupy (PTO) Certificate will be issued. In the case of rainfed arable land (i.e. land off the official irrigation scheme) and residential sites, the portion will usually be surveyed by the Deparunent of Agriculture prior to the PTO certificate being issued. There was no suggestion that a Magistrate had the power to refuse a properly-presented request from a chief to issue a PTO certificate. Virtually all respondents (chiefs and commoners) agreed that the ultimate power to allocate land rests with the chief, although many qualified this with statements such as ‘the ehief and the Tribal Authority’ or ‘the chief, on behalf of the community’. Chiefs were generally seen as bound by the need to consult the community on matters of importance and to protect the interests of all community members. There was no suggestion that chiefs were subject to democratic norms, but there was a strong sense that a chief was not free to act in an authoritarian manner or solely in his or her own interest. The chiefs were widely seen as having a responsibility to provide land for their subjects, and both chiefs and commoners expressed frustration that there was not always sufficient land for all those who wanted it. Thus, the theoretical (or traditional) right of every household to an arable plot must be understood in die context of land scarcity, betterment-era zoning and predetermined plot sizes on the formal irrigation schemes. all of which place limits on the power of the chiefs to allocate land and the rights of subjects to acquire it. Some minor disagreement was found around the question of who is entitled to apply for land, but a broad consensus could be discerned. Those with the strongest entitlement to land were male heads of households - that is, married men with their own homesteads. Such men were entitled to apply for land on their own behalf. All other social groups were seen as having lesser degrees of entitlement, the extent of which varied somewhat between respondents. Married women were considered by many to be entitled to apply for land, but others felt they were obliged to do so through their husband or other male relative. Unmarried women or widows with children and their own homestead were also seen as entitled to apply for land, but most respondents - including many women - believed that they should do so through a male relative. A number of respondents mentioned minimum ages at which such women became eligible, the highest, from Masemola Tribal Authority, being 50 years of age. No such age limits were mentioned with regard to married men. Unmarried people, without children or homesteads of their own, were generally considered ineligible to apply for land. A number of respondents, including one chief, suggested that exceptions could be made for unemployed young men who wanted to practice agriculture, but no such exceptions could be made for young women. The chief in question reported that if such a young man approached him, he would not grant him land of his own, but would persuade some other person to share land with him until he became entitled to land of his own. Thus, family status, gender and age all contribute to a widely- agreed hierarchy of entitlement to land, with married, older men at the top and unmarried, younger women at the bottom. Most of the households that obtained plots at the beginning of the scheme would appear to have held on to them to this day. Substantial numbers of households are today without access to land, and this can probably be attributed to an expansion in the number of households over time, whether due to natural increase or inward migration. With the exception of one area (Masha), no evidence could be found of households losing rights to land, although there were reports of households occasionally surrendering their plots. Thus, despite the elaborate rules governing allocation, there would in fact appear to be very little turnover in plot-holding on the scheme. This is supported by an examination of PTO records held in the ARDC office at Veeplaats. For example, on the farm De Paarl, with a total of 50 plots, just three changes are recorded for the period since 1983; two in 1992, both to people with different surnames to the original plot-holders, and one in 1998 recorded as a transfer to ‘the only widow’ of a deceased plot- holder. On Veeplaats itself, with 121 plots, in the period since 1987 four plots were transferred to people with different surnames to the original plot-holders, five more from men to what appear to be their widows and two from women to their daughters.’ As irrigated blocks are allocated to specific communities and every household has, in theory, a right to a plot, the main limitation on access to land is the number of irrigated plots available. The only other (hypothetical) reason cited, by a chief, for refusing land to a qualified applicant, other than land scarcity, was ‘disloyalty to the chief, which was explained as the refusal to accept the authority of the chief. In practice, however, no such case could be recalled by any of the informants. As noted above, irrigated plots generally remain within the possession of a single household indefinitely. On the death of a plot-holder, possession usually passes to a close relative, either a wife, son, daughter or daughter-in-law. No mention was made of land passing from a wife to a husband. No single pattern of inheritance could be identified, and it appeared to be influenced by who within the household was ’These judgements are made either on the basis of notes in the official record or by comparing ages, names and surnames of the various parties. actually using the land and the composition of the household at the time of transfer. According to one chief, the order of preference for inheritance from a male plot-holder was first his senior wife, followed by any junior wife, then a son and finally a daughter. A woman would be expected to leave her plot to her son, if one lived with her, or else to her daughter or daughter-in-law. Women tended to inherit land if the men in the households were not interested in using it themselves, or if they (the women) had already been using it. Primogeniture does not apply, but the eldest offspring still resident within the parental home would appear to have a strong claim on household land. In practice, land often passes to the last-bom, all other siblings having left to form their own households. Identifying the acmal ‘plot-holder’ within a household was not entirely straightforward, for a number of reasons. First, official records (e.g. PTOs) are not always updated following the death of a registered plot-holder and, over lime, there would appear to be a growing discrepancy between the written record and actual practice. Second, the registered owner, if alive, may not be the person actually using the plot. In particular, there would appear to be a tendency for plots to be registered in the name of men when they are actually used by women. It is impossible to say without further research whether this indicates a change in practice within households over lime, with land use transferring from men to women, or an inherent bias within the land registration system, whereby land is registered in the names of men even though it is to be used by women. Overall, the majority of plots on the Arable/Olifams scheme are registered in the narnes of men, although the majority of users are undoubtedly women. Formal transfers of land between households would appear to be extremely rare, although most respondents did hold open the possibility that a household could surrender its land (for whatever reason) and that it could be reallocated by the chief. Informal, temporary exchanges of land, however, would appear to be much more common, although informants were generally unwilling to discuss the matter in any detail. While many respondents claimed it could never happen, others admitted it was a common practice but refused to divulge any details. According to government officials on the scheme, the most common form of transaction is where someone requests the use of a plot from a neighbour for a single cropping season, and in return ploughs the land on behalf of the plot-holder the next season. Other transactions mentioned were temporary share-cropping arrangements, whereby a portion of the crop is handed over to the official plot-holder after harvest. Cash payment for land (i.e. rental) was said to be rare, but possible on occasions - a charge of RSO for a 1.28 hectare plot for one season was mentioned by one extension officer. Overall, however, respondents were agreed that the details of any arrangement to share land were a private matter between the two parties concerned, and that it was best not to publicise the fact that one was sharing ones land with others. The precise reasons for this reticence were not clear, but there was a general feeling that land sharing arrangements would not meet with the approval of the powers-that-be. There is a common perception among those on communal land that, because land is allocated on the basis of need (at least in theory), continued security of tenure requires that a plot-holder is seen to be using his or her plot. Indeed, many tribal communities in adjoining areas have rules that set time limits on how long a piece of land can remain unused (typically about three years) before it reverts to the community for reallocation. Within this context, long-term leasing-out of land could be interpreted as a lack of need for land, and could therefore weaken the rights of the official land holder. The leasing-in of land could also be perceived as an attempt to bypass the established channels for the acquisition of land and thus reflect badly on the lessee or tenant. The authorities involved in the provision of agricultural services, namely ARDC and the Department of Agriculture, also expressed some antipathy to land sharing, and plot-holders were under the impression that it was still forbidden. The reasons for this would appear, in part, to be simply bureaucratic, in that ARDC administrators encounter difficulties in cases where the name of the official plot-holder does not correspond with the name of the person being charged for services. Beyond this, however, it would appear that scheme officials in general have, over many years, discouraged land sharing as part of their mission to promote ‘good farming practice’ which, in the South African case, invariably carries a strong bias towards a system of ‘owner-occupiers’ and the concept of 'one-man-one- plot’. Informants at Arabie-Olifanis supplied a range of responses to the question "Who owns the land?” . The most common response was ‘the plot-holders’ or ‘the farmers themselves', an opinion shared by plot- holders and at least one chief. Other common response were that the land belonged to the ‘the Chief, ‘the community’ or ‘the tribe’, or some combination of these. One chief stated that the land belonged to ‘the community living under the Chief, emphasising the indivisibility of the two. A minority of informants mentioned ‘the government’ or ‘the Trust’ as having a stake in the land, but usually qualified this by reference to the moral right of the inhabitants as ‘the real owners’. The sense of ownership expressed by most informants can be linked to the strong sense of tenure security felt by plot-holders. A recent smdy by Mpahlele, Malakalaka and Hedden-Dunkhorst (1999) found that the great majority of plot-holders at Arabie-Olifants felt that their tenure was ‘secure’ or ‘very secure’. Objectively speaking, land holders would indeed appear to be secure, given the long occupancy of most plot-holders, the low turn-over of plots, the freedom to bequeath land to one’s chosen heir and the absence of reports of people losing rights to land (with the possible exception of Masha area, see below). Many plot-holders were conscious, however, of at least some of the differences between the communal tenure system under which they live and the free-hold system prevailing in the ‘white’ areas. This gave rise to a number of comments along the lines of ‘we feel insecure because we don’t have title deeds’, although most informants expressed their satisfaction with the PTO system. On further examination, all of those advocating title deeds in fact supported the continuation of the communal system, in that they did not believe that people should have to pay for land, or that people should be free to sell land to others. Rather, they wanted the same legal status as ‘white’ land-owners, both as a bulwark against arbitrary interference by the state (an experience common to all the communities at Arabie-Olifants) and as a means of accessing credit from commercial lenders. The apparent contradiction between land as collateral and as a non-tradable commodity was not raised. Overall, it would appear that plot-holders and community leaders at Arabie-Olifants wish to maximise the tenure security and access to commercial services of individual plot-holders, while rejecting the concept of a free market in land. Although not everyone was supportive of the current tribal-based system, there was strong support for some degree of ‘communal’ control over the allocation of land and for restricting access to land to members of the immediate community. 8. Changes AITecting Land-Holding Since the Beginning of the Scheme The land tenure system described above remains more or less intact today, albeit with some modifications. The main changes over the twenty-five years to 1997 (prior to the dramatic decline of the scheme) have been an increase in average plot size, growing technical sophistication of the irrigation system and what would appear to be a substantial rise in the proportion of plots held by women.* The rale of change in land-holding has, however, been limited by the lack of formal mechanisms for reallocating land (i.e. to people outside the households of the original plot-holders), and by the fact that reallocation can only be made within the communities (villages) located on particular farms. Where change has occurred, it has generally been linked to major changes in the design of the irrigation scheme itself Thus, apart from the original allocation of land in the 1950s and 1960s, allocation on the Arabie- Olifants scheme has tended to be a somewhat sporadic affair, and the information and opinions gathered on the allocation process needs to be seen in this light. While many informants were able to supply ‘The absence of comparative data for the period prior to 1983 meant it was impossible to quantify this change with any certainly. 0 0 detailed descriptions of the (theoretical) process whereby people applied for and were allocated plots, few were in a position to provide instances where this process had actually been applied. The Arabie-Olifants irrigation scheme has been extended, and the infrastructure upgraded, in a number of phases. The first phase of the scheme can be considered to be what existed on the scheme when the communities now present first took up occupation, in the late 1950s and 1960s. This consisted of a gravity-fed canal system, utilising furrow and flood irrigation. A standard plot size of 1.28 hectares was the norm and plots were allocated to ‘heads-of-households’ who were largely, if not entirely, male. In the years that followed, the profile of plot-holders would appear to have become more female, as plots were transferred to women within the existing households and some new plots were allocated directly to women. A major expansion of the scheme was implemented by the Lebowa government in the early 1980s. The farms Veeplaats, Goedverwacht and Nooitgezien which had previously been used for grazing were put under irrigation for the first time, using sprinklers. Sprinklers were also installed on farms such as Strydkraal and Wonderboom which had previously been using furrow or flood irrigation but others, such as Hindustan and Coetzeesdraai, continued to use flood irrigation. A number of centre pivot irrigation systems were also installed on the farms Goedverwacht, Veeplaats and Strydkraal. Further upgrading of the technology occurred in 1987, when pumps were installed at Nooitgezien to pump water from the Olifants river and at Goedverwacht to pump water from the Ngwaritsi river (a tributary of the Olifants, fed by Piet Gouws dam). The significance of these developments, from a land tenure perspective, is that they brought a substantial increase in average plot sizes, as well as in the total number of plots, and necessitated the reorganisation of plot boundaries. This in turn created an opportunity for the reallocation of plots among members of the affected communities. The biggest group of farms affected by these changes were those under Chief Masemola. On the newly created irrigation scheme at Veeplaats, the standard plot size was 2.5 hectares, virtually double the previous norm. Access to irrigated land for members of the Masemola community was greatly increased with the addition of 121 new plots on Veeplaats alone. In addition, thirty 5- ha plots were created under the three cenbe pivots on the farm Goedverwacht to accommodate an emerging ‘elite’ of more commercially-orientated farmers. At least two of these plot holders, both men, have since been able to acquire additional plots on the centre-pivot schemes, giving them holdings of ten hectares each. Five out of the current twenty-eight plots holders are women, each with a 5-ha plot. On farms such as Gataan and Wonderboom, where plots were redesigned to suit the sprinkler system, informants were of the opinion that virnially all of the original plot-holders were reallocated plots, and some further households received plots on the additional land brought under irrigation. The only households to lose land, according to the Tribal Authority, were those that were no longer interested in farming, although this could not be independently verified. Further downstream, at Strydkraal and Mooiplaals, the reorganisation of holdings would appear to have led to the dispossession of some plot-holders, particularly members of the neighbouring Nchabeleng community. This matter is now the subject of legal proceedings so it is not appropriate to draw any conclusions at this stage. On the basis of a number of different accounts, however, it can be said that the area falling under Chief Masha appears to be the only one where members of two different tribal communities - Masha and Nchabeleng - shared the same farms and that at some point in the 1980s the members of the Nchabeleng community withdrew or were excluded from Masha’s area. Apart from the increase in average plot sizes, and the increased supply of irrigated land, the biggest change in land-holding patterns arising from these developments was a reported increase in the proportion of plots registered in the names of women. According to the Masemola Tribal Authority, a deliberate effort was made by the then-chief to allocate plots to women in cases where they were to be the principle cultivators. This was intended to strengthen women’s control over plots their plots and to reduce conflict between husbands and wives over the fruits of the women’s labour. Recent studies conduced at Veeplaats suggest that 59% of plots are de facto controlled by women (Mpahlele, Malakalaka and Hedden-Dunkhorst, 1999), although an examination of the available PTO records would suggest that the proportion actually registered in women’s names is closer to 40%.’ On other farms, the estimated proportion of plots registered in women’s names ranged from 20% at Gataan to 26% at Wonderboom. Another development that has favoured the acquisition of land by women has been the rise of so-called garden plots adjacent to the main irrigation scheme. Five of these were identified on the farms in question: two at Wonderboom, started in 1988 with support from the NGO Operation Hunger; one at Vlakplaats, started in 1993 with support from the Department of Health and Welfare; and two at Nooitgezien. Another, the largest by far, is the Sepitsi scheme, on the farm Badfontein, across the Olifants river from Veeplaats. This was a Joint initiative by the local community, ARDC and the local Department of Agriculture, started in 1996 on land falling under Chief Mphahlele. All of these holdings are extremely small. At Wonderboom, one plot of 0.7 hectares is shared by 44 people (equivalent to 0.016 hectares per person) and another of 2 ha is shared by 50 people (0.04 ha per person). Both of these schemes are largely composed of women members (43 out of 44 and 47 out of 50, respectively). On the Mamakau scheme, on Vlakplaats, 17 women share approximately 1.75 hectares, equivalent to 0.1 ha each. At Sepitsi, where a total of 81 plot-holders have individual holdings of0.12 ha each, Mpahlele, Malakalaka and Hedden-Dunkhorst (1999) estimated that 70% of plot-holders were female. The formal land tenure situation on these garden schemes differs from tliat on the larger irrigation plots of the Arabie-Olifants scheme. The land was formally allocated by the Chief, and surveyed by the Department of Agriculture, and a single PTO was issued to each group. Such schemes are a recent development, and it is not clear what degree of tenure security they offer their members. One further change to the land-holding pattern on the Arabie-Olifants scheme has been the rise of ‘informal’ (or ‘private’) irrigation plots adjacent to the river and the main irrigation canal. A handful of such plot-holders have PTOs granted by the respective chiefs, but most would appear to be unapproved. This was reported to be a recent phenomenon which informants linked to the decline of production on the formal scheme and the break-down of local authority structures during the political turmoil after 1990. In a few cases, water is being pumped from the river, but most plot-holders rely on unauthorised (‘illegal’) connections to the main canal. The rise of ‘informal’ irrigation could be seen as evidence of a demand for land, but this must be set against the collapse of production on much of the formal scheme. 9. Land Tenure Reform: Now And In The Future The evidence from the Arabie-Olifants scheme would suggest that there is a high degree of satisfaction with the existing tenure system among plot-holders and community representatives (elected councillors and tribal leaders) and little demand for radical change to the system by which people acquire land, the conditions under which people hold land or the (largely unwritten) rules governing the transfer of land. ’p r o Certificates were found for eight farms, with some certificates missing. The Certificates indicate the name but not the sex of the registered plot-holder. Estimates of the sexual division of land are based on an analysis of first names but remain imprecise due to the prevalence of non-specific first names. Pressure for tenure reform is coming mainly from two external sources, arising from the unilateral withdrawal of government services to the scheme. The first of these sources is the government itself, or more accurately the Northern Province Department of Agriculture, l,and and Environment, in its role as custodian of the irrigation infrastructure and as the parent body of the Agricultural and Rural Development Corporation (ARDC). Paced with the imminent collapse of the ARDC, the Department is anxious to transfer responsibility for the upkeep (and possibly the ownership) of the scheme to the plot- holders themselves. Before this can happen, it will be necessary to clarify the ownership not only of the land but also of the infrastructure and other assets that make up the scheme. Transferring ownership rights in land from the state to the plot-holders (probably as a collective) has, until recently, been seen by the Department as a necessary step towards the transfer of the entire scheme. The second factor propelling the tenure debate is the new-found need for plot-holders to access production credit from commercial sources. Again, this stems from the withdrawal of credit facilities by the ARDC over the last two years and the failure of the authorities to make provision for the continuation of this and other services by alternative means. Commercial banks in South Africa generally will not lend to farmers who cannot offer title deeds as collateral. The state-owned Land Bank, which includes the provision of credit to ‘emerging’ farmers in its mission, has also been slow to lend to small-scale farmers on communal land. A transition from communal tenure to individual free-hold is being promoted by many commentators as the solution to this problem, and the idea has been taken up by some small-scale farmers. What is not being addressed, however, is the credit-worthiness of farmers currently holding land under communal tenure. It can be assumed that commercial lenders will take into consideration not only collateral but also issues such as financial history and ability to repay in deciding whether or not to extend credit to small-scale farmers. Thus, a transition to ffee-hold is unlikely to offer the panacea that many would suggest. Evidence from the Arabie-Olifants scheme suggests that what the majority of plot-holders want is permanent and secure tenure, and access to credit and other services on reasonable terms. The preference among all the plot-holders and community leaders interviewed for this study was to achieve this within the current system of communal tenure, or with a slightly modified communal system. The elements of the current tenure system that would appear to have near-unanimous support are: • that the community (through its leaders) should have the power to allocate land; • that land should neither be bought nor sold; • that access to land should be restricted to members of the immediate community; that plot-holders should have secure, life-long tenure and the freedom to bequeath their land to members of their family; * that the community should have the right to impose sanctions, ranging from verbal warnings to fines and repossession, on land-holders who abuse their position, either by failing to use their land or by Interfering with other users. This strongly suggests that, while plot-holders are anxious to find solutions to their most pressing problems, most do not support, or have not contemplated, a transition to free hold tenure or a market- based private-property system .This clearly presents problems for those advocating the use of land as collateral and points to the need to consider alternative forms of credit for farmers in the communal areas. Within the communal tenure system there is certainly room for reform, without necessarily compromising the features of the current system that community members find most desirable. In the area of land allocation, it was clear that elements within the community were not happy with the power of the tribal leaders and the way it is exercised. Discussions with the elected local councils (TLCs) revealed considerable dissatisfaction, especially among the youth and the landless, many of whom would appear to feel excluded by die current tribal leadership. This suggests the need for a more democratic or inclusive process whereby different elements witliin the community, not only those on good terms with the chiefs, could be involved in decisions regarding communal land. It also suggests the need to relax some of the restrictions on the right to apply for land, especially for women and young people. Land registration is another area in need of urgent reform. The issuing of PTOs offtcially ceased with the abolition of the governing legislation in 1991. Nonetheless, some chiefs and Magistrates have continued to issue PTO certificates since then but this has been without legal authority. Pending the introduction of new legislation, there is currently no legal mechanism for registering new plots on communal land, or transferring land to new plot-holders. Although the legal position is far from clear, it would appear that existing PTO certificates (those issued in 1991 or before) still carry some legal weight. The national government has declared its intention of abandoning the system of permits in favour of one that recognises the underlying rights of land holders on communal land but has yet to legislate in this regard. The South African Land Reform Programme is commonly described as having three legs: Restitution, for those who can prove they were unfairly deprived of their land rights since 1913; Redistribution, which assists communities to acquire land through the open market; and Tenure Reform, which aims to provide tenure security to those who were denied it under the previous regime. Of these, tenure reform has received the least attention to date and has yet to bring significant change to the millions of people living under communal or informal tenure systems. The legal basis for tenure reform (as applicable to people living and working on communal land) is composed of the following Acts. The Upgrading o f Land Tenure Rights Act, 112 of 1991, introduced by the former regime, allows for the ‘upgrading’ of PTOs to title deeds (or Deeds of Grant). Amendments by the ANC-led government since 1994 mean that, in practice, this now applies only to residential or business sites in urban areas and is not applicable to agricultural land in communal areas. The Interim Protection o f Informal Land Rights Act, 31 o f 1996, (IPILRA), is intended to protect people with insecure tenure from losing their rights in land, pending the introduction of long­ term tenure reform measures (i.e. the Land Rights Bill). This Act gives legal recognition to all long-term occupiers of communal land, regardless of the type of permit or permission they hold. It does not provide for any change in tenure, but does set down procedures that must be followed by anyone wishing to develop or dispose of communal land or otherwise affect the status of established occupiers. • The Communal Property Associations Act, 28 o f 1996, establishes a new form of legal body (a Communal Property Association, or CPA) through which people may collectively acquire, hold and manage property in terms of a written constitution. To date, the Act has mainly been used by communities acquiring ‘new’ land under the Restitution or Redistribution programmes, ratlier than as a means of changing the status of land already under occupation. Communal Property Associations generally acquire land in freehold and allocate use-rights to members based on agreed rules. The formal subdivision of land under a CPA, and the issuing of title deeds for individual plots, is a possibility, but is generally not encouraged. Other recently-introduced laws, such as the Land Reform (Labour Tenants) Act, 3 o f 1996. and the Extension o f Security o f Tenure Act, 62 o f 1997, also fall under the tenure reform banner, but are applicable only to occupants or tenants of privately-owned land. Within this legislative framework, there is little or no opportunity for plot-holders on schemes such as the Arabie-Olifants to achieve any change in their tenure position. The most important of these laws is probably tbe Interim Protection of Informal Land Rights Act (IPILRA), which gives blanket legal recognition to all established land users and effectively replaces the PTO system. This law did not feature in any of our discussions at the Arabie-Olifants scheme, or with provincial officials, and it can be assumed to be widely unknown among its intended beneficiaries. The legal vacuum created by the abolition of the apartheid-era land laws, and with it the system of PTOs, is also widely unacknowledged. While IPILRA does not grant any new rights to people on communal land, it does, for the first time, give legal recognition to those rights which have been established through custom and practice. This does not only include exclusive rights to residency and arable plots, but also shared rights such as the right to graze livestock, collect firewood or cut thatching grass, even where these rights overlap with the rights of others. An important aspect of the Act is that while it recognises the rights of individual occupiers, it also recognises the rights of communities to administer their affairs in line with established custom and practice, as long as this is in keeping with the provisions of the South African Constitution with regard to matters such as gender equality, due processes and compensation for those deprived of property rights. In the area of gender equality, in particular, IPILRA is clearly at odds with much established practice, but this has yet to be tested in the courts. Of particular importance is the requirement contained in IPILRA that all members of an affected community be consulted and give their approval before any changes in land rights can occur. As land on the scheme is vested in four tribal communities - Mallala, Mampana, Masemola and Masha - all the members of a group, not just the members who currently hold irrigated plots, would be entitled to a say in any changes to their portion of the irrigation scheme. Any attempt to transfer ownership of the irrigated plots to the current occupants, for example, could be seen as depriving non-plot-holders of their right to access plots in the future and could therefore be illegal. On the positive side, IPILRA provides a mechanism whereby communities can, in an open and democratic manner, express their opinions and influence proposed changes in land use or ownership. While the Act does not in itself initiate change, it does ensure that any body, including government, proposing change must first win the support of the community. 10. Conclusions Access to irrigated land on the Arabie-Olifants irrigation scheme is obtained primarily through membership of one of the four so-called tribal communities that reside there, subject to availability. Land is allocated by the chiefs and Tribal Authorities to heads-of-households for the exclusive and indefinite use of that household. Traditionally, the head-of-household is expected to be a man, but women can achieve the status of head-of-household through death of a husband, divorce or, in the case of an unmarried woman, through establishing a homestead for herself and her children. Land is usually registered in the name of the head-of-household, although there is a trend towards registering plots in the names of users, most of whom are women. Most irrigated land is in the form of plots on the formal irrigation scheme, ranging in extent from 0.7 to 5.0 hectares. Recent years have, however, seen the rise of a number of so-called food-plots, of a much smaller scale, most of which are allocated to women. In most of these small schemes, land is allocated by the chief to a self-managing group of would-be farmers, and access to land by individuals is obtained through membership of that group. There has also been a rise in private, or individual, irrigation away from the formal schemes. Such land is also allocated by the chiefs, with or without a PTO certificate, although irrigation is commonly by means of unauthorised connections to the canal system. Land, at the level of the individual plot, is entirely under the communal tenure system, a hybrid of traditional African practice and apartheid-era legislation. With the exception of two farms, the land is nominally owned by the slate, but allocated for use purposes to specific communities (‘tribes’). Occupiers may hold on to their land indefinitely as long as they use it and do not interfere with other users or antagonise the chief. They may bequeath their land to members of their family but may not sell, mortgage, lease or sub-divide it. Informal share-cropping and other land-sharing arrangements are tolerated on a temporary basis. There would appear to be no effective mechanisms for removing land from people once it has been formally allocated. The communal system thus offers occupants a high degree of tenure security although it continues to exclude them from access to commercial credit. The legislation that governs communal tenure was repealed in 1991, leaving a legal vacuum, but plot-holders and local administrators are largely oblivious to this development. The majority of plots on the Arabie-Olifants scheme are worked by women, who suffer from a range of discriminatory practices when it comes to formal land rights. This is rooted in African traditional views of women as subordinate to fathers and husbands, and was entrenched in the legislation that until recently governed the administration of communal land. Women gain access to land through applying to the Tribal Authority in their own name, through inheritance or by arrangement with their husband or other male relalive. All liie Tribal Authorities on the scheme dlscriiiiinale between male and female applicants, although the traditional bias against woineti holding land in their own right would appear to be dimlnisliiiig. Wives have no auloiiialic right to niliern land Iroin their husbands and can, in theory, be displaced by other wives, oti spring or other relatives on the death of a husband. Recent legislative changes give some tlieorelical proteclion to woiiien in teriiis ol sex equality but diis has yet to make itself felt at the local level. Women’s desire lor land is widely acknowledged on the schenie, and recent developments would appear to have enhanced women’s access to irrigated land. At least one Tribal Authority has a policy of registering plots in the name of the user, regardless of sex, and membership of the small-scale food-plot projects initiated in the past ten years is predominantly female. Such changes, however, raise the possibility of new forms of gender differentiation, with women concentrated on smaller, ‘subsistence- oriented’ plots and men dominating the larger, ‘commercially-oriented’ plots. The limited possibilities for tenure reform under current government policies are outlined above. Pending the introduction of new legislation, there is unlikely to be much change in the formal tenure arrangements on schemes such as the Arabie-Olifants. Plot-holders and community leaders were found to be largely content with the current communal system, but concerned that the lack of title deeds excludes them from access to commercial sources of credit. Tliere is strong support for continued community control of land and opposition to privatisation of holdings or a market in land. Pressure for tenure reform is coming mainly from government, and the absence of western style private property rights appears to present a problem to many officials with responsibility for rural development. Substantially improved access to land, for men and women, will require the enlargement of the areas available to black people. Despite the repeal of overtly-discriminatory legislation, black people are still, in practice, confined to the relatively small and over-crowded areas of the former homelands. It is regrettable that the government’s Land Restitution Programme, designed to restore the land rights of those who were dispossessed under previous regimes, has yet to have any impact on land-holding in the Northern Province. Some progress has been made under the market based Redistribution Programme, but not yet in the area of Arabie-Olifants. Without significantly improved access to agricultural support services, black farmers will continue to struggle to make productive use of available resources. The unilateral withdrawal of credit, ploughing, pumping and other services by government agencies in recent years has left many small-scale farmers in an impossible situation and led to the collapse of agricultural production in many areas. Any reform in land-holding or tenure will have to be accompanied by a complete overhaul of slate and other support services if black farmers are to overcome the legacy of poverty and underdevelopment and make full use of the land and water resources available to them. 0 Bibliography Baber, Rupert, 1996, 'Current Livelihoods in Semi-Arid Rural Areas of South Africa', in M. Lipton, F. Ellis and M. Lipton (eds.). Land. Labour and Livelihoods in Rural South Africa; Volume Two: KwaZulu-Natal and Northern Province. Indicator Press: Durban. Beinart, William, 1992, 'Transkeian Smallholders and Agrarian Reform'. Journal o f Contemporary African Studies, V ol.ll, No.2. Bembridge, T.J., 1986, 'An Overview of Irrigation Development in Some Less Developed Countries of Africa'. Journal o f Contemporary African Studies, Vol. 5, No. 1/2. Bennett, T.W., 1995, Human Rights and African Customary Law Under the South African Constitution. Cape Town: Juta. Bernstein, Henry, 1996 'South Africa's Agrarian Question: Extreme and Exceptional?', Journal of Pea.sant Studies, Vol. 23, No.2/3. Special Issue on The Agrarian Question in South Africa, ed. H. Bernstein. Bromberger, Norman, 1988, 'Cash-cropping, subsistence, and grazing: prospect for land tenure in KwaZulu', in Cross and Haines (eds.), op.cit. Budlender, Geoff and Latsky, Johan, 1991, 'Unravelling rights to land in rural race zones’, in M. de Klerk (ed ), A Harvest o f Discontent: The Land Question in South Africa. Cape Town: Institute for a Democratic South Africa. Cooper, David, 1991, 'Agriculture in the bantustans: Towards development policies', in M. De Klerk (ed.), A Harvest o f Discontent: The Land Question in South Africa. Cape Town: Institute for a Democratic Alternative for South Africa. Cousins, Ben, 1996, 'Livestock Production and Common Property Struggles in South Africa's Agrarian Reform'. Journal o f Peasant Studies, Vol.23, No.2/3. Cross, Catherine and Haines, Richard, J. (eds.), 1988, Towards Freehold: Options for Land and Development in South Africa's Black Rural Areas. Cape Town: Juta. Cross, Catherine and Rutsch, Peter, 1995, 'Losing the I.and: Securing Tenure in Tribal Areas’. Indicator SA. Vol.12, No.2. Davenport, T.R.H. and Hunt, K.S., 1974, The Right to the Land: Documents on Southern African History. Cape Town: David Philip. Delius, Peter 1983, The Land Belongs to Us: The Pedi Polity, the Boers and the British in the Nineteenth Century Transvaal. Johannesburg: Ravan Press. Delius, Peter, 1996, A Lion Amongst the Cattle: Reconstruction and Resistance in the Northern Transvaal. Johannesburg: Ravan Press. Department of Land Affairs (South Africa), 1997, White Paper on South African Land Policy. Pretoria: Department of Land Affairs. de Wet, Chris, 1995, Moving Together Drifting Apart: Betterment Planning and Villagisation in a South African Homeland. Johannesburg: Witwatersrand University Press. Haines, Richard and Cross, C.R., 1988, 'An historical overview of land policy and tenure in South Africa's black areas', in Cross and Haines (eds.), op.cit. Haines, Richard and Tapscott, C.P.G., 1988, 'The Silence of poverty: tribal adtninistration and development in rural Transkei', in Cross and Haines (eds.). op.cit. Harries, Patrick, 1989, 'Exclusion, Classification and Internal Colonialism: The Emergence of Ethnicity Amongst the Tsonga-speakers of South Africa', in Ixiroy Vail (ed ). The Creation o f Tribalism in Souihern Africa. London: James Currey. Hendricks, Fred T., 1990, Die Pillars o f Apartheid: Land Tenure, Rural Planning atul the Chieftaincy. Uppsala: University of Uppsala. Lacey, Marian, 1981, Working fo r Boroko: The origins o f a coercive labour system in South Africa. Johannesburg: Ravan Press. Lahiff, Edward, 1997, Land, Water and Local Governance in South Africa: A Case Study o f the Mutale River Valley. Manchester: Institute for Development Policy and Management. Levin, Richard and Weiner, Daniel, 1997, 'From Apartheid to Development', in R. Levin and D. Weiner (eds.). No More Tears: Struggles for Land in Mpumalanga. South Africa. Trenton, NJ: Africa World Press. ( Mpahlele, R.E., Malakalaka, T.M. and Hedden-Dunkhorst, B., 1999, Potential and Constraints of Smallholder Irrigation in South Africa: A Case Study o f the Olifants River Irrigation Scheme. Report prepared for the International Water Management Institute, Colombo. Nieuwenhuysen, J.P., 1964, 'Economic Policy in the Reserves since the Tomlinson Report'. South African Journal o f Economics. Vol .32, No. 1. Robertson, Michael, 1989, 'Dividing the Land: An Introduction to Apartheid Land I,aw', in C. Murray and C. O'Regan (eds ). No Place to Rest: Forced Removals ami the Law in South Africa. Cape Town: Oxford University Press. Shacklelon, C M., 1993, 'Are the communal grazing lands in need of saving'.’'. Development Southern Africa, Vol.lO, No.l. Solinjani. SB. , 1986. The Dynamics o f Land Allocation in Transkei. Paper prepared for I jn d Tenure and Rural Development Workshop, Institute of StKial and Economic Research, Rhodes University. Surplus People Project, 1983, Forced Removals in South Africa: Volume 5, The Transvaal. Cape Town: Surplus People Project. Union of South Africa, 1932, Report o f the Native Economic Commission 1930-1932 (U .G. 22/1932). Pretoria: The Government Printer. Union of South Africa, 1955, Summary o f the Report o f the Commission fo r the Socio-Economic Development o f the Bantu Areas within the Union o f South Africa (UG 61/1955). (The Tomlinson Report) Pretoria: The Government Printer. I Vink. N., 1986, Systems o f land tenure: Implications for development in Southern Africa. Research Report No. 6, Development Bank of Southern Africa. Watkinson, Eric, 1996, Enforced Agricultural Change in South Africa: The emergence o f a small class of commercial African farmers. Master's Thesis, Development Studies, University of Natal. Westaway, Ashley, 1995, Land and Local Government: Policy Paper 27. Johannesburg: Land and Agricultural Policy Centre, Yawitch, Joanne, 1981 Betterment: The Myth o f Homeland Agriculture. Johannesburg: South African Institute of Race Relations. o