Land tenure, territoriality and ecological instability: a Sahelian case study

I Jan van Dijk

Land Tcnurc, Terrilorialily, and Ecological Instahility:

In recent policy documents and poliey oricnled research il is consislenlly a i -gued Ihal Ihere nccds lo hc more allenlion lo local land Icnurc icgimcs in ordci lo arrive at inoie suslainahlc fornis of nalural resource management in the Sa-lie!.2 Solutions for environnicntal management aie sotighl in apprcciating the diversity of local tcnure regimes and in sponsoring a variely of forms ofco-tnanagenient of resources by the administration and (lie local population (Ilesscling and Ba 1994). Olhcrs are more hésitant on the grounds ihnl cnslomary tcnure regimes have been dceply inflticnccd by state inicivcnlinns. loading to iion-sustainablc management end ecological degiadalion (Mooie-liead 1991 ), und Ihat the complcxity of customary tcnure regimes is difficull lo caich in wrillcn and codifiée! law (Vcdcld 1995).

One of the problems is (bat Ihere are few detnilcd .studies of what (hese lo-cal Icniire regimes amount to. Espccially the rclalion of land tcnure lo ecologi-cal conditions is problematic. In (scmi-)arid areas wilh a low dcnsily and pie-diclabilily of resources more or less strict rulcs of leiuire give way to moic flexible ways of rcgulating acccss to resources, such as tciritorial control over gra7inp arcas. Particularly in Ihe Sahcl the production of food and forage for human and animal consumplion fluctuâtes heavily in lesponse (o impiedictable raiiifiill palterns. On the onc hand rules of acccss lo resources are critica) for pcoplc (o deal wilh thc.se fluctuations. On Ihe olher hand (lic mechanisms of icgulaling access to resources are closcly inlcrconnected lo Ihe dynamics of so-cio-political organisation. Typically, ecological anlhropologisls, agionomists and geographcrs have taken up the sludy of-ihe management of space. i.e. the nalural environment wilhotil much rcgaul for the,legal complexitics involve<l, whcieas sludents of law have conccntinlcd rïh-jhc complexilies of propcrtv and teniirc and legal pluialism, wilhotil payrng nmch altcnlion lo the agro-ccologi • cal background or ecological conséquences, 'l'his papci will Ihcrelbie concen-Malc on hmv individuals and j'.mtips of pcople g;iin acccss to agriniltuial hind. 1 'Ihe field research on tvJiich Ihis ntticle is basod was eomlficled wilhin Ihe fiainrwfiik of a icseaich piojecl "Pulhc society in :i chaiïpini» wnrld: ('cnlnil Mali" which \v;ti linaiiccd by llic Nelheilands I;oiiii(la(ion loi Ihe Advancemeitl of 'ïrnpical RrM'aich (WO IKO. p.rani ntiinbci W 52-'f9'l). In Iltis research piojrcl, excciiled by llic aiiihoi and Mirjam di1 Uitlijn, llic tlniveisily ol Ulirchl (l)rpmlmeiil of ('iiltnral Anlhuipdocv). and Ihe AjiricitlUiral Univcisity of Wagcninpen (Deparlincnls of Apiaiian Law and l'oteslry) pailiripalcd. Acknowlcdgcnicnt is made of ihe input of Mit jam de Hiuijn in all phases of Ihe project, and the hclpful commrnls of Oerli llcssclinp, Mctino van der Velde and Melanie Wilx:r for Ihis article.

2 Anionp olhers Mcl ain (1992), .Sliamimpaialnnm et nl. (1991), Kint? and Tinoié

18 Van Dijk

water and paslure, and (hè relations belween these processes and historica! and ecological factors at the regional Icvel. This will be donc by discussing a case sludy of ffulhè (agro-)pastoralisls in Central Mali.

In the first section the interaction belween land lennrc and territorial control is discussed fioni a Iheorelical perspective. In the second section a short sketch is given of the agro-ccological environment and the ltiulnric.nl dcvclopinenl of land use in the région. The third section contains a hislorical description of the dynainics of land (cniire and conlrol over paslure. I;inally, soine concln.sions for the sludy of land tenurc regimes will he drawn.

Tcnure and TcrritorialUy

Land (enure tnay be conceived of as sets of rules and practiccs specifying whom is lo get access to land at which time in which place (cf. Lane and Moorehead 1995:116-117). The rights defined may range from private to coinmon propcrty. They may overlap in space and time, may be deiivcd from different ideological Systems and bodies of law (Herry 1988:53), and aie dceply engrained in the social organization of the people involved (Shiplon 1994:349). In the Salie), there is a pressing need for the coordination of the actions of indi-viduals given (he fact (hat often the herding of animais and the cropping of ce-reals are combined in the same area. If cultivators werc to clear land eveiy-where they saw fit in the Inish, the available space for herding the animais dur-ing the growdur-ing season would soon be exhausted. Bolh activilies have to be se-gregated either in time or in space, if only because of the likelihood of damage to crops by wandering herds.

People do nol begin from scratch in solving such problems. They starl with perceptions of tenure, and righls over resources which were developcd by their predecessors. This results in a spécifie type of management of space and allocation of resources for spécifie ends. The landscape is not just a iiattiral Substrate on which people imprint their activities. It is humanised in thé sensé that rights of access to resources provide thé population wilh coordinates to direct their actions in a purposeful manner. Of course, givea climalic fluctua-tions or olher forms of instability, thèse coordinates are subject to change, ne-gotiations and conflict all the time (Berry 1988:65-66, Shiplon 1994:351).

Our social science concepts need to capture bolh the continuity of land tenure, in the sense that it is a historica) process in which people and their pre-decessors have invested labour, political power and so on, and its flexible character in the sense that access to resources must be continually re-assesscd given the constant changes in agro-ecological and socio-political conditions.

The interplay between the need for secure access to resources and the need for their flexible use is hère delineated by the (win concepts "tenure" and "territoriality". Tenure is more likely to be the mode of resource appropriation in cases of more permanent forms of resource use such as the croppiiig of cc-reals, and the construction of water resources whcre people invesi Ihcir labour or olher as.sets in, whereas lerrilorialily is the dominant way of icsotircc ap-propriation in cases of flexible modes of natura) resource management such as the grazing of animais and the gathering of bush products, where people lypi-cally do not invest labour or assets in the ecological environment. Tertitoriality

Land lenui e, tcrritorialily, and ecological instability 19

is (luis qualilatively different from lenure because it engages society in a sys-lem of nalural relations, and because it dénotes 'a process Ilial conlinually goes on'. .Social relations and lennre can be said to have ü more permanent chaiac-'ler, whereas lerrilorialily is more f luid and refers to a 'succession of synchronie States' (Ingold 1986:136). Willtin a spécifie social and cultuial context, nature, i.e. the environmenlal conditions, détermines possibilités for thé use of ic-sources, and Ihus thé size of thé territory needed for survival, rallier lhan Ihat Ihc users of these resources are able lo détermine thé use of pa.stiires and gath-ering groiinds in advance. The density and variety of resouices, which vary enoimously from one year to (lie nexl, sel (lie possibililies for use, and Ihe boundaries and surface which are needed in particular circtimstances. Althougii thèse concepts were developed for a pastoral context in boréal conditions Ihey may be applicable to all kinds of olher ecological zones and production sys-lems.

In most of the literature dealing with "lerritoriality" thé behavioural aspect is slressed. Territorial behaviotir is conceived of as a response to the ecological conditions, i.e. thé density and predictability of the resources, and Ihe changes herein (Dyson-Hudson and Smith I978,Cashdan 1983, Casimir 1992a. I992b). In Ihis way il is disconnected from "tenure", which is defined in légal ternis, or thé maintenance of territorial boundaries is defined as a component of a général framework for land tenure. Casimir (I992b) for example shows how différ-ences in rainfall conditions coïncide with varions (6) types of access to pastoral grazing lands, ranging from open access wilhout any restrictions to private ownership of grazing land.

However, wilhin Ihis larger scheine valions groups of people may work ont quite différent solutions for thé maintenance of (lie boundaries under sirnilar conditions. To account for spécifie modes of behaviour with explanalions of corrélation is soiïiefhing différent lhan explaining thé causes for thé choices people make (Silberbauer I994). Territorial behaviour ultimately must go back to a prédisposition to act, which is cultural, and is mostly bascd on a claim on Ihe territory, which is to be maintained vis-à-vis others. This claim is always juslified by some ideological, moral, légal or political reason, which is not ecological, but is hased on thé perception of the people of their ecological and socio-political conditions (Schlee 1990:13). When ecological or societal condi-tions change this claim in ternis of type of subsistance resources and size of terrilory may be adjusted witli obvions conséquences for territorial behaviour. It is thèse changes in territorial behaviour and underlying légal constructs which are of thé greatest interest when we are dealing wilh climalic variabiliiy bolh in Ihe long and thé short terni. Il is there where thé potentiell for intensil'ied compétition for resources and political and military conflict is the highcsl.

20 Van Dijk

the worst year (worst case scenario), which is siniilar to prcscriplions of conventional range science (e.g. Breman and de Ridder 1991). But, as will be sliown in the course of this article, ecological, hislorical and socio-politica! clynainics have lo hè Ircated in an integratod Framework or at least as inlcrdcpcndent variables loexplain terrilorialily.

So the dynamics of territorialily are essentially different from lentirc, bc-cause of the ways in which boiinclaries are maintained. Tenure can be under-stood as the assignmenl of a spécifie right, for a specified aniounl of time In an individtial or a group over a demarcated object, a pièce of land a certain tree species etc. This is not the case witii (erritorialily, bccause the boinularies are (luid both in social (Cashdan 1983) and geographical lenns (Dyson-lludson and Smith 1978). Another important aspect of the tenure of resources is (hat pcoplc invest in these. In fields, which are held in tenure, people invesl labour or manure and try to conlrol natural processes by consciously changing the végétation cover. Because of this the claims on agricultural land are nuich more fixed and codified in legal mies and political hiérarchies and therefore nol subject to the fast changes as posttilated in the case of territorialily. Anolhcr différence is (hat in the case of paslure, not the range ilself, hut the (prodticts of the) pastoral animais are the prime object of appropriation. In the case of a field the land itself is appropriated.

The différence betwcen the two concepts is of course relative. As wc will see below all kinds of intermediate situations exist. Moreover, conlrol ovcr ler-ritory is often closely associated with tenurial rights over other resources, ('on-trol over a source of waler may facilitate and make possible thé use of a spéci-fie territory (Thebaud 1990). Wltile the boundaries of a lerritory may bc ill-de-fined, access to the water resource may be limitée! (Kinlz and Traoré 1993). Especially in thé Salie! rights of access to water resources aie exlremely inipor-tant and hâve a major political dimension. Conflicts betwcen pasloralisls and scdcniary cullivalors and among pastoralisls oflcn hâve Ilieir background in thé management of walcr resources. The devclopment of ncw waler resources is of-ten a reason for the eruplion of conflict and conlrol over water lias been tised as political instrument by governmenls to control spécifie groups of pastoralists and to favour others.

On thé other hand territoriality, and ils result, conlrol over Icrrilory, is a necessary precondition for thé exploitation and tenure of resources witltin the boundaries of thé territory. This control is often not obtained (and challcnged) by légal means, but instcad by bribery, violence, raiding, military delencncc, polilical relations, and rilual (Johnson and Anderson 1988, Schice 1992).

So, tcnurc and lerrilorialily slaml in a dialeclical relation to cach olher. Thcy can bc saicl to bc thé Iwo sidcs of Ihc saine coin, namcly (lie conlrol over the nicans of existence. They are différent in quality and in llicir dynamics, bul are inlimalely connecled, because thé one rentiers (lie olhcr effective and in re-verse.

Land leniirc, terrilorialily, and ecological insfahilify

Inlcnsificnlioii of Land Use Jind IÇcologicsil Dynamics

2l

Krxenrrh Ai en

Tlie research area is located in Mali, a land-locked Sahelian country (sec map I). Administratively Mali is subdivided in 8 régions (7 on thé map), which in turn are subdivided in cercles (districts) and in arrondissements (sub-dislricts). 'The country gained ils independence in I960 afler more lhan 60 years of French colonial rule. Before colonisation Ihis part of Africa was mied by scv-eral savannah-enipires. Mali belongs to ihe poorest connlrics of Ihc world, wiih a inean ODP per capila of $ 270 in 1989. Within Mali thc région of Mopti, of which the research arca is Ihc poorest part, is worst off in (crins of lifc cxpccl-ancy and chikl morlality.1'

M n p I: Location of Central Mali and the research area (adaplcd from De Brtiijn and Van Dijk 1995).

22 Van Dijk

In the recent past the research area, called Hayre localiy (see map 2), was the scène of spcclacular climatic variability. It is looalecl in the central dry-land part of the Niger Demi, soutJi of the inoiinlaiii range Connecting (lie Bandiagara plateau with Mount Honinibori. The région is vtilncinlilc to periodic dioughl. Menu anntial lainfall over the peiiod 1976 1986, ineasincd on the t wo nearesl wealher stations was 286 nun in Booni and 338 mm in Monndoro willi Mandai d déviations of 33% and '11% icspecljvcly. Oscillations in biomass pioduclion are piobably bighcr than in rainfall.'' The distribution of rainlall over the légion is equally insecure. Over the pas| decades these variations in rainfall seeni to have increased, along witli a decrease in total rainfall (Hesse and Thera 1987:

15).

Our field research was done in Serma, a gronp of seini-pernianent setlle-ments, inhabited mainly by Fulbe pastoralists of the clan Jallube, a smal) mnn-ber of Riimaybe families (former slaves) from very diverse origin and a couple of families, bclonging to the ruling class of Fulbe society, localiy called the Weheebe. The core of Serma is formed by a sedentary hamlet (Debere) of the Riimaybe cultivators, the Weheebe families and some impoverished families of Fulbe pastoralisls. The sedentary hamlet is the social and commercial centre of the Community. Around this core 8 temporaiy settlements are localed, whcre the pastoralists can be found during the rainy season. Tlie total number of in-habitants is very hard to fix with certainty, but is somewhere around 500. All inbabitants of Serma engage in the cultivation of millet the main staple erop. The Riimaybe cultivators regard this activily as Iheir main mode of subsistance. This is in contrast to the Fulbe pastoralists, who regard transhumant livestock keeping as their main activity. Ilowever, due (o the décline in livestock num-bers they are oflen erop cultivators as well in practice.

The inhabitants of Serma exploit (wo agro-ecosystems.^ Norlh of the scl-tlenients np to the motintains at 30 kilomètres there are flallands overgrown with very dense biish, called tiger bush or broute tif>iée in the htcralure (e,g. Gallais 1975), and Ferro in the local language (sec also Marchai 1983). The soil of these flats is loamy with latérite concrétions in the subsoil. In belween the tracts of bush land bare land can be found due to the combined effects of drought and poor soil quality. This alternation of végétation and bare sliips gives this végétation type ils typical tiger skin pattern, when viewed fiom aerial photograpbs, from which it dérives ils name. In between the bush, mainly an-imal grasses, including wild fonio, and sometimes perennial grasses can be found. The végétation provides good fodder for browsers like camels and goats, but the het bal layer is loo sparse to afford niuch grazing by caille and sneep. Under the present rainfall regime exploitation of this area for erop culti-vation lias only limitée! possibilités.

The second agro-ecological zone exploitée! by the inhabitants of Serma is localed south of Serma and is called Seeno-Mamigo This term désignâtes the 4 Spcctacular fluctuations in biomass production wcrc ine.isurc«! hy a research le.im of CIPHA (Centie Intcintitionalpmir l'Elevage en Afrique) In the Noilhcin .ind Soiilliern Gurma whfie the research aiea is located coefficient1; of variiinrc 1110.111 m;mmum biomass production wcre measuted of 86% and 64% respective ly apainsl coefficients of vnn.mcc of incan animal r.iinfall of 37% and 28% (De I .eenwel al 1991. MS). 5 If we do nol coiint the agio-ecosystems which are exploiled hy the l'tilhe hcrdsmen when on lianshiiniance during ihe dry season.

/ and te nui c, tcnitotïality, and ecological iintahiliiy 21 sandy soils, and in particular (lie fixed dunes thaï form the dominant geological formation. The Seeno-Manngo offers a very open rolling sleppc-likc landscajx:. The végétation is dominated by animal grasses. Trees are scarce. The paslures, consisting of animal and perennial grasses are of excellent qualiiy (dallais 1975:126). The soils of thé Sceno-Manngo arc also veiy .snilable for thé giow-ing of millet piovided thaï thé rainfall is abundant and maniirc lias bccn ap-plied, for (hcir nalural ferlilily is very low.

Map 2: The research area (adapted from De Bnn'jn and Van Dijk 1995)

SEFMO MANNOO

Over the past decades the végétation bas coiisidcrably degracled due to the im-pact of drought. A large nuniber of tree species and perennial giasses have be-come extinct. Water resources have dried up and biomass production lias signi-ficantly declined. Rspecially in the 1980s erop production consislently faik-d to feed the population, cxcept for 1988 when a bumper erop was harvcslcd Livestock was decimated during the drought of 1971 -1971 and again m 1981-1985, by as much as 75% on the lader occasion The hcrds have h a r d l y rc-coveied since Ihen. As a lesult a large number of pcople have nugt.ilod to the south. The remainmg people suffer from chronic m a l n u t r i t i o n .

Van Dijk Land lenitre, tenitoiialily, and ecological inttability 25 appear during (hè rainy season, bul these dry oui quickly in Ihe dry scason.

Given Ihe perméable soils of (hè Seeno-Manngo Ihe prospects for (he collcclion of surface waler are even more limiled Only in Ihe border zone belwcen these two agro-ecosyslems larger pools are formeel due (o run-off water from the flats Ihal flows inlo Ihe small dépression helween the Ferro and Ihe firsl cluncs II is precisely there where human setllemenl was possible in the pasi. In Ibis tiansi-lion zone we find a doren temporary and permanent settlemenls of Fulbe pasto-ralists, Riimaybe and Hummbeebe6 cullivators from west (o cast. Sernia is oitc of (hese seltlements locaied near onc of the large ponds. Only in the 1970s per-manent wells were dug in this (ransilion zone.

Map 3: Sketch map of the village Serma and suiroundinps (from De Bruijn and Van Dijk 1995).

„ _ _ _. Tnol path

Irtfiporary pond Camp Rlimayt* hftmlff

At present all agrirullural fields are in permanent use for the cullivation of millel, Ihe mam slaple erop and nre all located in the dépression scp.iratinp the Ferro from the Secno-Manngo or on (lic Seeno-Mannpo (see map 1). 'l hete is only one field loralcd on the Ferro.

6 Hummbeebe are frrc cnltivators. Unlikc Ihe Riintayhc Ihcy have nevcr Ix-cn enslavccl, anti xvcrc allies r.iiher ihan slaves (o ihc Faillie. Tlicir Innguage and culliiic aie relate«! l<i (lic Dogpn. \vlio rnn l>e found in «ome villages in (lic ninunlniiK, nc:u Ihe Bamhagaia pl.ikMii lo Ihr wesl and in Ihe plains lo (lie

Intensification of Land Use: A Ilixtmical Oveiview

The hislory of the intensification of land use can be snbdividcd inlo three phases. In the first phase, which lasted welt inlo the 20lh ccnlury, Serma was a lernporaiy cultivation hamlet, due to the absence of dry season waler resources. Fulbe herdsmen and their Riimaybe dépendants praclised an extensive buUi-fallow cultivation system with some permanent fields on the best accessible plots near water resources which retained waler for some linie during the dry season, e.g. near Debere. As a result of tbc temporary character of habitation the politica! slalus of Sernia was ambiguous. It was dependent on a larger vil-lage, 30 kilomètres to the north/Booni, where the eilief ofthe Fulbe pastoralists rcsided. This political dependence lias remained unchanpcd, even fhough Scr-ina lias long since becomc a permanent village.

As a conséquence Üiere is no lineage or group in Serma who can claim lo be the founding lineage and is Ihus able to control access to agricullural land as among cullivating neighbours or sedentary Fulbe villages, or can lay claim lo exclusive use of pasture tcrritory. The Torodbe, who were the first to setlle on the land where the sedentary hamlet of Serma, Debere, is now posilioned, weie and are a prominent lineage in the village, but they cerlainly do not perform Ihe lasks assignée! (o the "diefde (ene" in neighbotiring Dogon, fluininbecbe. and Sonrai villages (sce e.g. Gallais and Boudet 1979. Kinlz and Traoré 1991). Within the Fulbe political hierarchy, the cbief was the ultimale autliority over Ihe allocation of agricuitural land and pasture, hè was the master of tenurc and lerritoiiality. This is quite logica) for he beid ultimate responsibilily for Ihe coordination and timing of herd movemenfs, which should nol interfère with Ihe cultivation of cereals which provided his stibjecls wilh their basic food.

During (lie colonial pcriod (l893-1960) Ihree independent factors emcrgcd which piomoled the inlcnsification of land use. The first was the graduai libé-ration of the slaves (Riimaybe), so thaï the pastomlisls had to face incrcasing laboui shoilages for the watering of the animais and the cullivalion of rereals. Most of (hem possessed too few catllc to rely on liveslock kecping alonc.7 Secondly, as a conséquence of the suppression of warfare and plunder, popula-tion and liveslock numbers begnn (o grow, so thaï pressure on the exisling ic-souices around Serma increased." The third factor was the abundanl rainfall in Ihe !950s and the 1960s, which togelher with applying cattle mannrc (o the fields created a potenlia! for the intensification of land use nrouml Serma. 'l hè only (hing lacking was a permanent source of water for human and animal con-siimplion, allowing cattle and their hctdsmcn lo romain longer in Serma lo pm-duce (lic ninnure neccssaiy for Ihe inlcnsification of ccreal cullivalion.

The second phase starled wheit undër|round waler reservoirs wc ie 'discovered' in the impénétrable layers of c!a*y and latérite in the bed of tbc pond adjacent lo Debere aroiind 1935 (see map 3). These reservoirs were

pmh-7 This is confirmed by (lic ntimbcr of animais Ihat were laxed in Ihe rhiefdom of Doom diiiing the colonial period.

26 Van Dijk

ably dug by ancient inhabUants of Ihe sile, who had disappeared before Ihe Fulbe occupied it. Jn the course of linie the openings of the reservoirs had caved in, so that they were difficult to discover and had to be oj^ened and re-conslnicted. The water reservoirs bccaine Ihe properly of Ihe people who dis-covered Ihem and invested their labour inlo their reconstruction. This allowed a certain mimber of people and animais to reinain longer in Senna during Ihe dry season. Some time later, at the end of the 194()s the discovery of flint stone on Ihe Seeno-Manngo sorne kilomètres form Debere suggesled Ihe présence of water in Ihe subsoil. With some difficulties the Fulbe and Riimaybe dtig a well, wliich began to function in 1956, and was called Yaraama.

The water reservoirs and the well at Yaraama enabled herdsmen and Iheir cattle lo remain all year round in Serma. When the ponds dried up afler the harvest Ihey watered Iheir animais at the waler reservoirs near Deberc. During this period they corralled their livestock on their fields around Ihe hamlet at night, so that at Ihe start of the rainy season a thick layer of manure was lef!. This mantire ensured a higher productivity of cereal cultivation than Ihe bush field they cultivated before. When the water reservoirs were finished they moved willi their animais to the well at Yaraama, where they also had fields to fertilise. When the rainy season started Ihey moved their homes and animais to newly established caltle camps where Ihey camped during Ihe rainy season.

This development was accompanied by the immigration of a mimber of rieh pastoral families from the east, belonging to the Seedoobe lineage.9 These Seedoobe had left the area at the turn of the Century after a conflict wilh the Fulbe chief in Booni. They spread across the whole zone. They simk a well at 20 kilomètres east of Serma and re-occupied the land of their grandfaihers in Serma. They participated in the sinking of the well at Yaraama. They dug water reservoirs and bought otlier ones from their original owners in Debcie. They drove their cattle into the territory of Serma, and established caltle camps hère. Soon they were among the most prominent and powerful inhabitants of Serma, because they possessed large numbers of caltle and were relatively independent of Ihe chief of Booni, because Ihey had been allowed lo form Iheir own admin-istrative unit by the colonial government.

The effects of all these changes on the land allocation, land lenure and terri-toriality were tremendous. The bush fields had to be abandoned. l'eimanent fields were concenlrated around Debere and aroimd the well in Yaraama. Land began to change hands via commercial transactions. New semi-permanent fields were established at the sile of the newly eslablished caltle camps. Claims on land were reactivated. People were forced out of the village. Territorial dis-putes arose, which led to new power balances in the village. A new hybiid complex of land teniire mies developed, which gave shape lo the spatial alloca-tion of land as depicled on map 3.

In the following section the land teniire mies which emerged oui of these changes will be discussed. We will see how different ride complexes aie hislor-ically juxlaposcd upon each olher. The next phase which is characierised by Ihe récurrent droughts of Ihe 197()s and I980s, chronic food shortagcs and

in-9 11 is ilifïicull lo say whether Ihis wave of mignilion prnmpird llic chanpi". in land lenure and llic ilcvclopmcni in wnler lesourccs, 01 the levcise.

Land lenure, tciritoriality, and ecological imtabiliiy 27 creasing state involvement in natural resource management and land teniire will be discussed hereafter.

Forty Vears of Change: Local Dynamics of Teniire and Territoriality

Permanent fields

At present the greatest concentration of agricultural land can be fotind around Debere, the Riimaybe hamlet. Most land worked here is owned by Ihe paslo-ralisls. Given the increasing pressure of liveslock (predominantly cai(le) and the increasing risk of erop damage on the bush fields in the 1950s and 1960s, a inunber of Fulbe pastoralisls and Riimaybe cultivators had to abandon bush fields which extended too far into the pasture area claimed by the newcomers. Tliey, as well as the newcomers, nished to get land and water reservoirs al De-bere, where there was less risk of damage lo the ciops by livestock. This led to the émergence of a lively land market. The land nearest to the pond, Debere and the water reservoirs was regarded as the most precious, because of Ihe ma-ntire left behind by animais which pass the fields on Iheir way to Ihe pond and Ihe waler reservoirs and because of the village wasle of Debere. Status is an addilional factor which makes land near Debere more valuable. Those who have land near the hamlet are closer to the founders of the village and are said to have more power. Of the 76 transactions in land that were recordcd over Ihe period 1955-1990,'° 50 transactions involved the purcha.se of land. The olher 26 involved inheritance, borrowing or renting of land. So we may safely con-clude that Üie distribution of land at this site is thoroughly commetcialised.

llie distribution of agricultural land is not problemalic al present. In con-Irast lo Ihe past there is no shortage of land for cultivalion, ihough (hè quality lias decieased over time. Rather there is excess land and a polential for allow-ing more people in Serma, because many people leave. This is also reflecied in the priées for land, which are rallier low, if compaied for example to the bush near lowns, where numerous ecological refugees have sellled. Despite ihe un-dcr-use of erop land and the low stocking density on Ihe pastures around Serma. there are no people from ouiside who have recently boughl land or wa-ler reservoirs. The fact that there is a markei for agriculiural land and (hè faci lh.1t Fulbe of various lineages and Riimaybe bought land in the pasl, does nol mean Ihat at present someone can enter the village and buy himself an esiale. There can only be one explanation for this phenomenon and that is Ihat Ihe land market is only by appearance a free compétitive market.

Inheritance of land for erop cultivation is reslricled lo male k i n . I f an inliab-ilant of Sel ma wants to seil his land," hè is expected lo offer it lo Ins kinsmcn first. If tbey do nol want to have it, or if they do nol have Ihe money, hè may offer it (o olher people. The purchase of land is, howevei, slill subject lo Ihe

10 The sample includecl only part of Ihe fields.

28 Vfin Dijk

control of the stntdu banba, i.e. the comnuinily of men in Sorma.12 This makes it difficult for an outsider (o obtain land or control over olher assels in Sernia, unless hè is prepared to settle in the bush, where no-one can stop liiin, and (o wait until hè is accepted as an inhabitaiit of Serina. But in practice there are other ways to get land in Serrna. Fields may be borrowed, and may be obtained through marriage, if the wonian does nol have brothers to inlicrit the land. However, for someone who is nol related lo people in Sernia, or does nol réside in Sernia, it is almosl impossible to get land.13 In the past only people who had a new or indispensable contribution to village life to offer, such as Islamic clerics, Tuareg wood werkers, and a tanner, were admittcd (o the village and were given land.

The suudu haaba controls land transactions through indirect but effective nieans. In Serma this communily control may go so far as to force someone lo seil his land. When an inhabitant of Wuro Kaaral wanted to seil the land hè inberited from his father, as hè planned to go away,14 his elder brother objccted. Tlie latter wanted to add Üiis field to his own field to broaden his own (meagre) base of existence. Moreover, as the oldest brother, hè slood on his rights. However, the siwdu baaba forced him lo consent. Tliey wanted a respected Is-lamic clcric, famous for his rain-making and healing capacities to buy the land. This cleric livccl in Wuro Kaaral, bul did not originale in Serma. The men of Serma wanted (o give him a permanent basis in Serma to keep him in the vil-lage. Otherwise, no Diimaajo and no outsider has ventured to buy, lease, or borrow any field at Yaraama in the 35 years since its establishment. Control by the suudu baaba prohibited this through mechanisms that are relalively unclear. Thcre is no organixed forum for discussing problems and most of the negotiations take place in small group discussions on varions occasions in dif-ferent locations. Whcn consensus over a décision émerges, the desircd course of action is somehow made clear. H scems thcre is no public bidding and nego-tiation about the details of the transaction. Everything appears lo be pre-ar-rangcd, the buyers, the price, and the timing of Ihe deal. It is vcry difficull for powerlcss outsiders to gain access to this circuit.

Tliis set-up allows the inhabitants of Serma to slay in conirol of Iheir own village territory. The fact (hat all kinds of deals are concluded by informai, al-most secret, negotiations and thaï heavy pressure is exertcd to influence the outcome of the negotiations, indicates that territoriality is a major concern. An outsider who gains access fo agricultural land and water in Serma also bccomes a compclitor for grazing and olher resources. According lo the old men tcrrilo-12 For want of spncc we will not discus.? (he funclioning of the suudu haaba (lilcrally Ihe house of Ihe fnlhcr) hcre. The (erm, howevcr, serves originally lo indicate male kin gronps, hul mny also br applied to orgnniMlional Icvcls bascd on résidence such ns Ihe camp and Ihe villnpc (even includiiig Ilic Riimaybe). This should lx-, sufficirnl lo iiidi-c;ilc thaï il is a veiy flexible org.inirntionnl dcvicc, whicli can bc nianipulnlrd lo serve Ihe purposcs of Ihosc in charge at Ilinl nioment (sce also De Bruijn and Van Uijk 1995: 145-146, 165).

13 Thnl is thosc onisidcrs who gain access lo a pièce of land aro firsl made inlo insiders bcfore Ihcy can borrow or buy land (sec also De Bruijn and Van Dijk 1995:310).

14 This stiatepy of 'asset depletion' is often followccl in cases of extreme dcprivalion. It also indicafe* Ilial Ihe ncrson involved does nol inlend lo return, which nctually happens

^1^^^

Land leniirc, f en ilarialily, and ecologiral inslabilily 29

rial concerns are indeed the main reason for the high degree of consensus in land tenu re tnatlcrs in the suudu baaba. Only by mainlaining consensus the

stiiidu baaba Serma can exert control over commercial transactions and

in-heritance matters. It is helped in this respect by the fact that there are no year-round open-access water resources in the immédiate neighbourhood of the set-tlement, so that outsiders are not able to stay in the neighbourhood of Sernia all year round unchecked. In order to keep it that way interférence from outside is avoidcd as nuich as possible.

The government may attempt to interfère whcn conflicts willi large hcid owners escalate. Thus mechanisms have also been developed to résolve conflicts over damage to crops by wandering herds. No compensation is de-manded if the damage is moderate, and even if the damage is great, the formal fines outlined undcr cuslomary law are not imposed. Rallier people negoliale by rneans of intermédiares. If a landowner persists in demanding cuslomary law penalties against the owners of herds, the conimunily responds with econo-mie and social sanctions. In one case, now over 20 years old, a fonner slave <lc-manded compensation and has since been boycoltcd by the pastoral hcrdsmen. Tliis forced him to rely on passing herds for. access (o valuablc manure. Bul if herdsmen passing Ihrough the area remain loo long and cause loo many prob-lems, fines may be levied loo. Dut the local people are al ways aware Ihat lliis may bring unwanted government interférence into Iheir lives. Government agents not only corne to solve problems, but also to line their own pockets with the fines for putative transgressions of the law, payment for unwanled and un-necessary services and even through intimidation. Since the inhabitants of Ser-ma have unclergone impovertshmenl over (hè past several decades, Ihere is no money to bribe such government officials, These considérations and Ihe asso-ciatcd territorial concerns play an impôt tant rôle, in aulhori/.ing the siintlu

baa-ba Serina lo keep conflicts aboul erop damage-and access lo resources

cnniain-cd wilhin thé village.

The well af Yaraama

The way agricultural land around a new water point, where no previous land rights were defined, is appropriated by society is'cnlirely different from what we saw wilh respect to Ihe permanent-fiéftls. 'HitS is clear from the way Ihe in-habitants of Serma dividcdlhe Innd nröitnd the well of Yaraama in the I950s. Participation in the sitlking o f t h e welfwas af pfe-condilion for gctling a sliarc in the land. Résidence in Sernia was n'of'a" ïeces'sary condition, since scvcral people who oblaincd land at Yaraama did not réside in Sorma al Ihat linie. Thcre aie still people living clscwhere who have land al Ihe well. So, Ihetc wcrc, ,'ipail fiom the valions l incages and social nilcpoi ics involved, Iwo main blocks that participalcd in Ihe division of the land: on the onc hand the people who had been living in Serma for a long time, such as Ihe foumlcrs of Serma. and olhers from diverse lineages; and on Ihe other hand a group of ncwcomcrs belonging lo Iwo other powcrful lineages.

com-30 Van Dijk

the politica! elite (Üie Weliêèbe) under wlïich the Riiinaybe were also groupecl; and the rest, aniong whoin also immigrants of lesser standing, Itach group oc-cupied a sector of 90". In fact no niaster-servant relation exislcd beiwecn the Wchccbc'iUMl Riiinaybe, bul lhe-l«ilbe paliloralisls in Serma had no longer any bonds wilh Iheir own Riimaybev'and no bonds with other Riiinaybe as yet. In the end the Riiinaybe never used Iheir- fields at Ihe well, because tlie Wehecbe look all the fields nearest to)the*we!I,".where the passing cattle leave their drop-ping«. Oiven the high status-and the'political influence of Üie Weheebe and the low status of the Riiinaybe, the latter cöuld <!o nothing bul accept.

The land wilhin each sector was divided along lines of senioiily. The oldest of the lineage (ook the (lelci'nearëst to the well, so (hat he wou ld profit most froni (lie droppings of the animais watered at the well. Then his juniors follow-ed, then the elders of the next génération and so on. In fact the lineage organi-zation is attached tb the well ¥ith ils apex at the well and its base in the bush. 'l'his way of dividing may besregarded"»as typic.il of a pastoial segmenlary line-age system, which is typical"for many-African pastoral societies.

However, the mode-of inheritance-of the land at Yaraama does not follow tliis model which prescribes that each time the eldest dies the land around ll)e well bas to be redistribtited to fit the *new^ lineage organization. Tliis wou ld make the investtnents of an individu.il in the land in the form of maniire worthless. Instead succession to land at the well is arrangée! in a form Ihat may be termed Islamic. The sons each inherit from their father an equal portion of the land. In this way the land becomes increasingly fragmentée! over time. The original lineage organization also tends to be blurred by the fact Ihat many fields of elders are worked by their sons.15 Land lias been lent to others and even sold, for example when people migrated under the impact of drought. Be-cause of this the lineage organization around the well is no longer intact.

The water of the well is neither an open access water resource, nor a private property resource. In principle only people who helped sink the well and who still possess a field near the-well,.have unlimited access to the well. Those who seil their land around the wel! lose the right of unlimiled access to the water. With Ihe purchase of land one buys access to the water al the saine time. How-ever, in accordance with pastoral custom, water cannot be refused to a passing herdsman with his herd, so anyone trekking wilh his berd may use the waler of the well, provided that hè asks permission and that hè wails until the people who "own" the wel! have finished drawing water for their animais. This right can be exercised for a week or so. If the passing herdsman wants to slay longer and exploit the pastures around Yaraama, he will have to ask permission of the owners of the well, or make a contract with one of Ihe owners of (hè land around the well to manure his field in return for water. However, only in case of great need the landowners will seek to engage someone else.

15 U shoulil howcver hè remarked Ihal the exclusion of women from tho inheril.mce of land is nol Jslamic at all.

Land tentire. lenitoriality, and ecological instahility The wiinde and the rainy season camp

31

Yct anolher land access arrangement can be fourni on the land of the biile (sg.

wiinde), and on the rainy season camps on the Seeno-Manngo. The most literal

translation of wiinde is "desertecl campsite". lts most essenlial feature is a hislory of use for corralling livestock. The use of such areas as a rest place créâtes rights, but what is appropriated is not the land, since Ihis can only be seized by putting the hoe to the soil. Rather Ihe appropriative act involves the mamire the animais deposil; tliis dung constitutes the basis of a spatial unit and serves to appropriate the space that is occupied by the herding family.

A deserted campsite is not a fixed point in space. Most camps have tin ce locations from which and to which the camp is moved every three to four ycars. After a couple of years the cattle is removed and the land is laken inio cultivation. The manurc left by the animais penned at night at the campsite serves to produce a big harvest on a relatively small plot. When soil ferlilily is exhausted one moves on to the next site.

Most (deserted) campsites are localed on the Seeno-Manngo. When the sandy soils are well fertilised they are ideal from an agriculiural point of view. They do nol silt up as the soils on the Ferro do. They demand much less labour than clayey soils of higher natura! ferlility. And the water rétention capacity is better than the clayey soils of the Ferro. At the time of the fieldwork the desert-ed campsiles were not cultivatdesert-ed because of lack of rainfall. The aniount of mamire in the soi! is so high that the crops "burn" because of drought periods during the rainy season.

The wiinde is, however, mucli more than a deserled camp sile. When in use il is the place of the rainy season seulement a place of intense social interac-tion, to which the Fulbe fee! some attachment. Given the fact Ihat a nurnber of families, from one or several lineages, occupy the same site in the rainy season, it can be inferred that within this unit coordinating links must be established, at least between the men. In order to make efficient use of the nianure deposiled on the present habitation site and the former sites, Ihe movement from camp site to camp site must be simultaneous.

When a deserted campsite is taken into cultivation, the rights of every par-ticipant have to be assessed anew. The reason for this is that (hè rainy season carnp is arranged on a circular plan, and fields to grow millet on a reclangular plan. Within the circle rectangular plots have to be laid out. In principle each family gels a field of almost the same size, unless there are reasons to reconsid-er each family's share.

Given the extremely flexible organization and the lack of leadership in Fulbe society it is logical that frequent changes in the composition of camps are bound (o occur. This poses the problem of what to do with the righls on Ihe mamire and the land of Ihose who left. In principle Ihe male head of Ihe extcnd-ed family, who was die first to seltle on a non-occupiextcnd-ed sile, is ils owner. All the others who followed h im and wanted to join h i m had to seek his permis-sion. Räch new occupant is assigned his own sector. In a rainy season camp Ihe huls of Ihe various families occupying Ihe site are arranged in cluslcis, so thaï each family deposits the manure on his own seclor. Righls in the land of a

wiinde are, however, not for eternily, allhoiigh everyone would have liked tis to

32 Van Dijk rainy season seulement lasls, plus the time needed for the dung lo décompose. After that period the rights and the image of the family in (lie seulement faded away, as it were.

The reason for such dwindling of rights is this. When someone leaves his part of the wiinde uncultivated the manure décomposes and soil fertility dé-clines. As a conséquence the interest of someone who h as left also déclines with time, for what hè added to the soil withers away. When aflcr a number of years his interest lias diminished significantly, someone else may ask to setllc on his sector, and will subsequently acquire rights in Ihe soil as bis share in Ihe manure rises.

Conflicts over campsiles like this are rare. There are relatively few people involved. As long as there is no source of water, permanent .seulement during the dry season is nol possible. If dissent occurs, leaving the camp is relatively easy and happened regularly in Üie past. There are few formal lies binding the people together in this unit, except for brothers, and fathers and sous, who are expecled fo co-operale in olher domains as well. The departure of people does nol necessarily indicate that they do no longer feel at home. Rcasons for joining and leaving a rainy season camp or an extended family may bc an entirely pri-vate affair. What is relevant from the perspective of land lenure is Ihe almosl automatic diminution after one's departure of rights in the soil fertilily created throtigh animal droppings. So it is not the soil that is approprialed but the producls of the animais, and this explains also Ihe Iransient nature of rights on the wiinde.

The suppression nf ctiltivation on bush fields

So far rights on land that is cultivaled at present have been discusscd, i.e. the claims on the land are constantly made effective by niaking Ihe soil pioduclive. liiere remains, however, one category of land, on which claims are laid, i.e. the bush fields, which had to be abandoned during the second phase of the intensi-fication of land use. As was shown many grotips had been ctiltivating bush fields everywhere on the Ferro and the Seeno-Manngo for a considérable pe-riod before the intensification of land use gained momentum. The bush fields of the Riimaybe were large and were dispersed over a vast area. Productivity of these fields was low and feil quickly after one year of exploitation, hecausc of Ihe poor sandy soils. Their dispersai made exploitation of the aren as paslure impossible. The Riimaybe cultivators clearcd the majorily of these fields aflcr thcir libération and as such the appropriation of these lands was an important marker of their newly acquired independence.

During the 1950s, however, a number of Fulbe pastoralisls cslnhlishcd Iheir campsites in bctwccn these bush fields. The présence of large hcrds made the protection of Ihe bush fields vcry problemnlic. The cullivalion of dcscilcd camp sites allowed Ihe concenlration of crops inlo higlily productive small fields. Moreover, the pastoralists coordinated the migratory movements of Iheir setlle-nienls on the campsites, so that never more Ihan a few per cent of Ihe total land arca was ctillivated. In this way (hey were able to combine cereal cullivalion and (lic grazing of herds at the same time, whereas Ihe Riimaybe cultivalors

Land tem<re, terriforialify, and ccohgical inxlahilily 33

had to face an increasing pressure of cattle and were pushcd lo givc un thoir fields.

Understandably tliey were nol prepared to leave these fields uiuler the pres-sure of Ihe herds of the pastoralists. The latter, however, outnumbered Ihe Riimaybe cultivators by far and were politically dominant at the time. Tbcy claimed Üiat these fields had been worked by their anccstors, and thaï they wanted to reconvert (hem to paslure. This claim was reptidialed by Ihe Rii-maybe, bul Üiey were powerless against the obduracy of Üie pasloralists. When damages to crops occurrèd the cultivators were nol compensaled. Herds were driven into their fields. Complainte to the Fulbe chief of Booni, who was nominally in charge of these affairs, led to nothing, for the pasloralists bribcd him with catlle. In Üie end the Riimaybe cultivalors had to give up Iheir fields in Üie bush.

Several conflicts arose over the passage of herds from paslurcs to water re-sources. To solve these conflicts cattle routes were opened by the administra-tion over fields, while other fields were allowed to reniain in existence. Among Ihose who lost their bush fields somc had to give up living in Scrma, if Ihcy were not able to acquire land in loan or ownership elsewhcre.

In this way the Fulbe pastoralists stopped agricullural encroachment of cul-tivators on the posture lands surrounding Serma and were even able lo push them back. They look these sites into production for precisely the same reason as the Riimaybe cleared bush-fields, i.e. to produce cereals. Only the Jallube did so in an entirely different manner. The Riimaybe praclised a bush-fallow System, in which the soil was worked for a couple of years and Ihen abandoned. The Jallube invested in the land by bringing in the manure of Iheir livcsiock. Until today the campsites make Ihe cuilivation of Ihe bush fields impossible. Ilowcver, Ihere is no Fulbe pastoralist who would dare to occupy such a sup-pressed bush field. They retnain the property of thcir original owneis or Iheir lieirs. even while they are not in use.

Control over range

Control over pastures is arranged on a different basis than control over agricul-tural land or over water. In principle accessjo range is open forevcry one. K is difficult (o get to know what territory mcans in Serma. The hcrdsrncn have no idea whcre to locale Ihe boundary of the grazing tcrrilory of Scrma. Wc sup-pose the strcngth of (ciritorial claims is relative lo Ihe distance from Scima, and the distance to the next settlemcnts, Felcsambo in the cast and Pclcgudu in tbc west. When everylhing proceeds without pröbtcms Ihe nolion of lerriloiy is liardly discernible. Only when calamities occur, for cxamplc droughl, does Ihe concepl of (crrilorinlily bccome relevant. Thcn a largei Ihan noininl arca is claimed, bccausc of the nccds of Ihe animais.

34 Van Di/k Land tenwe, teiritoriality, and ecological instabihty

large. Moreover, tlie capricious nature of the ecology places a premium on mobiiity and tlie mtitual granling of access (o eacli olhers paslures.

During Ihe rainy season there is harclly any conlrol on gra?ing. The multi-tude of water resources in the form of small ponds precludes Ihis. At fiist glance there are hardly any rules governing access to open water in pools. Rveryone is free to use the ponds (hat appear in the rainy season as long as Ihey last. No righls of access are dcfined. Outsiders cannol be excluded, allhough thèse ponds are mostly located in thé neighbourhood of selllements. The inhab-itants of thèse communities hâve, however, riglits of priority of access. In prac-tice thèse riglits are hardly put in effect, becau.se for most of the time during tlie growing season there is more than sufficient opportnnity to water the herds al-most everywhere in thé bush. Soon after tlie rainy season, often as early as Sep-tember, water has disappeared, so that outsiders are not able to remain in tlie area for more than a couple of months. When the water situation becomes more tight, there needs to be some coordination, to prevent all heids entering the pool in Debere at the same time. One after the other the herds are watered at the pool. The aclual work of leading the herd back and forth from the pool is done by tlie young herdsmen, while their elders supervise their work from be-neatli the rcmaining tall trees, sitling and chatting, and only intervening when someüiing goes wrong.

Tlie only limitation on the movements of herds during the rainy season are the fields, from which the herds have to be kept at some distance, in order to prevent damage. In the past the herds were away for the annual sait cure at the sait licks north of Booni for an extensive period during the growing season. Livestock from other herd owners in the Hayre used the pastures of Serma to some extent, when they were on their way to the sait cure, or when pastures around their own settlemenls were insufficient. This was welcomed by the in-habitants of Serma, as long as these herds did not bring any diseases, because Üiey would have tlie saine riglits of access on the pastures around other settle-ments when their own range was of low quality. Livestock from outside the Hayre was rarely seen, and this had been so ever since the 19th Century.

Understandably control over grazing is much more strict in the dry season. Access to the grazing ilself is still free. In the past as well as in the present wa-ter has been the most limited resource, at least in Senna. Most of the foreign herds move to the Inner Delta, where beller grazing may be obtained Given the location of Ihe wells and water reservoirs in Serma, and the well in Bunndu Jaabi, and tlie journey a herd of cattle is able lo niake in a day's time, il is easy to see that the Seeno-Manngo and the Ferro could only partially be exploited during the dry season. To compensate for this, the herdsmen fed their animais on the fresh re-growth of perennial grasses after burning. In this way the ani-mais were able to graze for two days on end without water, as long as the tem-pérature was not too high. As long as the waler resources were conlrolled by the inhabitants of Serma themselves, gra/ing territories were only a Iheoretical entity. Dry season grazing resourres were so vast that there weie no territorial disputes.

§

Teiriiotial d'npute

This situation changée) fundainenlally when Ihe French colonial governmcnt intioduced a new element in the control over territory by the developmenl of tiew water resources that enablecl outsiders to encroach on village terrilory In 1957 the colonial administration starled to drill 7 boreholes spread over the Seeno-Manngo, which were equipped wilh windinills to pump tip waler. 'I hese boreholes were meant to open up the paslures of Ihe Seen-Manngo, which the administration considered under-exploited (Chapuis et al. 1972). Two of these boreholes were positioneel in the neighbourhood of Serma. Botli boreholes be-came the subject of conflicts with clear territorial and politica! dimensions. The ecological conditions do not seem to play a prominent rôle, but will re-appear later on in the 1970s and 1980s.

In the first case (the ßunndu Hinndu borehole, see map 2), some Fulbe herdsmen were given permission by colonial administrators to make fields aiound the well since they were using Ihe area for the graring of liveslock at the time. This enraged the people of Serma who were not involved in thé déci-sion and they rebelled since they viewed Üie area äs their territory. After pro-longed disputes involving attempts at negotiation by bolh Ihe Fulbe chief of Booni and thé more distant colonial administration, the matter was only resolv-ed when the windmill broke down and could not be repairresolv-ed The intrusive herdsmen were forced to leave their fields and to lake their herds to pasltire elsewhere. The second case, however, remained very contentions during the pe-riod of field research and only a few people were prepared to discuss if for fear of a récurrent flare-up of trouble. The second borehole was located at Wiinde Kokulu (see map 2), between the settlements of Felesambo, Bunndu Jaabi and Lennga, which were dominated by the Tamankoobe and Seedoobe lineages. lts location had been used in the past by both groups for the cultivation of millet without conflict, but the borehole precipitaled trouble when the Seedoobe foiind out that a few Tamankoobe had starled to sow fields at the wiinde with-out Consulting them. Since Üiey had planned to graze their cattle there they told the Tamankoobe cultivators to leave; bul given that these cultivators had no other millel fields planted Ihat year, they refused to move. The herds of Ihe Seedoobe subsequently damaged Üiese millet fields and Ihe leaders of the two lineages carne to blows and insults over tlie issue. Several décisions by neigh-bouring chiefs did little to conlrol the escalating anger and violence, and the dispute settled into an ongoing feud between tlie two parties. Tlie outcome was that the contested area has not been cullivated after the conflict had been sup-pressed.

36 _{Van Dijk}

decision-making authority over the wiinde. Tliis is made plain by (hoir refusai to withdraw their herds from the area around of the wiinde.

Both of (hese conflicts were the result of the unclear situation with respect to access to government-owned water resources. In the first instance it secms tliat tlie pasloralists from the east won the politica! game over diese water re-sources by playing the chief against the administration. The nulhority of the chief does not extend over Ihem and the administration lias (o take over. 'l hè colonial administration assunied ultimate decision-making aiithority over the management of grazing lerritories through drilling the boreholes. With the control over the boreholes administrators coukl décide who was allowed to make «se of the range and the land around the boreholes. Tliis clearly interfcred with existing territoria! claims on the pastures areas, such as those of Serma in case of Bunndu Hinndu and the two lineage in tlie case of Wiinde Kokulu. The colonial government crealed insecurity with respect to control over resources in this way, and at the same time created, consciously or unconsciously, the con-ditions in which more interventions, i.e. the allocation of user rights on (hc land dépendent on the borehole, were necessary.

Ecolngical calantity

Over the last few decades the above situation with respect to the organization of grazing has been forced (o change cotnpletely. The perennial grasses, which were the cornerstone of dry season grazing on tlie Seeno-Manngo have disap-peared because of drought.16 The remaining annua! grasses do not re-grow when burnt, and in case fire occurs the range is lost unti! the next rainy season. These végétation changes have dirninished the value of grazing enormotisly during the dry season. The trek to the sait licks has become less important. Only the few remaining large herds départ for tlie annual sait cure. The small herds stay in Serma, and are given sait which is bought on the market.

In 1975, a couple of years after the first major Sahetian drought (197!-1973) a special programme, called Üie Opéiation de Développement de

l'file-vage dans la région de Mopti (ODEM),17 was starled by the liveslock service and financed by the World Bank, to help alleviate the problems in the pastoral sector. This programme lias to be seen in a larger context. Tlie rationale behind tlie attcmpts to develop the research area is that the Inner Delta of the Niger (see map 1), which produces the most valuable dry season pastures in Central Mali is hopelessly overcrowdcd due to the decreasing rainfall and the décline of the flooded area in the rainy season. By developing dry lands such as the Seeno-Manngo it is hoped that pressure on (lic Inner Delta can bc diminishcd.

W i t h i n the franicwork of Ihis programme government wclls were sunk in ncighbotiring scttlemcnls of Serma in Ihc border area belwccn Ferm and 16 It was recently reportée! that AnHropvgon gayanus, the most important percnnial prass in this area, is rccslalilishing itself aflersevcral ycars of ahundanl rainfiill (Douma cl ni. I99S).

17 ODIîM was suspcixled in 1991. In the research area ils drvrlonment arlivilies were laken pvrr In m NCO b.ised in Doucnl7a, Ilie rltslricl rapil.il, finançai by Ihc CCC'E

l Mild leintrc. lenitonalily, and ecolagical instabil! ty 37

Seeno-Manngo. An altempt to drill a borehole at (lie site of the water reservoirs in Serma failed. The inhabitants of Serma had opposed this, for Ihey did nol want foreign herds on their pastures. The geology helped (hem: the bil mellcd in an impénétrable layer and no water could be found. Another intervention in the neighbourhood of Serma was Üie organization of a grazing reserve around a borehole on the Seeno-Manngo. This grazing reserve, 12,000 ha, was located on the lerritory of Serma and Felesambo. Access to these pastures was prohib-ilecl in tlie rainy season, and for most of the dry season. Use of this scheine was for the inhabilants of nearby villages only, so lhat the pasltires were not lost for grazing. The scheine was managed by a pastoral association licaded by the chief of the Fulbe and his entourage, who résides in Booni. The inhabitants of j Serma, however, were no longer in conlrol over this pari of Iheir territory, as j became clear during the second major Sahelian drought in 1983-1985 (see also

\ Van Dijk and De Bruijn 1995).

f According to an annual repoit of ODIïM the badly timed rainfall in 1983, only allowed a satisfactory growlli of the pastures in Ihc Mcma (West of Ihc In-ner Delta of the Niger), and on the Seeno-Manngo. From the tnonlh of A p i i l 1984 on there were heavy concentrations of animais aiound the boreholes lead-ing to critical overgrazlead-ing and the exhaustion of pastures (ODEM 1984:8). And Ihis was only 1983. In 1984 ihe rains failed complelely and pastures were bad all over the Sahel and Sahara. During the dry season of 1985 herds from the North flooded the Seeno-Manngo, and galhered around (hè governmenl wells. In an atmosphère of polilical coercion and corruption free access was allowed lo outside herdsmen to waler Iheir animais a( (hè borehole of the grazing man-agement project. Tlie abandoned borehole of Bunndu Ilinndu. also posilioncd wilhin the grazing scheme was opened for a Tuareg chief, wilh thousands of head of caille. As a result of having lo share their pastures with so many olhers (lie inhabitants of Serma lost about 75% of their cattlc in that year. This led (o tlie collapse of the local economy, enormous émigration and widespread im-povciishmcnt(De Bruijn 1994, De Bniijn and Van Dijk 1992, 1994)

Not all of this livestock dicd; an unknown portion of these lost catllc have found their way to new owners among agriculltiralists, urhan daders, and civil servants. These new owners bought them cheap, and managed to keep pari of their new acquisitions alive.1" When (hè drought was over and pastures rc-covcred the new owners soughl ways to manage their new assels. They cm-ployed young herdsmen, and penelraled into new arcas. Remole areas such as Üie Hayre and the Seeno-Manngo, where paslure was good and liveslock densi-ty very low, were soon discovercd. During the growing season these herdsmen do nol nced lo negoliate access to pastures and water. As a result minicmiis Iicrds now roam over Ihc Seeno-Manngo in the rainy senson. To give an exwn-plc: wlicn we wcrc celcbraling the Islamic slnuj'htcr fpasl in Wmo Boppo on (lic 23rd of hinc in 1991, more Ihan 30 herds passcd Ihc camp on (hal day alone. Each of Ihese licrds was accompanied by at leasl (wo herdsmen. so wc

38 Van Dijk

may estimate their size al 200-250 heacl of caille each. The people <?ƒ Wuro Roggo were baffled: never in their life Iiad they seen anything like this. Our host got so angry that he-ordered the Herdsmen to take another route, for lliey were damaging tlie hutsriWhen-óne of-the herdsmen asked for wiitcr hè refused lo give il, whirh is omrof Ihe greatcsf offenses one can Ihink of in Ihis area of waler shortage.

During the rainy season a number of these herds remain in the area around Senna, and cause a great deal of tension in the village. 'Plie herdsmen exploit the pastures of Serma'with their herds, and in this way diminish the already smaller resource base for the dry season. They also neglect the boundaries of the grazing reserve (cf: Shanniugaratnam et al. 1993:22). There are mimerons cases of damage to fièlds. Tlie-'hérdsmen, who spend Iheir days in Debere, dominate the scène completely. They possess cassette players, with which they dislract the youths of Serma from their cultivation tasks (so the elders com-plained). They are the main customers of the Riimaybe wotnen, buying their foocl in Debere in the form of cous-cous, which is prepared from millet. During the fieldwork this added to the scarcity and the high priées of food in the vil-lage. Contagious diseases in the herds lead to incidents. Somelimes herdsmen were beaten up by the youths of Serma when they refuse to leave the area. When they complained to the police in Booni, people were arrested and delain-ed in Booni for some time. The herdsmen also courtdelain-ed young women in Serma, which created another source of mirest and jealousy.

The strugglefor control

It is now understandable why the inhabitants of Serma want to stay in control over territory.19 Over the past few decades they have gradually been disowned, nol only of their cattle, but also of their pastures. Tliis struggle goes so far lhat even the open water resources, which were nominally free during the rainy season are drawn into it, as the following conflict shows.

In 1987 part of the main pond in Serma was deepenecl by ODF.M, so that the availability of open water was extended from Oclober lo Jamiary, reducing the work load of watering the cattle in this period. An incident (hat occurred in October 1991 showed that such an intervention in control over territory may create ambiguity in thé définition of rights of access. Even open waler is nol a free for all resource as it appeared to be in the rainy season of 1991.

In the neighbouring cattle camp, Petegudu, liiere was an oulbreak of bo-vine pleuropneumonia, probably brought by a herd from the Inner Della of Ihe Niger which had been passing the grovving season on the Seeno-Manngo and Ferro. Because of the outbreak the herds of Petegudu had to be isolated from Ihe herds of Serma and Felesambo in the east. Normally all these herds were walered at the Serma pond, because the ponds at Petegudu and Felesambo were already dry at this time of the year. In agreement wilh the velerinary service it 19 This is fiirlher lestified by Ihe fact Ihnt rccently Ihe inhabitants of Serma linve even refusctl the help of ai) NGO, which look over the aclivilics of ODIiM in this aiea, foi (lie repair of Ihcir own well al Yaraama. ïliey sus|>ecl Ihat outsiders will pel conlrol over Ihe waler, even if Ihey pay part of the repair of Ihe well.

Land lenure, lenitonality, anti ecological instabil!ly 39

was decided that the herds of Pelegudu should be assigned a pond f tirilier away on the Ferro, so that (lie water at Serma wotild not be containinated. This did nol work at all. A combination of resentment aniong the herdsmen of Pelegudu at their bad luck and at being denied access to an open access (govermnental) water resource, with fcelings of jealousy, and the absence of a structure lo ré-solve conflicts in Üiis kind of situations, led to a very lense situation, in whicli lerritoriality played an important rôle.

A couple of days after the animais of Petegudu were banished to Ihe pond on the Ferro,20 a herd from Petegudu was driven into the pond by a herdsman, in order to provoke conflict. The elders held back the young herdsmen, who wanted to attack, and told Üie herdsman lo go away. They promised nol to beat him up, hut instead to scare the wits out of his animais if hè ever returned. The lader measure would be much more sévère, because the whole herd would ihcn disperse, and hè would probably lose half of it, because of thefl and thirsl, be-fore hè found Üiem all. A week or so later, another incident happened. Despite the guards, who were probably sleeping or absent, the herdsmen of Petegudu somehow managed to drive a sick buil into the pond. The young men wanted to slaughter it out of revenge. The elders and common sensé once again prevented escalation of the conflict, for such an action would surely hâve brought a counter attack, more fighting and finally thé intervention of tiie police from Booni. In that case everyone would have had to pay dearly, because these offi-cials would not have left Serma and Petegudu in peace before they and other authorities had pocketed large sums of money.21

This example shows that even ponds are not just open access resources. In principle everyone has the right to water his cattle in Ihe pond. When, however, a calamity occurs, the people residing in the neighbourhood have preferential access to it. Those from outside were excluded. Even if the illness had been among the cattle of Serma only, the animais of Petegudu and Fetesambo would have been assigned another pond and not the reverse. An incident like this brings territorial groups into being almost at once, albeit on a lemporary basis, but wilh a high degree of internai cohésion and motivation. All Üie young men, even the Riimaybe, were prepared to f i glit against the herdsmen of Petegudu, while Ihey were of course friends and comrades in daily life. Equally remarka-ble was the fact that the elders were aremarka-ble to hold back the young men from tak-ing revenge, despite the fact that they cannot always command their sons. Probably nobody feit sufficiently protected by the livestock service, and sure enotigh about Uie status of the pond to take a more firm posture, and lo resorl to violence to résolve the conflict.

Discussion

As has becotne clear, access to agricultural land and grazing resouices was or-ganized in a very flexible manner. In the 1950s (he nieaning and value of Lind, 20 Which was located much farther away, putting extra slrain on the animais, wliich wcre already sick. •

40 Von Dijk pasture, and waler points changed in a short span of lime, due to thé sinking of a well, the drilling of a borehole, and abundant rainfall. Instead of expanding the cultivaled area around the village, cultivation was reslricted (o Ihe confines of permanent water points and the sites of rainy season settlements, under the pressure of growing numbers of livestock. Probably cultivated arca dccreased in Ihis period, rallier lhan increased as in olher régions (cf. Gallais 1975, Mar-chai 1983). The orbit of transhumance changed considerably, so as to ailow thé boosting of soil ferlility of the village fields, which came under a permanent cultivation régime.

The political and ecological conditions of the time were partictilarly fa-vourable for this development. The rains in thé 1950s and 1960s were very abundant in comparison with thé présent. The millet grew very well on thc ferlilised sandy soils of thé Seeno-Manngo, particularly on thc biile. so thaï people even produccd surpluses of millet, which they sold, something which is unthinkable loday. The cattle were grazed on (lie re-growlh of burned perennial grasses on thé Seeno-Manngo, which made the paslures far more productive tlian at présent.

The most remarkable feature of land tenure and territorial rules scen in this historica! perspective is their indelerminacy and flexibility. There is nol a sys-tematic body of rules and procédures governing access (o village resources. If we analyze all kinds of events we find thaï Ihe présent légal framework consists of an amalgam of customary, Islamic and modem éléments tinusual for this part of Mali (see for comparison Gallais 1975, Gallais and Boudel 1979, Kinlz and Traoré 1993). However, looking at history we would hâve to add lhat Ihis framework is only a temporal construclion, probably only completely valid in this communily. Il is a plural framework, nol only plural in a légal sensé, but also in a spatial, a social and a historica! sensé. Différent plural légal frame-works apply to varions constellations of social groups, resources under rules of tenure and thé surrounding "open access" type of resources. Their dynamics cannot be thé resuit of Ihe dynamics of ils légal sources. The rules arc nol dc-libcralcly designcd lo restilt in a cohérent and consistent légal framework lo regulale social relations of individuals and groups wilh respecl to resources in a satisfactory manner. They are mainly an expression of how people solved prob-lems and regulated access to resources in llie past. In this way Ihey very inuch reflect thé considérations and conditions of lhat spécifie period. What we ob-serve in thé présent, i.e. Ihe allocation of land for varions purposes and the amazing variety in land tenure arrangements is based on lineage ideology, ts-lamic inheritance rules, free market principles and governmenl law, wilhin thé frame of the present power structure and daily concerns for survival of thc in-habilants of Serma. This is a property which emerged out of Ihc décisions, nc-gotialions and conflicls of individuals and groups who Iricd lo confond wilh po-litical change, ecological hnzard, and thé need for mobilily.

So, thc land tenure situation in Serma may better be analyzed as a "nego-tialcd order" (Bcrry 1988), which is closely conneclcd to Ihe polilical hierarchy as was shown by sevcral evenls. The ncgoliations and procédures for the trans-fer and allocation of access rights lo resources are neccssary lo résolve Ihc ten-sions arising out of Ihc simultaneous practice of herding and cropping, Ihc co-ordination, liniiiip, and spacing of ctillivation itself, and thc rontingencies

Land tcnttre, lernlorialiiy, and ecological inslahility 4l fields and their number of livestock the Jallube have different inlerests in farm-ing and herdfarm-ing, and hâve more, or less, to fear from wanderfarm-ing herds. Wilh thé development of new water resources the need for negoliations and coordination increased, which led to a growing complexily of land tenure arrangement.

In some ways thé allocation of land appears to be nolhing more Ihan thé exercise of power. The varions conflicls over land arc inevilnbly lost by Ihose lower on thé social hierarchy, despite thé facl lhat such people play a crucial rôle in (lie labour force of thé comnuinity. In Ihis way poor Fulbe herdsmen could be pressured into selling their land, and Rtimaybe cultivalors coiild be forced lo give up their rights to bush fields. In this exercise of power, définitions of access rights are often manipulated to achieve objective goals. This was the case when a younger meinber of a family was able to alienate "privale rights" lo land to and Islamic cleric favoured by those in power, des-pite his elder brother's opposition. A sitnilar situation was oblaincd when thé land around the Yaraama well, which was officially lineage properly, was cov-erlly alienated by lineage members. Under conditions of extreme environmcn-lal stress, conflict over lerritory and waler lias potenlially mucli graver consé-quences, as larger power blocks become involved - more people, entire linc-ages, and political loyallies. Involvement of the administration lias become un-avoidable. In the course of 40 years the administration, and ils représentatives such as the ODtîM and olher services, have oblaiiied a firm grip on the situa-tion. The people of Serma are no longer able to conlrol who is to use Ihcir pas-tuies in llie rainy season and a considérable period (hereafler. In thc dry scason Ihey are officially co-managers of thé grazing reserve logelhcr wilh ODfiM. In rcaiity they have to deal with a permanent unconlrollablc source of walcr, which very much Ihrealens to make a définitive end to village autonomy wilh respect to natural resource management.

The basis for this loss of control was laid in thé colonial period. The chief as central aulhority over land and pasture lost his power lo the colonial admin-istration. However, llie basic hieiarchical sluicture remained intact. Al picsrnl the now independent govcrnmenl uses Ihis hierarchy lo manage the giiizing scheine, nominally a so-called participatory development organisation, but in realily almost a privale enterprise of the "traditional" chief. Now thé Fulbe pasloralists, who fought territorial disputes over the boreholes in Ihe 1950s, hold back llieir sons who wanl to défend Iheir water in Ihe pond against contamination by sick animais.