Zambezia (1997), XXIV (ii).RURAL-URBAN LINKAGES: MASVINGO'S DOUBLE-ROOTED FEMALE HEADS OF HOUSEHOLDSV. N. MUZVIDZIWADepartment of Sociology, University of ZimbabweAbstractThis article results from research on female heads of households in a smallZimbabwean city. Urban-rural linkages constituted an important, on-goingsurvival strategy for women in the research sample. Despite my respondents'desire to stay and to source a living in town, most maintained rural linkagesin order to cope with various problems in town. Without a foot in the ruralarea, most women would not have been able to pursue their desired objectiveof being permanently urban. Although the women were urban-oriented, theyused kin networks in ways that maximised their chances of surviving in town,against the structural constraints imposed by central and local-levelbureaucracies. The article explores ways in which the women pursued thestrategy.THE STUDY ON which this article is based was carried out in Masvingo town,a provincial capital with a population of 52 000 according to the 1992Census (CSO 1993, 13). Masvingo lies in the southern part of Zimbabwe,ravaged by repeated droughts in the decade extending from the mid-1980sto the mid-1990s. Masvingo is Zimbabwe's oldest town, founded in August1890. At one time in the 1890s, it was also the largest colonial settlement.It has twice changed names: Fort Victoria was the town's name up to 1982,then for a brief period it was Nyanda, a name rejected by the inhabitants ofthe town. From 1982, the town assumed the name Masvingo, and is now aregional economic centre as well as the provincial capital.According to the 1992 census, only 33,7% of Masvingo women aged 15to 64 were economically active (CSO, 1993,59). Given the high participationof women in operations in the informal sector that were not enumerated inthe census, it is not surprising that the census produced a very low femaleparticipation rate in the urban economy. The town had an officialunemployment rate of 25%.Masvingo had its economic roots in the free capitalist market thatexisted prior to the independence of Zimbabwe in 1980. This waspredominantly a male-oriented and male-controlled economy. The urbaneconomy supported a migrant male population. Africans were consideredas permanently rural. Job opportunities for women were very rare. Womenworked in the domestic sector and service jobs such as nursing andteaching. Few of these could be absorbed into the two primary schools97I 98 RURAL-URBAN LINKAGESI and one secondary school (in Mucheke, then the town's only high densityI suburb), one clinic and a limited number of private medical practices. InŁ the informal sector, beer-brewing, prostitution, and marketing of freshI produce, together with domestic service and child care, were activitiesŁ characteristic of town women.I This study is based on urban anthropological fieldwork over a periodŁ of 14 months from early November 1994 to the end of December 1995. TheŁ results presented in this article comprise part of a larger project. I usedŁ multiple methods in order to gain greater insights of struggles by myI Masvingo respondents not only to cope with poverty but at times to moveI out of it. Apart from informal interviews and observations in the community,Ł I selected for in-depth and intensive interviews, 58 female heads ofhouseholds (eight married bread-winners were later on removed from theanalysis as they were not unequivocally female heads of households), andI also interviewed 31 officials.The female heads of households who participated in the in-depthinterviews from January to May were found by 'snow-ball sampling'. Thesnow-ball sample drew all its study population from Rujeko and Mucheke.These were Masvingo's only low-income or high-density areas, where thebulk of the town's population lived.Kin are quite important at the time and point of entry into an urbanarea. Only 6% of the research participants were born in Masvingo town.The rest were migrants to the city. For my respondents, urban-rurallinkages constituted an important on-going survival strategy. I argue inthis article that despite their desire to stay and to source a living in town,most respondents maintained rural linkages. Without a foot in the ruralarea, most women would not have been able to pursue their desired,permanently urban objectives. Although female migrants were urban-Ioriented, they used kinship networks in ways that maximised their chancesof surviving in the city, against the structural constraints imposed bycentral and local bureaucracies. This article explores the ways in whichfemale heads were turning to rural-urban linkages as a survival strategythat enabled them to cope with Masvingo town life.There were some for whom the urban strategy was dominant. Themajority (68% of my sample) maintained the double-rooted strategy: theseoften had an ultimate aim to return to their rural villages at a later point inlife. Irrespective of the dominant survival strategy, urban-rural interactionsdid accomplish the desired goals.Basic data on Masvingo research participantsDuring data analysis I developed an analytical category to determine thesocio-economic status of respondents. Initially I had adopted animpressionistic division based on total income, but this provedV. N. MUZVIDZIWA99unsatisfactory as a basis for categorising the behaviour of my respondents.I subsequently developed criteria that provided a classification ofrespondents into four groups, namely: 'burnt-out', 'hanging-on', 'copingwith poverty', and 'climbing out of poverty'. These categories in manyways revealed that Cheater's and Jackson's (1994, 143) 'straddlingstrategies' in the rural context were being replicated in the urban contextby my Masvingo respondents. Table 1 indicates the composite variablesaggregated to produce scores that led to an objective definition of thesocio-economic status of participants. It also shows the value assigned tothe response in each score category. Finally it demarcates divisions onthe basis of attaining a certain number of scores by a respondent.Table 1THE POVERTY INDEX VARIABLESVariableEducation (self)HousingNumber of roomsPersons per roomClothingNumber of meals per dayMeat intake (last wk)Number of dependentsTotal incomeRemittances per annumIncome minus expenditureValue = 1primary or lesslodgershared5+only donated104+$>6000negativeValue = 2Forms 1-4rentalone3-4only 2nd hand21-4 times2-3601-9991-3break-evenValue = 3post-secondaryownertwo+2some new35+ times1 or less1000+4+positive'burnt out' = total score < 14: those who could hardly sustain their urbanexistence'hanging on' = total score of 15-20: in difficult circumstances but struggled tostay in town'coping' = total score of 21-26: could balance the household budget buthad no savings'climbing out' = total score >27: had a healthy domestic budget, had savingsand investmentsTable 2 presents the two distributions of respondents on the basis ofthe composite score, compared to my initial hunches based on incomeonly. Those with an income below $300 per month I had considered the'burnt-out'; the 'hanging-on' had an income between $300 and $600; the'coping' ones had an income range of $601 to $999. It appears that the100 RURAL-URBAN LINKAGESmore complex criteria captured the real situation of my respondents. Veryfew were in the 'climbing-out' group. The majority struggled to exist in thecity and were thus in the 'hanging-on' or 'coping' group.Table 2SOCIO-ECONOMIC STATUS OF MASVINGO RESPONDENTSClimbing-outCopingHanging-onBurnt-outTotalUse of Income OnlyNumber171410950%34282018100ObjectiveNumber81720450Criteria%1634408100Table 3 shows that the research participants comprised five maritalstatus groups.Table 3DISTRIBUTION OF FEMALE HEADS OF HOUSEHOLDS BY MARITALSTATUSDivorcedNever-married single mothersNever-married single personsWidowedMapoto (informed union)TotalTotal Number191177650%3822141412100I noticed in only one case a discrepancy between the participant'sself-classification of her marital status and that of the person who referredthe person to me. One never-married, single mother identified herself as amarried woman. She had spent at least nine months at her would-be in-laws' rural homestead. Her prospective husband, a soldier in the nationalarmy, never paid bride wealth, though he did pay her a nominal amount of$100 per month for maintenance. At the time of the initial interview, shewas staying with her parents in Masvingo, having rejoined her family twomonths prior to the interview, but only in October 1995, did this researchV. N. MUZVIDZIWA 101participant indicate to me that her prospective marriage had failed tomaterialise. In a situation as in Zimbabwe, where marriage is idealised asnear universal, a woman must be seen to have tried before giving upattempts at marriage.Table 4 shows the distribution of the respondents into three agegroups. It is not surprising that there were more women aged 26 to 41 thanin the other brackets. This represents a phase in women's lives when theywere more likely to seek gainful work in town in order to support themselvesand their dependents. During this phase women try to invest in theirchildren's future.Table 4DISTRIBUTION OF FEMALE HEADS OF HOUSEHOLDS BY AGE-GROUPAge Groups Total %18-25 8 1626-41 34 6842+ 8 16Total 50 100Table 5 shows that the majority in my sample had very little education.The percentage of women with a primary only or less educational levelwas high, only 8% had completed a post-secondary professionalqualification. Hence the majority of the women in my sample could noteasily sell their labour in the formal market. They relied on their ownresourcefulness and entrepreneurial skills in order to earn a living.Table 5DISTRIBUTION OF FEMALE HEADS OF HOUSEHOLDS BY EDUCATIONALLEVELTotal %64288100Table 6 shows the distribution of the respondents by source of income.My sample had a high rate of participation of respondents in multiplePrimarySecondaryPost-secondary trainingTotal3214450102 RURAL-URBAN LINKAGESincome-earning activities. The attempt to ward off poverty by engaging inmultiple income-earning opportunities is what Cheater and Jackson (1994,143) refer to as 'straddling' strategies. Poorer respondents relied more onmultiple incomes. One woman had five sources of income. She was across-border trader, sold vegetables and tomatoes, rented out a room,was a mupedzanhamo operator (second-hand clothes dealer) and aretrenchee on a pension. She was one of the few successful ones whoengaged in multiple income sources. It is no surprise, given the location ofthe research, that the majority of the research participants, in Table 7identified themselves as Karanga.Table 6SOURCES OF INCOMEMain Source of Income Total %Food vending 19 38Cross-border trade 13 26Formal job 6 12Prostitution 4 8Carrier bags 3 6Hairdressers 2 4Others 3 6Total 50 100Other = one in each of the following categories: a traditional beer brewer, arentier and a recipient of maintenance money.Table 7RESPONDENTS' ETHNIC BACKGROUNDEthnic Group Total %Karanga Shona 35 70Non-Karanga Shona 6 12Ndebele 4 8Other 5 10Total 50 100Non-Karanga Shona included 3 Zezuru, 2 Manyika and 1 Korekore. Other categoryhad 2 Shangaans and 3 Malawians.Table 8 shows that most of my respondents had a rural backgroundand originated from rural districts surrounding Masvingo. This highincidence of persons with a rural background partly explains the reasonsV. N. MUZVIDZIWA . 103why many of the respondents chose to have a foot both in town and thevillage. The proximity of the women's rural homes to Masvingo town madeit easy for most of them to maintain rural links through remittances, visitsor other regular contacts with those in their villages. Not only was itconvenient for most of the participants to maintain one foot in Masvingoand another in the village, but it was a part of their survival strategies. Forinstance splitting the household into the urban and rural componentreduces costs of urban reproduction. Keeping the rural links active is likean insurance against unforeseen future problems.Table 8RESPONDENTS' ORIGINAL HOME BACKGROUNDS Š PLACE OF BIRTHTotal %Rural Masvingo 33 66Rural non-Masvingo 6 12Masvingo urban 3 6Urban non-Masvingo 7 14Born in South Africa 1 2Total 50 100Unstable marital unions meant a reduced risk of pregnancy or a greaterwillingness to prevent pregnancy on the part of the women in the study.The fertility rate for the sample is well below the provincial average of 6,7children per woman according to the 1992 census: my respondentsaveraged 2,2 children per woman. Even if we exclude the seven childlesswomen, the average number of children was only 2,6 per woman. Theseresults paralleled Finan and Henderson's (1988, 94) observations thatsingle-mother households are small, less educated and impoverished.All of my respondents contributed to the upkeep of their dependentchildren, whether the children stayed with them or not. Eighty-eight percent of respondents' children were below the age of 18, the official age ofmajority. In some cases, child support continued well after attainment ofthe age of majority. However, 14% of respondents reported that they didnot support their children. Their children were either grown-up or stayingwith their fathers. All those with dependent children staying with theirfathers made irregular remittances to their children and yet they did notconsider this as 'material' support. They thought that their contributionsdid not significantly count towards a child's upkeep. In one case a woman,who was the sole care-giver to her daughters aged three and six years,reported that she did not contribute to the upkeep of her children, because104 RURAL-URBAN LINKAGESTable 9RESPONDENTS' OWN CHILDRENTotal %One child 11 22Two children 14 28Three children 8 16Four children 5 10Five children 3 6Six children 2 4No child 7 14Total 50 100her only source of cash income for herself and the children wasmaintenance money from one of her children's fathers. She did not considerher labour input as 'in kind' income.Seventy-two per cent of the women contributed to their children'seducation. Schooling was a common topic amongst the women whentalking about their children. In a way they saw their children's educationas an investment they had to make in their struggle to escape frompoverty.In a few instances single mothers with a grown-up child or children,tried to secure for themselves an investment in old age by partially caringfor a brother's child or children. Hence for some female heads ofhouseholds, reproductive responsibilities did not necessarily end withtheir own children.Whilst the size of the urban household in many instances did notpresent a complete picture of the demands on the incomes of myrespondents, it gave a glimpse of the day-to-day demands on resources.Table 10DISTRIBUTION OF SIZE OF URBAN HOUSEHOLDSTotal %Less than 5 persons 13 265-9 persons 20 4010+ persons 17 34Total 50 100V. N. MUZVIDZIWA 105Many of my respondents had lived in Masvingo town for quite sometime, yet 68% had retained their rural roots.Table 11RESPONDENTS' LENGTH OF STAY IN MASVINGO7bta/ %__Less than 5 years 11 225-9 years 16 3210+ years 23 46Total 50 100Many respondents, despite having lived in Masvingo for a long time, had avery high rate of residential mobility, as shown in Table 12.Table 12RESPONDENTS' PERIODS OF STAY AT THE RESIDENCES DURING INITIALINTERVIEWTotal %Less than a year 19 381-5 years 20 406+ years 11 22Total 50 100The single room for my respondents was a multi-purpose unit. Itfunctioned as a kitchen, a bedroom, a lounge or sitting room and astoreroom. In my sample, 54% cooked inside the one room, 20% cooked ina room designated as a kitchen (the kitchen sometimes was used as abedroom at night), and 26% cooked outside in an open space or in a shackor shed, the weather permitting. At times some of the members of thehousehold used such shacks as bedrooms.I solicited information concerning my sample's health. Up to 16% hadbeen ill within the six months preceding the initial interviews. It wasnoteworthy, however, that all these women had applied self-care andhome-made remedies or procured over-the-counter medicines. Only twohad sought treatment, one from Mucheke Clinic and another from aherbalist. My respondents minimised or avoided altogether health-relatedexpenses.106 RURAL-URBAN LINKAGESHowever, 40% of the respondents also reported that in the six monthspreceding the initial interview a member of their urban household hadbeen ill and needed medical attention. Of the 20 respondents who had asick household member, 15 noted that treatment was sought from a clinic,while three had received treatment at the Masvingo provincial referralhospital. Only two had sought traditional medical intervention. Thirty percent of the respondents had spent part of their time caring for a sickdependent. Nine cared for own children, four for a sick parent and twocared for a relative other than parents. Eleven had spent more than threehours a day caring for a sick person. On the other hand, four had spentless than three hours a day caring for the sick. Twelve respondentsreported that the sick person had recovered, five said that they wererecovering and three noted that the sick person had died.Access to rural resourcesAndreasen (1990, 164) demonstrated the importance of access to ruralland by a family or household head as a major determinant of splitresidential patterns. In Masvingo the situation was very different: only onerespondent, a widow, claimed ownership of some 12 acres of rural land.Ncube (1987, 20-21) noted that the Matrimonial Causes Act (1985)strove to establish a fair and equitable formula for the re-allocation ofmarital property at divorce: nevertheless, whilst the marriage lasts andupon the death of the husband, the woman remains unprotected by thelaw, leading many to argue that a woman was better off divorcing herhusband than to wait for his death. For Ncube, the Matrimonial Causes Actexcluded property by way of inheritance as well as that acquired in termsof any custom. Thus communal land appeared excluded from re-allocation.The system of land tenure and ownership denies women control ofthe only form of meaningful property in communal areas. At marriage,women acquire only rights of use of communal lands. Most of myrespondents had no access to land, mostly due to their status as femaleheads of households. A paltry 8% of the sample had access to rural land,owned by their male kin. Those widows and divorcees who had previouslyhad access to land through marriage, had lost their land-use rights. Onlyone widow still maintained her rights to use land. Her rural homesteadand livestock were looked after by a paid herdsman. She used hired labourto till her lands. This was despite the fact that she visited the village onlyinfrequently. This small fraction of the respondents with access to landproduced food crops for sale as well as urban consumption. For theserespondents, food production was useful in balancing their urbanhousehold budgets.With the exception of one respondent, a widow with 16 cattle, none ofthe respondents owned cattle, though four women owned goats. KinV. N. MUZVIDZIWA 107looked after the livestock of the research participants. Only one womanemployed a herd-boy whom she paid $60 per month.Widows and divorcees appeared to be the greatest losers in respect ofgetting a share of the property they had helped to acquire. Thirty-six percent had built a house in their husband's homestead. They had lost thisproperty to the husband or to his kin upon divorce or widowhood. Up to40% of my respondents owned a hut, or one or two rooms in the rural area.Their children or kin, especially parents and families of brothers, usedtheir rural housing units. Although the rural housing units were of littlemonetary value, they strengthened the respondents' double-rootedness.However, despite the fact that female heads had no access to rural land,they still maintained village ties so as to retain their right of return to thevillage should things not work out in town or for when they might retire.Access to urban resourcesOwnership of, and access to, urban land and housing in Masvingo clearlydemonstrated the vulnerability of my respondents. Only 16% had an urbanhome, all but one in Masvingo. The one respondent who owned a one-roomed unit in Mupandawana growth point lived in a mapoto relationshipin Masvingo and sold carrier bags for a living. All Masvingo home-ownersexcept one were widows who had inherited their property from theirdeceased spouses. An urban home was thus beyond the reach of most ofmy research participants.The research participants had two major uses of urban space. First,urban housing units were used as residences, and second, the housefunctioned to generate an income or as a place from which to operateone's business activities such as servicing clients by prostitutes or forverandah food sales. In my study, half the respondents produced cropsfrom their urban gardens, mostly for their own consumption although fewproduced for sale.My respondents had very limited access to urban facilities and services,including health facilities. Most women could simply not afford the cost ofhealth provision even at the cheapest level of the clinic. Each visit to theclinic cost them $10. Hence self-care became the primary means to maintainhealth. In a similar manner these cost-related problems prevented any ofmy respondents from pursuing further studies.Only one respondent received $100 per month from the Departmentof Social Welfare. She was a widow and had to supplement this grantthrough selling food. The rest did not qualify for Social Welfare assistance.As able-bodied women, they all failed the means test for Social Welfareassistance. Not only were they ineligible to receive welfare grants, theyhad no access to credit and financial institutions in town. Only threerespondents had ever benefited from loans obtained from formal credit108 RURAL-URBAN LINKAGESinstitutions. Despite being in town, female heads of households had verylimited access to urban institutions and facilities.There is a tendency to associate towns with formal employment, butonly 12% of my respondents derived their primary source of income froma formal job. My respondents depended instead on informal alternativesources of income. My respondents had no durable assets to fall back on.For most the only form of durable asset consisted of their furniture(generally a single or three-quarter bed), and a few utensils such as pots,plates, cutlery and dishes. Even when sold in times of crisis, these goodsrealised very little. However, 52% of respondents intended to accrueurban assets, another 20% wanted to accumulate rural assets, especiallyto build a rural home.I collected data on the sources of energy used by female heads forcooking as well as lighting and heating. For my sample, 88% used electricityfor lighting their rooms. Thirty per cent used only paraffin for heating andcooking purposes, 50% generally used electricity for heating and cookingpurposes, 20% used both firewood and paraffin. Occasionally, the peoplewho used electricity also used either paraffin or firewood. All therespondents whose source of energy was firewood or paraffin spent lessthan $20 per month on energy. Costs for electricity consumption variedbetween $5 and $80 per month. Many respondents used electricity forlighting purposes only.Investments in rural and urban networksAmongst the many strategies adopted by my Masvingo sample was thedeliberate investment in rural-urban networks. Continued utilisation ofrural-urban networks as a survival option was a pragmatic response toscarcity of resources, including jobs. Through an examination of rural-urban linkages, one is able to see how institutionalised regulations wereside-stepped, manipulated and even resisted by the women. On the otherhand, laws that had a bearing on migrants, especially women, such as theVagrancy Act, conditioned and provided the arsenal, language andideological space for understanding rural-urban linkages.Kin connections underpinned rural-urban networks. Whilst the logicof urban living meant side-stepping some obligations to kin, the realitywas that for most respondents kin continued to be significant. Myrespondents fell into two groups, the double-rooted and the permanentlyurban. These categories distinguished my respondents in terms ofperceived attitudes towards urban stay. I constructed an index for double-rootedness and permanently urban on the basis of a composite measureof ten variables. The ten variables were intended to measure the degree ofcommitment to the rural village and town. The categorisation of double-rooted or permanently urban resulted from the total score of the tenV. N. MUZVIDZIWA109Table 13THE DOUBLE-ROOTEDNESS AND PERMANENTLY URBAN MEASURESVariable1 Housing situationin town2 Split residence3 remittances4 attitude towards kin5 length of stay intown6 frequency of visitsto th«»^Hage7 total income8 number ofdependents9 urban attitude10 rural attitipdeValue = 1ownerall in town0negative10+ years0$1000+1 or lesspositivenegativeValue = 2rentalsome in vill.1-3 per annumnon-committal5-9 years1-2 per annum$601-6992-3non-committalnon-committalValue = 3lodgerall in village4+ per annumpositive0-4 years3+ per annum<$6004+negativepositivedouble-rootedness = total score 23-30urban-permanency = total score of 10-22measures. For my study there were 68% double-rooted respondents. Myresearch participants continued to use rural-urban networks as part oftheir strategy to ensure their continued stay in the city. Only 32% saw thetown as their permanent home. Generally they considered the town amarket place from which one derived a living.My double-rooted respondents continued to activate rural-urbannetworks in their day-to-day living experiences. They intended eventuallyto return to their villages. In this context, none defined the villages of theirex-husbands or deceased husbands as 'their' homes. When asked wheretheir 'home' is (Kumusha kwenyu ndekupi?), they invariably gave theirvillage of origin, although as we shall see this does not apply in all contexts.Respondents saw their stay in town as temporary and believed that theywould eventually return to the village. Consequently, they perceived thattheir rural village was their home: in the women's ordinary talk, 'home' isreserved for the village. Many respondents participated in villageceremonial activities such as weddings, funerals and other get-togethers.For instance, a never-married single mother noted:It is quite good for one's well-being to be constantly in touch with one'svillage. One should be present at family functions especially funerals. Itis a pity that most women cannot afford to do so these days.^0 RURAL-URBAN LINKAGESInterestingly, although this woman had no intention of ever returningto her village even in old age, she showed commitment to her village oforigin. This is in contrast to conclusions reached by Schlyter (1990, 188)thatit seemed impossible for women householders to maintain two placessimultaneously as a survival mechanism; they were urban, or they leftand resided for a longer period in rural areas.The 68% of women who were double-rooted seem to be saying,'Although I live in town, I shall maintain contact with my village of origin.'They wanted to hang on in Masvingo town for as long as they could, andhad no immediate plans to return to the village, but regarded the urbansituation as insecure, and knew that they would eventually return to thevillage in retirement or when everything else failed. This is the pattern ofearlier generations of immigrants observed by people like Mitchell (1970).Those who opted for the double-rooted strategy thought that urban lifewas too expensive. They realised the impossibility of continued urbanstay in old age. Many agreed that at some point they would go back to thevillage. Some female heads in the study contributed to the education oftheir village-based brother's children. This ensured easy re-entry intovillage life at some future date. Most continued to communicate with andcare for village-based parents and kin.The women who intended to pursue the permanently urban strategyhad socio-economic benefits in mind. They perceived the standard ofliving to be better in town compared to the village, and that there weremore survival options in town compared to the village.Double-rooted/permanently urban and other variablesFive variables (respondents' age, ethnicity, number of children, educationallevel and frequency of rural visits) were not significantly related to eachother even at the 10% level. Another three variables, namely the povertyindex, sources of income and size of urban residential household, werefound to be weakly related to the variable of double-rooted or permanentlyurban (DR/PU). They had significance levels of 0,0922,0,07576 and 0,09181respectively. However, Table 14 shows that DR/PU cross-tabulated againstlength of stay in town was significant at the 5% level. The longer one hadlived in town, the more likely that one would be in the urban permanentcategory. Yet overall, despite the length of stay in urban Masvingo, mostrespondents maintained links with their villages of origin.Marital status was also significantly related to the PU/DR variable atthe 5% level. As shown in Table 14 all but one widow were in the PUcategory. Widows in my sample had acquired ownership of urban propertyafter the death of their husbands. This explains the high preference byV. N. MUZVIDZIWA 111Table 14DOUBLE-ROOTED/PERMANENTLY URBAN AND LENGTH OF STAY INTOWNDRPUTotal%<5 yrs1011122DR = Double-rooted; PUChi-squarePearsonValue5,660755-3 yrs124163210+ yrs12112346= Permanently urbanDegrees of Freedom2Total341650100Significance0,05899%6832100100widows for the urban strategy. They knew that even when things get toughthey were unlikely to get the support of their affines. Yet widows referredto the homes they helped to build in their late husbands' villages as 'their'own, sometimes commenting that they had lost their homes to their in-laws. The never-married and single persons and those in mapotorelationships were predominantly double-rooted. A high proportion ofdivorcees and the never-married-single mothers opted to be permanentlyurban.Table 15DOUBLE-ROOTEDNESS/PERMANENTLY URBAN AND MARITAL STATUSDRPUTotal%NM741122D1541938Widows16714LWM51612Single Total6 341 167 5014 100%6832100100NM = never married-single mothers; D = divorced; LWM = living with a manChi-squarePearsonValue12,08195Degrees of Freedom4Significance0,01675Double-rootedness and permanently urban are cross-tabulated againsthousing status in Table 16. These variables are strongly related. All house112RURAL-URBAN LINKAGESowners are in the permanently urban category. House ownership had astrong influence on the decision to stay in town permanently. Half of thetenants were in the permanently urban and another half double-rooted.Twenty-nine out of 33 lodgers opted for the double-rooted strategy, arational response given their limited options.Table 16DOUBLE-ROOTEDNESS AND PERMANENTLY URBAN AND HOUSINGSTATUSLodger Tenant OwnerTotalDRPUTotal%2943366551020077143416501006832100100Chi-squarePearsonValue22,35684Degrees of Freedom2Significance0,00001Available housing space for use by the respondent also had a bearingon DR/PU as shown in Table 17. The more space a respondent had, thegreater the probability of urban permanence.Table 17DOUBLE-ROOTEDNESS/PERMANENTLY URBAN AND HOUSING SPACEOneRoomSharedRoomTwo+RoomsTotalDRPUTotal%1742142152173421012243416501006832100100Chi-squarePearsonValue19,3498Degrees of Freedom2Significance0,00006Table 18 shows that split residential practice (a situation in whichsome dependents reside elsewhere Š this is what Croll [ 1987,489] refersV. N. MUZVIDZIWA113to as 'segregated families') was the single greatest predictor of double-rootedness or permanently urban. Split household practice necessarilyled to double-rootedness. Respondents had to cater for the welfare oftheir dependents, particularly children, who lived elsewhere, and thismeant high rates of remittances, rural visits, etc. Resource ownership,whether of rural or urban property, significantly related to the DR/PUvariable at the 5% level. Those with urban property, especially housing,were more likely to opt for permanently urban. The propertylessrespondents were predominantly double-rooted.Table 18DOUBLE-ROOTEDNESS/PERMANENTLY URBAN AND SPUT RESIDENCEAll dependentsresiding in thevillageDependents splitbetween townand villageAll the dependentsresiding intown TotalDRPUTotal%1501530141153051520403416501006832100100Chi-squareValueDegrees of Freedom SignificancePearson28,477330,0000Table 19DOUBLE-ROOTEDNESS/PERMANENTLY URBAN AND RESOURCEOWNERSHIPDRPUTotal0/Chi-squarePearsonHas Property391224Value13,4167No Property3173876Degrees of Freedom1Total341650100Significance0,00025%6832100100Respondents' total income was also significant at the 5% level inrelation to DR/PU variable. Most respondents with an income above114RURAL-URBAN LINKAGES$1000 per month opted for the permanently urban strategy, while a higherproportion of those with incomes below $1 000, were double-rooted.TOTAL INCOME ANDDRPUTotal%Chi-squarePearson<$6001451938Value9,87854Table 20DOUBLE-ROOTEDNESS/PERMANENTLY$600-999 $1 000+13 71 1014 1728 34Degrees of Freedom2Total341650100Significance0,00716URBAN%6832100100Table 21 shows that the higher the frequency of remittances to therural village by a respondent, the more likely was the adoption of thedouble-rooted strategy. Respondents who remitted less favoured the urbanstrategy. At the 5% level the relationship between the double-rooted/permanently urban variable and frequency of remittances is significant,showing that these variables had an effect on each other.Table 21DOUBLE-ROOTED/PERMANENTLY URBAN AND REMITTANCESFREQUENCYDRPUTotal%Chi-squarePearsonNone481224Value9,649341-3 yrs16622444+ peryr1421632Degrees of Freedom2Total341650100Significance0,00803%6832100100Investments in time and moneyAn important element that showed commitment to the village and one'sdependents was the issue of remittances. Remittances were part of theV. N. MUZVIDZIWA 115rural-urban support network. According to Bigstein and Kayizzi-Mugerwa(1992, 1428), 29% of households in Kampala remitted money or materialmeans to children living elsewhere, mostly in rural areas. Remittancesconstituted at least 9% of total income. Figures elsewhere (e.g. Nairobi)showed that remittances could be as high as 20% of total income. Myrespondents remitted cash mostly, and sometimes material goods.Some 76% of my respondents remitted money or material means tothe villages where their children lived. They contributed to the welfare oftheir parents in part or in full. It was not possible to state categorically inquantitative terms the volume of remittances. On the basis of conversations,it seemed many respondents found it increasingly more difficult to remitsomething to their dependents living elsewhere. Especially those in thecategory of 'climbing-out of poverty' shunned connections with kin.Investments in networks of kin were more deeply entrenched amongst the'hanging on' and 'coping' groups.In addition to remittances, it is important to notice kin networks inexamining Masvingo respondents' investments in terms of money andtime. Hospitality to kin had its costs. It was clear that the respondents feltthey could not sustain all the demands of rural kin on their meagrehousehold resources in town, yet they felt they could not deny kin anyaccess to their resources. Around half housed kin when they visited town.In most cases they also provided food for kin during these trips. Theyhoped that their rural kin would understand that they were also in difficultsituations.I observed that the double-rooted strategy pursued by most of myrespondents involved the exchange of goods and services, a reciprocalflow of resources from both rural and urban directions. These issues arediscussed in the next section.The role of kinKin influenced the mobility of women in my sample, especially their initialmove to, and arrival in, town. All but two respondents had discussed withkin their plans to migrate to Masvingo. Kin made it easy for new arrivals toexplore the urban environment.Most often rural-urban ties are articulated through kinship networks.On arrival in town for the first time, 84% had stayed with kin. New migrantsto the city stayed more often with maternal relations than paternal relations.In fact only one respondent in my study had stayed with a paternalrelative. The period of stay with kin varied from a week to three years. Myrespondents were without exception from a patrilineal and patriarchalsociety. A woman was thought more likely to compete for status, economicadvantages and group leadership with paternal relations as opposed to116 RURAL-URBAN LINKAGESmaternal relations. Competition over access to resources was regarded aslikely to generate hostilities that undermine possibilities of co-residentialoptions with one's paternal relatives. It is usually one's maternal relativeswho are supportive in a non-antagonistic manner. More generally, a womanliving with her paternal relatives would be expected to perform manymundane household chores and would generally be at a disadvantage.During the initial phases of urban migration, kin assisted some of myrespondents to get a job in the formal or informal sectors. In my sample,20% had been assisted by their siblings to earn an income. They werespared household financial contributions during the initial phases of theirentry into urban life, to enable them to build up capital to set themselvesup in the urban economy. However, the role of their siblings in their lifehad diminished by the time of my study.Not only did kin ease my respondents' initial urban stay by assistingwith urban residence, jobs and income. The contributions had continuedin 28% cases, in relation to children's residence, education and welfare. Atthe start of my research amongst my respondents, 8% had a child stayingwith kin in Masvingo, 6% received assistance for children's education and14% noted that their kin assisted in the general welfare of children. Tworespondents were assisted with baby-sitting by urban kin on a regularbasis at least monthly. Another had had kin baby-sit once in the last threemonths preceding the initial interview.However, as a general rule my respondents did not pool financialresources with urban kin. Income pooling occurred only among 10% of mysample. These were the cases where kin made a part contribution towardsthe household budget. Although they did not pool resources, aid from kinin the form of money as well as food occurred during my fieldwork. Some26% of respondents had received monetary grants from kin in the yearpreceding the interviews. Whilst the amounts involved were quite small(almost without exception these grants were less than $50), theyrepresented a continual re-enactment of kin obligations in town. Food aidwas more common: 30% of my sample had received food donations fromkin in the six months preceding the initial interviews.The most frequent form of inter-household exchange of goods andservices involving kin was in the form of moral support. Within the courseof a three-month period, 50% of the respondents received advice from kin.In most cases the advice concerned personal matters. Only 8% noted thatthey got business-related advice from kin.The extension of loans or credit by kin indicated that kin continued toplay useful roles in the lives of my respondents. Twenty-two percent hadborrowed money at least once in a 12-month period, 12% had received aloan two to three times over the same period and another 12% got loansfrom kin three or more times within a year. Bus fares to attend funeralsV. N. MUZVIDZIWA 117and children's education (school fees and uniforms) were the main reasonsfor loans from kin.According to Mutambirwa and Potts (1990, 677) most urban familiesin Zimbabwe had a strong perception that the maintenance of rural linkswas essential as a form of economic security for the eventualities of oldage, unemployment and hard times. The majority in my sample supportedthis view.Table 22 shows the frequency of visits by my respondents to theirvillages within a six-months period. High-frequent visitors were those withthree or more visits during the six-months period. Lack of financial meansor not being connected to any village were the two reasons cited for notvisiting communal areas.Table 22FREQUENCY OF VISITS TO THE VILLAGEFrequency of Visits Number of Visits %None 16 32One to two visits 23 46Three or more visits 11 22Total 50 100The existence of split households was the major reason for ruralvisits. The women went to the village to see their children, siblings,parents and other relatives. A family function such as a funeral, weddingor ritual occasion was often the reason behind the rural visit. The resultsdiffered from Holm's (1992, 248) findings that showed that 65% ofrespondents in his study went to their home village to attend funerals andweddings. My respondents noted in their general talk that due to the everincreasing, Aids-related mortality levels, funerals were no longer the mainreason for going to the village. There were too many funerals and it hadbecome too expensive to visit the village every time a relative died.Village-based kin also visited their kin in town during fieldwork. Someof the respondents noted that kin visited town to conduct business suchas buying agricultural inputs, for banking purposes during month-ends, insearch of employment, in search of schooling for their children, for someto meet administrative bureaucrats at district and provincial levels, andsometimes for health reasons. There were times when kin travelled fromthe communal areas to seek financial assistance from urban kin. Up to 40'V.of respondents had hosted visiting rural kin within a six-month period. Kinvisits occurred whether one was a lodger or tenant.Split residence involving some or all the respondent's children livingelsewhere, notably the rural village, was common: 60".x of the research118 RURAL-URBAN LINKAGESparticipants had split-households, much higher than in Uganda whereBigstein and Kayizzi-Mugerwa (1992,1436) found that 47% of their samplehad dependents mostly children, living elsewhere.For my sample children mostly stayed with maternal grandparents orsome other maternal relations. It was very rare for paternal grandparentsor relations to take the responsibility of looking after children; this occurredonly in 6% of my sample. The split household ensured continued interactionbetween the female head of household in town and her village of origin.Children were in the village mainly as a cost-cutting measure. Therange of ages of children in the village was from infancy, through those atsecondary school, to school leavers. Nearly all the women with children inthe village wanted regular monthly contacts with their children. However,it was economically impossible to carry out monthly visits to rural villages.Attitude towards urban permanency and double-rootednessIt was clear that whatever option the women took, whether it waspermanently urban or the double-rooted strategy, it enabled them to planhow best to survive in town. The women in my sample regarded the urbancontext as offering better opportunities for earning an income. Hence for60% of the respondents, a successful pursuance of the urban strategydepended on the rural strategy of split-residence. But a preference for oneoption did not entail abandoning the other. The women maintained bothreal and symbolic links with their villages of origin. Even in situationswhere respondents did not visit their villages, they frequently continuedto perceive the village as their 'home'. Table 23 shows the women'sattitude on whether they preferred their children to live permanently inMasvingo or not.Table 23PREFERRED PERMANENT HOME FOR CHILDRENPreferred Home for Children Total %Masvingo as permanent home 15 30Masvingo not preferred 18 36Indifferent up to children 7 14Inapplicable no children 7 14Total 50 100The respondents preferred children to make Masvingo a permanenthome because of the perceived existence of attractive income-earningopportunities. They noted that Masvingo was a rapidly growing town, withincreased opportunities for earning a living. Some preferred MasvingoV. N. MUZVIDZIWA 119because of its easy access to its rural hinterland and other towns inZimbabwe. It was convenient staying in Masvingo.On the other hand, those who disliked Masvingo noted that the townhad experienced serious economic decline following the closure of severalindustries, notably the Cold Storage Commission, at one time Masvingo'slargest employer. They did not wish their children to live in such adepressed state of economic affairs. Some of those who had a negativepicture of town life tried to paint a rosy picture of the rural village. Thisgroup consisted mostly of those who were finding it difficult to stay intown, the 'hanging-on' category. Their attitude appears to be arationalisation of their eventual and inevitable return to the village. Theyrealised that their stay in town was only temporary and so tried to createa more positive view of their villages. It was clear that most women had noimmediate plans to leave town for the village: they wanted to hang on intown for as long as possible.Whilst many respondents acknowledged experiencing hard times,they had no intention of abandoning Masvingo. It was only those in the'burnt out' category who conceded that they might be forced to go back tothe village. I also observed that for the 'burnt-out' women, linkages withurban kin were weakly articulated or non-existent. The importance ofexchanges that flow from urban to rural areas, mostly through networksof kin, was absent amongst those who needed assistance the greatest, i.e.the 'burnt-out'. Going back to the village for most respondents was not apreferred option. The women's attitudes were in opposition to thegovernment policy of discouraging unemployed women from becomingpermanent urban residents.Another indicator used to measure the women's attitude towardsbecoming permanently urban was single motherhood. Masvingorespondents considered single motherhood to be a survival option fordisadvantaged women. Despite its shortcomings, marriage remained adesired, but unlikely option for many women: all but one home-owner inmy sample were widows. Housing was short, and the rules of access tohousing put single women at a disadvantage. Preference went to those onthe waiting list with full families. Also, the length of stay in town was aconsideration and generally women have been coming to town morerecently than men. So the criteria for the allocation of limited housingfavoured men over women, and helps to explain why most of myrespondents preferred marriage. Lodging and renting remained the onlyoptions for unmarried women.At times the women used their rural networks to provide food fortheir own consumption and their households in town, as noted above. Attimes rural-urban interactions were mutually beneficial. Through kinnetworks in urban areas, some villagers managed to gain access to urban120 RURAL-URBAN LINKAGESresources, like housing or an urban job. Urban kin had accommodatedmany of my study respondents until they found a job and their own home.However, those respondents who opted for permanently urban statusviewed kin negatively, seeing them as potential claimants to their resources.Yet, in my study the PUs were relatively well off compared to the rest inthe sample.Managing urban poverty: The 'burnt-out' in townThe poorest in town, the 'burnt out', have the fewest resources to invest inthe countryside and to service kin networks. This is part of the povertytrap. This might leave them with no option but to squat. But with draconianpolicies, where? The second paradox is the way those who are precariousmobilise and use resources in such a way as to enable them to continue tostay in town. The rural strategy or double-rootedness is not about going tothe village but how to delay or even avoid going back. The immediatething most women do is to try to generate an income and this requiresservicing both the urban and rural networks. Sometimes children are senthome or pulled out of classes. Some kind of 'delayed reciprocity' takesplace as a strategy to stay in town. Through the activation of relationswith kin and non-kin, the women try to resolve the paradox of beingreceivers only by engaging in some form of return payment. It is in thecontext of the problems that women confront in their attempts to mobiliseassistance and resources to stay in town that the double-rooted strategyis adopted.Respondents Q and W are singled out to illustrate how the poordevelop strategies to deal with urban poverty, both in the 'hanging on'categories at the start of my fieldwork. They represent different strategiesand outcomes pursued by my respondents in order to stay in town.0 is a 40-year-old woman who had never married. She had four childrenwhose ages ranged from 11 to 20 years. Her eldest children, two sons,stayed with her brother and sister-in-law in her village. They both attendedsecondary school. Her two other children, daughters aged 11 and 13,stayed with her in the one room where she lodged. Her two young daughtershad to forgo schooling in 1995, to enable her mobilise resources for theeducation of her two eldest children. For Q the sex of her children did notmatter: if the eldest had been girls she noted that she would have used thesame strategy of temporarily taking the young children out of school. Herpriority during 1995 was to mobilise resources to stay in town, althoughshe was double-rooted. She mobilised the labour of her daughters to raiseher urban income Š they acted throughout my fieldwork as mobile townvendors. Q was able to raise the examination fees for her two boys, payrent, buy food and remit some money to her village. The strategy of splitresidence helped her cope with town life. Q's kin were a resource thatV. N. MUZVIDZIWA 121enabled her to hang-on to town life. She was planning to return herdaughters to school at the beginning of 1996, something she was workingtowards: she showed me the school uniforms she had bought for them inNovember 1995.Respondent Wwas a 34-year-old widow and mother of four sons. Shewas also in the 'hanging on' category at the start of my fieldwork anddepended then for her income on rentals. By the end of 1995, she was'climbing out' of poverty. Her four children ranged from a 14-year-old inform one to an eight-year-old in grade three, and all stayed with her. Shehad become the owner of the four-roomed residential unit after the deathof her husband in 1994. Before the death of her husband, the family hadused the whole unit.At the beginning of 1995, Ws household was in dire financial need.She depended on income from rent ($240 per month for the two rooms shewas letting out to a five-person lodger's household and a two-personhousehold) and also the sale of vegetables, mostly to neighbours, fromher house garden. Her husband had not allowed her to get a passport,since he thought cross-border female traders engaged in immoral sexualactivities. W was thus denied the chance to develop her business skillsduring her husband's lifetime. Her first ever passport was issued at theend of February 1995, and she started cross-border trading at the end ofApril 1995. Her first trip to Pretoria, undertaken with a neighbour, was afiasco. However, all her subsequent monthly cross-border trips to SouthAfrica had been very successful.For nearly a year after her husband's death, she relied on the familynetwork and house rentals to sustain her town life. She was able to use herkin networks to raise her cross-border start-up capital. As soon as she wason her feet she started servicing her kin network especially her affinalrelations. By the end of my fieldwork it was becoming clear that W had anegative view of kin and wanted to minimise connections with kin. She hadstarted making investments in non-kin networks.Respondents Q and W highlighted two paradoxes that affected myrespondents. The first was that the poorest in Masvingo, the 'burnt-out',had the fewest resources to invest in the countryside, in pursuance of thedouble-rooted strategy. Yet, the poorer respondents were in greatestneed of keeping a foot in both the urban and the rural settings. This waspart of the poverty trap; my poorest respondents found themselves caughtin the trap. The second paradox as illustrated by W was the way thosewho are precarious (for instance the recently widowed respondents)mobilised resources to ensure their continued stay in town. It was in Wscase as if through widowhood, the resources the women had sat uponduring their married life were released for immediate use. Widows, unlikeother categories of single women, were in greatest need of financial as well122 RURAL-URBAN LINKAGESas emotional assistance, especially soon after the death of their husbands.In all cases their kin and friends rather than affines came to their side.Widows also sold some assets to meet their immediate needs. Turning thehouse to an income generating asset was very common amongst thewidows who were house-owners.Those respondents who were in precarious situations reacted byturning to social networks of kin or friends, which they used as resourcesfor survival. Despite the determination of the 'burnt-out' women to stay inthe city, in the few cases that I recorded, they were forced to return to thevillage, unlike the women in the 'hanging-on' category. The very poor hadthe least access to social networks. Servicing rural and urban networks ona more permanent basis was the wish of most respondents, but was notalways possible, hence some kind of 'delayed reciprocity' when they actedas 'takers' in anticipation of paying back later.The strategy of double-rootedness enabled most of my respondentsto continue to stay intown. At the same time processes of returning to thevillage by the 'burnt-out' concealed the full extent of urban poverty. Theadministrative policing of poverty through statutes such as the VagrancyAct, and attempts by the Municipality to regulate marketing throughlicensing informal traders, forced the very poor women to leave the cityand take their poverty to the villages. The administrative policing ofpoverty created the impression that things were better off in town incomparison to the villages. This may be part of the processes that fuelrural-urban migration.Officials reinforced the need for the double-rooted strategy byperiodically carrying out clean-up operations to remove the public soresof poverty. The strategies used by the Municipality to prevent squattingdid not achieve the desired results. The report by Masvingo's municipalDirector for Housing and Community Services (1994/95) revealed in relationto squatters:During the past year the shacks they had put up by Mucheke river bankwere demolished. This was not a solution as they are still squatting (p.19).These policies had nevertheless managed to prevent massive squattingin Masvingo and Zimbabwean towns generally.The official regulation of urban poverty, dating back to the colonialperiod, by off-loading the urban 'burnt-outs' onto the rural areas haseffectively meant these areas, and especially the communal villages,continue to act as rural 'dumping grounds'. For many people, includingthe women in my study, migration to town does not lead to a break withthe village (see Logan, 1981, 238). Instead, it creates a continual exchangebetween city and country. Rural-urban migration, as demonstrated in myV. N. MUZVIDZIWA 123study, has thus reinforced rather than destroyed rural-urban socialnetworks. New urban arrivals use resources of their kin based in bothurban and rural areas. In Masvingo many in my sample had learnt how tosurvive in town by adopting such a double-rooted strategy.ReferencesADREASEN, J. (1990) 'Urban-rural linkages and their impact on urban housingin Kenya', in Baker, J. (ed.) Small Town Africa: Studies in Rural-UrbanInteraction, Urban-Rural Linkages and Their Impact on Urban Housing(Uppsala, The Scandinavian Institute of African Studies), 161-171.BIGSTEIN, A. AND KAYIZZI-MUGERWA, S. (1992) 'Adaptation and distress in theurban economy: A study of Kampala households' World Development,XX, (10), 1423-1441.JACKSON, J. C, AND CHEATER, A. P. (1994) 'Contract farming in Zimbabwe:Case studies of sugar, tea, and cotton', in Little, P. D. and Watts, M. J.(eds.) Living Under Contract: Contract Farming and AgrarianTransformation in Sub-Saharan Africa (Madison, University of WisconsinPress), 140-146.CROLL, E. (1987) 'New peasant family forms in rural China' The Journal ofPeasant Studies, XIV, (iv), 469-499.FINAN, T. J. AND HENDERSON, H. C. (1988) 'The logic of Cape Verdean female-headed households: Social response to economic scarcity' Urban.Anthropology, XVII, (i), 87-103.CENTRAL STATISTICAL OFFICE (1993) Census 1992: Provincial ProfileŠMasvingo.(Harare, CSO).HOLM, N. (1992) 'Survival strategies of migrants to Makambako Š Animmediate town in Tanzania' in Baker, J. and Pedersen, P. O. (eds.)The Rural-Urban Interface in Africa: Expansion and Adaptation (Uppsala,Scandinavian Institute of African Studies), 238-257.LOGAN, K. (1981) 'Getting by with less: Economic strategies of lower incomehouseholds in Guadajara' Urban Anthropology, X, (iii), 231-246.MASVINGO DIRECTOR OF HOUSING AND COMMUNITY SERVICES (1995) Annual Reportof Housing and Community Services for the Year Ended 30th June 1995.MITCHELL, J. C. (1970) The Causes of Labour Migration. Black Africa: ItsPeoples and Their Cultures Today (London, Macmillan), 23-37.MUTAMBIRWA, C. AND POTTS, D. (1990) 'Rural-urban linkages in contemporaryHarare: Why migrants need their land' Journal of African Studies, XVI,(iv), 677-698.NCUBE, W. (1987) 'Underprivilege and inequality: The matrimonial propertyrights of women in Zimbabwe', in Armstrong, A. and Ncube, W. (eds.)Women and Law in Southern Africa (Harare, Zimbabwe PublishingHouse), 3-34.SCHLYTER, A. (1990) 'Women in Harare: Gender aspects of urban-ruralinteraction', Baker, J. (ed.) Small Town Africa: Studies in Rural-UrbanInteraction (Uppsala, The Scandinavian Institute of African Studies),182-191.