h.u 1.Aspects of Status and Mobility among Farmers and TheirFamilies in Msengezi African Purchase LandAngela P. CheaterDepartment of Sociology, University of Rhodesia, SalisburySociological studies of mobility, here de-fined as the movement of individuals andcategories of people within society or 'socialspace' (Sorokin, 1959), have tended to con-centrate on urban dwellers, particularly inEurope and North America. To a consider-able extent, this urban focus is also found inthe literature on migration, or geographicalmobility, in African countries, although somestudies have dealt with certain of the effectsof such geographical mobility on the ruralareas to which the migrants return (e.g. Scha-pera 1947; Watson 1958; Winter 1955). Butwith the exception of those studies of socialcategorisation and social movement which arebasically theoretical and may, therefore, applyin principle to rural as well as to urban dwel-lers (e.g. Lloyd 1966; Tuden and Plotnicov1970), there is a dearth of information onmobility in a rural context in African societies.This paper is thus intended to contribute tothis hitherto neglected field.The politico-administrative intention behindthe creation of the freehold farming areasknown as Purchase Lands in Rhodesia in thenineteen-twenties was, inter alia, to create a'yeoman peasantry'. It is generally consideredthat this 'yeoman peasantry' has not, in fact,materialised; so in order to indicate what sortof 'class' categorisation and mobility patternsare actually arising in these areas, an examina-tion of one old-established Purchase Land with-in one hundred kilometres of Salisbury will beundertaken. In order to make this examination,however, it is necessary first to identify theobjective components of social categorisationand then to consider the types of mobility rele-vant to such classification.It is frequently asserted (e.g. by Lloyd,1966) that social divisions and social categori-sation in African societies run along ethnic,lineage and age lines, and that internal percep-tion of horizontal distinctions, in terms ofaffluence, occupation and status based on thesecharacteristics, is lacking in these societies.Lloyd notes, too, that political considerationsin the form of ideologies of classlessness, maybe partly responsible for this alleged non-recognition of socio-economic divisions inmodern African states. In Msengezi PurchaseLand, however, the explicit recognition ofsocio-economic divisions among farmers isbeginning to be verbalised, and a rudimentaryhorizontal stratification, based on distinctionsof wealth, education and occupation, is readilyidentified.Since wealth (measured in terms of mate-rial possessions) and occupation are the mostimportant determinants of social status, thesefactors have been used in assessing and cate-gorising the status of individual farmers. Theindices of education and occupation have beenused in an attempt to demonstrate socio-economic mobility among farmers' children,and to indicate some factors affecting suchmobility. Because mobility is a dynamic con-cept, an attempt has been made to incorpo-rate a time perspective by using employmenthistories and by considering the more import-ant factors (such as education policy) whichmay affect the future occupational mobilityof the second and subsequent generations ofPurchase Land origin. Before considering socio-economic factors, however, one must first exa-mine the physical movement or geographicalmobility which brought these farmers to thisparticular Purchase Land.51GEOGRAPHICAL MOBILITYAll landholders have come to the PurchaseLands as immigrants, selected by successiveland boards, from widely dispersed areas, andare thus ethnically and linguistically hetero-geneous. Table I shows the farmers' areas ofbirth and the areas from which they actuallymoved to settle in Msengezi. People who camefrom different areas generally display import-ant differences in cultural outlook: in particu-lar, people from mission stations and townsshow significant differences in outlook' fromthose who came directly from tribal areas.Such differences in behaviour patterns andvalues are significant determinants of expendi-ture patterns, and all three influence socialclassification.Table IFARMIIOLDERS' BIRTHPLACES AND 'LASTPORTS OF CALL'Area Category Birthplace 'Last port of call'Reserves/TTLs 75,6 53,9Mission Stations 9,3 15,5European-owned Farms 3,8 5,8Purchase Lands 4,5 2,6Towns 5,5 22,3Outside Rhodesia 1,4 ŠTable I shows that the majority of farmersrelocated from one rural home to another,although considerably fewer moved fromtribal areas than had been born there, indi-cating considerable geographical mobilityprior to settling in Msengezi. This past move-ment reflects occupational mobility and ex-posure to new employment as wage earners.Most employment opportunities were in towns,and this distribution is reflected in the propor-tion of farmers who moved into Msengezifrom various towns. In general, those whomigrated from towns were not migrant lab-ourers, but relatively well-educated and stabi-lised urban workers in white-collar, lowerprofessional or uniformed services employ-ment. Many owned town businesses and somestill do, including those wealthy businessmenwho have bought farms from previous title-holders and who manage these farms fromtheir town homes. This movement outwardsfrom the towns into this Purchase Land byrelatively well-educated men holding respon-sible positions, does not represent a basicallyrural orientation, as some Rhodesian Whitesbelieve, but a much more fundamental searchfor land security in the form of freehold title,oriented towards retirement free from theconstraints of a tribal system to which thesemen do not subscribe. Indeed, the search forland security and a desire to live in privacy,together with a dislike of tribal authority,have motivated men from the Tribal TrustLands to settle in Msengezi as well.It is generally believed that Purchase Landfarmers enjoy a higher standard of living thanis characteristic of the tribal areas and thereis some evidence to support this belief* In-deed, some purchase land farmers enjoy ahigher standard of living than do the majorityof urban workers. The orientation of Msengezipeople is towards what they perceive as 'townstandards', and at the same time they con-sider their position to be immeasurably betterthan that of tribal area residents. In the lightof these views, it is interesting to note thepresent areas of permanent residence of Msen-gezi farmers' children born before 1955 andthus of employable age.Table IIPRESENT PERMANENT RESIDENCE AREAS OFFARMERS' SONS AND DAUGHTERSVArea CategorySons: Daughters: Total:Reserves/T.T.L.s 4,0 10.8 7,4Mission Stations 0,5 1.0 0,8European-owned Farms 1,0 0,4 0,7Purchase Landsf 56,2 58,6 57,4Towns 30,7 23,9 27,3Outside Rhodesia 2,4 2,1 2,2Boarding Schools 5,2 3,2 4,2A comparison of Tables I and II indicates avery small return flow of farmers' children*For instance, in the much higher populations ofpurchase land farm homesteads than of those intribal areas (cf. Hughes, 1974) ; in the much in-creased proportion of crops sold from Purchase Landfarms (Hunt, 1960) ; and in Msengezi particularly,by ownership of a functioning motor vehicle by 18per cent of the farmholdcrs.ŁfThis category includes sons and daughters aged 19 to25 years who are single and working on their fathers'farms; older daughters married to other purchaseland farmers; sons who have bought their own farms;and sons and daughters working as teachers, nursesand other local council employees in Msengezi orother purchase lands.52p.to the tribal areas, with a substantial differ-ence between sons and daughters in this respect:more than double the number of daughterslive in tribal areas than do sons. The majorityof farmers' sons who are resident in tribalareas are in fact skilled or professionalworkers (notably agricultural extension assis-tants and teachers): very few are tribal culti-vators. Among farmers' daughters resident intribal areas, in contrast, a minority are tea-chers or nurses: most are married women,their husbands being either cultivators ormigrant workers. Thus rather more femalesthan males are 'shipped back' to tribal areasfrom Msengezi Purchase Land, and this differ-ential is related to marriage: women are morelikely to 'marry down' and move to a tribalarea if they marry men from these areas,whereas farmers' sons who marry womenfrom tribal areas either bring their wives tothe Purchase Land or take them to town. Thefact that fewer daughters of farmers marryspouses whose homes are in tribal areas thando their brothers (64 compared to 71 per cent),may be related to this differential movementbetween areas which are differentially rankedin terms of socio-economic status.Although Table II includes both marriedand unmarried children, those ages range up-wards from nineteen years, it would appearthat most of these children remain within thePurchase Land category, while the next largestproportion establishes town homes. These pro-portions may alter slightly with movementthrough the age structure and with marriage,but the broad pattern is likely to remain un-changed. If we consider this geographicalmovement in the light of the socio-economicranking, in 'upward' order, of tribal areas-Purchase Lands-towns, found among PurchaseLand residents, then it would seem that thereis very little 'downward' mobility to the tribalareas. What downward movement there is inthis respect affects mainly women, throughmarriage. In contrast, nearly one-third offarmers' sons and nearly one-quarter offarmers' daughters move 'upwards' to thetown level. Of course, geographical movementof this nature does not always coincide with achange in socio-economic status, which isheavily dependent on occupational mobilityin present-day Rhodesia. For instance, thesocio-economic circumstances of extensionassistants, teachers and nurses working in thetribal areas do not reflect downward mobilitydespite geographical movement to a less highlyregarded land category. Likewise, the positionof unskilled labourers in town does not reflectupward socio-economic mobility, irrespectiveof the move into a more highly regarded en-vironment. Nevertheless, the trend of geogra-phical mobility does reflect, broadly, the socio-economic values of Purchase Land residents ex-pressed in the tribal areas-Purchase Landstowns rank order.INCIPIENT INTERNAL STRATIFICATIONSubstantial differences in incomes, incomesources and standards of living exist withinmost Purchase Lands. These differences restto a large extent on differing educationallevels, which have influenced past and presentemployment experience in both town andcountry, as well as on differential agriculturalproductivity. Contemporary patterns of strati-fication and mobility therefore have their rootsin the (non-agricultural) past occupations ofmen who are now farmers. As such, one mustbe careful not to assume that the present situ-ation is 'indigenous' to this purchase land: itis in fact related closely to conditions in thewider, external society of Rhodesia and hasrelatively little to do with farming itself. In-come levels in Msengezi range from subsis-tence plus to several thousands of dollars perannum, with relatively few cases at either ex-treme. A composite economic index is a moresatisfactory method of assessing economicstanding than is claimed income, however, notthe least because such an index reflects ex-penditure patterns and, therefore, values. Interms of such an index, incorporating type ofhousing, transport ownership, business inte-rests, income sources, farm enclosure andmechanisation, individual farmers have beenclassified into four economic categories,A, B, C and D in decreasing rank order. Re-spectively, 6, 16, 76 and 2 per cent of farmersfall into these categories, and for all practicalpurposes, category D is of negligible signifi-cance and may be ignored.Having identified these differing economiccategories, educational and occupational le-vels were then plotted against economic cate-gory (see Diagrams 1 and 2). From these dia-grams, it is clear that the educational andoccupational levels of farmers in economiccategory A are considerably higher than those53100 -rDiagrain 1EDUCATIONAL LEVEL OF FARMERS BY SOCIO-ECONOMIC CATEGORYsocio - economic categoryKey to educational categories1 five years of primary schoolingor less (up to standard 3)2 between six and eight years ofprimary schooling (standards 4-6)3 eight years of primary schoolingplus further vocational training4 some secondary schooling5 some university education1 2345123451234 512 34 5educational categoriesDiagram 2LAST OCCUPATIONAL CATEGORY OF FARMERS BY SOCIO-ECONOMICCATEGORYsocio - economic category7550..25--ADKey to occupational categories1 unskilled2 skilled or supervisory3 white collar, uniformed servicesand lower professional4 self-employed in business5 higher professional, managerialand executive123 45 123 4512345 1 234 5occupational categories54of farmers in any other economic category:indeed, the distribution patterns for categoryA farmers are quite distinct from those ofother categories. The differences in educa-tional and occupational levels between farmersin economic categories B and C, however, areof degree rather than kind.Marital status was also plotted againsteconomic category, revealing the distinctivepattern of 100 per cent monogamy amongfarmers in economic categories A and D andover 20 per cent polygynous marriages amongthose in economic categories B and C. Pre-sumably the monogamous status of those incategory D is explained primarily by theirfinancial inability to marry several wives, butthe same argument cannot be advanced forthose in economic category A, where themonogamy rate must be seen as a result ofthe adoption of European-model culturalnorms by generally better-educated men, mostof whom have been employed in professionaljobs.Having shown that the indices of educa-tion, occupation and marital status add somemeasure of confirmation to economic distinc-tions in Msengezi Purchase Land, one mustask whether these socio-economic categoriesare simply heuristic devices or whether theyactually reflect some kind of incipient classstructure. In order to answer this question,one must look at patterns of interaction amongthe individuals within these categories. 'Inter-action', in this context, includes both socialvisiting and common membership of commit-tees and voluntary associations, but excludesall forms of involuntary and formal behavi-our enjoined by public expectation and/or byimpersonal role enactment.INTERACTION ]Table IIILINKS WITHIN DIFFERENTSOCIO-ECONOMIC CATEGORIESCategory Intra-Categoryrandom av.A 6B 16C 76D 2Actuallinks% 7c2623770% increase overrandom-expectation433144_The figures in Table III suggest that thereIs a developing tendency for persons of highersocio-economic standing in this Purchase Landto interact among themselves rather than withwith persons of lower status: this tendency ismost pronounced among those in socio-economic category A. It would seem, then,that these categories represent what might betermed 'proto-classes' in a developing, or inci-pient class structure. Indeed, some farmersexplicitly recognise 'rich', 'average' and 'poor'farmers as socio-economic classifications,while many others distinguish 'educated' and'uneducated' as identifiable categories. Thislatter distinction indirectly points to distinc-tions of wealth as well. Furthermore, in iden-tifying important individuals in Msengezi,many farmers specified wealth as one reasonfor their importance: all of those whosewealth was thus explicitly identified fall intocategories A or B above. There is thus consi-derable direct evidence to support the conten-tion that a class structure based on socio-economic factors is emerging in this PurchaseLand, both in terms of the recognition of socio-economic differences by the farmers themselvesand in the reflection of these differences inpatterns of interaction.Further supporting evidence for a develop-ing class structure may be found in interactionclaims: in aggregate terms, people in the lowersocio-economic categories claimed more linksto those in higher categories than were reci-procated, which suggests that, as farmers im-prove their socio-economic standing, they tendto sever links with erstwhile friends and withclassificatory and affinal relatives who remainat lower levels. Those in the lower categories,however, tend to maintain their claims to in-teraction with those who have moved upwards,partly for the 'rub-off' in status that such claimsconfer.Finally, although there exists in Msengezian egalitarian ethic which is expressed in suchterms as 'There is no-one important here' and'We are all just farmers: varimi chete', thisassertion of equality contradicts the observ-able reality of social relationships. In effect,this assertion of equality on the basis of presentoccupation is used to deny status differencesbased on past occupation, by farmers in thelower socio-economic categories. Such asser-tions may be viewed as aggressive attempts todeny, even to obliterate, status differentialswhich are becoming increasingly patent.55OCCUPATIONAL MOBILITYAs noted earlier, socio-economic status isrelated both to educational level and, moreimportantly, to occupational background,which is largely dependent on educational level(or has been in the past). By using data onoccupational histories, it is possible to tracepatterns of past occupational mobility amongpresent-day farmers and thereby to indicate,very broadly, the means by which currentsocio-economic differentials may have arisen.Only 7,6 per cent of farmholders havenever been employed, and these men (somestill minors at school) have not been conside-red in respect of occupational mobility. Ofthe remainder, the majority (65,1 per cent)were either horizontally mobile between jobsat the same occupational level, or else wereoccupationally static, being employed in onejob only. However, 27,3 per cent were vertic-ally mobile, moving between different occupa-tional levels. Most of this vertical mobilitywas upwards (22,1 per cent): only 5,2 per centmoved downwards. Thus over one-fifth ofthese present-day farmers showed at leastmoderate success Š and in a few cases, specta-cular success Š in improving their occupa-tional positions during their period of employ-ment. Only 43,4 per cent of all Msengezi farm-ers ended their employment period in unskilledjobs.A number of different 'strategies for mobi-lity' have been used by these farmers. There is,firstly, the 'promotional strategy', whereby anunskilled worker seeks promotion to a super-visory position, in which position he will havereached his ceiling in formal employment.Secondly, there is the 'entrepreneurial stra-tegy', whereby savings from meagre unskilledor supervisory or even skilled employment areused to open a small business, a move whichincreases both income and occupational pres-tige without requiring additional educationalinputs. Thirdly, the 'educational strategy'seems to have been used most frequently:starting at the unskilled level, savings are in-vested in further education including, usually,a teaching qualification. Using one or moreof these strategies, it has been possible for asmall minority of farmers to move from un-skilled, through supervisory, lower professionaland self-employment, to higher professionalor administrative positions in a relatively shortspace of time. Some of these highly successfuloccupational climbers have endeavoured toensure their children a head start in the occu-pational stakes by sending them to private,multiracial boarding schools, a move whichhas the greatest chance of protecting the child-ren against future downward mobility in thesocio-economic sphere.Having considered the past occupationalmobility of the farmers themselves, it is nowpossible to use the occupational level at whichthese men retired from wage employment as abase to assess the occupational mobility oftheir children who, being born prior to 1955,have already entered the labour market. Un-fortunately, the data required to assess the pastoccupational mobility of these 'children' is notavailable, so this analysis must be static in theintra-generational sense: an attempt will bemade to measure only the distance between'children's' present occupational levels and thelevels from which their fathers retired. It isrealised that the 'children's' occupations maywell alter before they themselves retire fromemployment, and that this measurement is,therefore, interim and not final. From thisinterim consideration, however the followingpoints emerge:1. If the father left wage employment atthe unskilled level, his sons are morelikely to achieve higher occupationallevels than are his daughters. This situ-ation results mainly from differentialattention to the education of male andfemale children by fathers of this occu-pational level.2. If the father left wage employment atthe supervisory or skilled level, hisdaughters are more likely to move upthrough higher occupational levels thanarc the daughters of men formerly inunskilled employment; and relativelyfewer of his sons are employed at theunskilled level.3. The pattern of occupational distribu-tion among the daughters of farmerswho retired from lower professionalemployment is very similar to thatfound among daughters of farmersformerly in skilled or supervisory posi-tions. The sons of men formerly inlower professional positions apparentlyhave the greatest chance of attaining56rhigher professional status, and are morelikely to be found in the lower profes-sional stratum than are the sons offarmers in any other occupational cate-gory.4. Almost without exception, both sonsand daughters of farmers who were orare self-employed businessmen havemoved down from their fathers' occu-pational level. The occupational levelsof daughters approximate those fordaughters of skilled/supervisory or lowerprofessional ex-employees, while thepattern for sons' occupations is similarto that for sons of formerly unskilledworkers. Presumably this downwardtrend among the children of the self-employed is related to the fact that themajority of these entrepreneurs haveeducational qualifications of a similarorder to those of previously unskilledor supervisory workers.5. The children of farmers who are orhave been employed in higher profes-sional or administrative jobs have alsoexperienced a general decline in occu-pational status, although the majorityhave not moved below the level of lowerprofessional employment. Only thedaughters of these men seem to standa reasonable chance of maintainingtheir fathers' level in the occupationalstructure, generally as state registerednurses.6. Although fewer daughters have movedupwards from their fathers' positionsthan have sons, they seem to stand agreater chance of achieving higher pro-fessional ranking than do their brothers,since over 60 per cent of those 'children'in this category are women. This situ-ation is primarily a result of the inclu-sion of state registered nurses in thiscategory. The unusual position ofAfrican state registered nurses in SouthAfrica has been analysed by Cheater(1974), and their position in Rhodesiaappears to be essentially similar as aresult of the structural similarity be-tween these two societies.7. Very few farmers' daughters are foundin supervisory/skilled or self-employ-ment, mainly because of restricted op-portunities for women in these fields,and because independent entrepreneur-ship on the part of women has nowell-established precedent in this partof the world, in contrast to the situa-tion in West Africa.8. Sixty-six per cent of daughters andsixty-one per cent of sons of Msengezifarmers are employed at occupationallevels the same as or adjacent to thosefrom which their fathers retired. Overone-third of these farmers' 'children'are thus two or more steps away fromtheir fathers' final occupational classi-fication, which suggests that the circu-lation of persons through the differentoccupational levels is perhaps morerapid than might be expected, and thatthe people under consideration are em-ployed in an open and mobile societyin which ascription of socio-occupa-tional status does not operate, althoughracial status, ascribed by the nationalstructure of Rhodesia, obviously affectsoccupational achievement in an absolu-tely vital manner. This rapid occupa-tional movement, both intra- andinter-generational, is also indicative ofthe lack of a stable class structure (inthe Warnerian sense of classes as statuscategories).9. In aggregate terms, there are remark-able differences in the apparent direc-tions of occupational mobility betweenthe sons and daughters of these farmers.As shown in Table IV, sons haveachieved occupational rankings higherthan the levels from which their fathersretired, four times more frequentlythan have daughters, despite the pointnoted above (point 6) that women attainhigher professional status more frequen-tly than do men. This distinction be-tween the upward movement of men asa category and women possibly as in-dividuals, probably reflects relativelyfixed and conservative male ideas re-garding the proper roles of women insociety.57Table IVDIRECTIONS OF OCCUPATIONAL MOBILITY,FROM THE BASE OF FATHERS'PRE-RETIREMENT LEVEL, OF SONS ANDDAUGHTERS PRESENTLY IN EMPLOYMENTDirection ofOccupational Mobility Sons: % Daughters: %Vertical: UpwardHorizontal: NilVertical: Downward31,342,726,08,056,835,2Occupational mobility, especially in deve-loping states, is heavily dependent on botheconomic development and education. Whilekinship and friendship networks may operateas effective recruitment agencies for relativelyundemanding jobs, skilled and professionalpositions generally require minimum qualifi-cations and open, competitive recruitmentprocedures. In the past, it was relatively easyto manipulate this system indirectly, by utilis-ing personal ties to get one's children intosecondary school. With the increasing bureau-cratisation of the educational system, particu-larly as it affects Africans in Rhodesia, it hasbecome virtually impossible to manipulate thesystem to the advantage of one's children. Onlyby taking his children out of the governmentsystem and sending them to private, multi-racial schools, can a man hope to ensure hischildren against dropping out of the systembefore completing secondary education: suchtactics are open to few. Once the childrenhave completed secondary schooling, furthertraining is possible and it is this training onwhich upward occupational mobility is increas-ingly dependent.Since education is a partial key to occupa-tional mobility, educational differences be-tween farmers and their children should, logi-cally, have been discussed prior to intergene-rational occupational movement. However,the relationship between educational leveland occupation has weakened considerablyin the time span separating these generations,for the expansion of primary education faci-lities to serve a much wider population hasresulted in a general rise in educational levels,such that a primary education is no longera guarantee of unskilled employment, muchless a white-collar job. Thus a farmer with aprimary education may have been a teacherof lower professional status, whereas his son,also with a primary education, may be a ma-chine operator, petrol pump attendant orgeneral labourer. Given this need for increasededucational attainments in order to stand stillin the occupational structure, it is hardly sur-prising that the vast majority of farmers' child-ren are better educated than were theirfathers. Decreased educational attainmentamong the second generation of Msengezi'spopulation is virtually negligible, although itis somewhat more likely to involve girls thanboys.Children of both sexes have increasedchances of reaching the upper levels of secon-dary school, and even university, when theirfathers have had at least a standard one edu-cation (three years of formal schooling) them-selves. In respect of less educated fathers, how-ever, the situation is different as a result ofconservatism, lack of knowledge of how toavail one's children of the available opportu-nities, low levels of ambition for one's children,and financial factors. Children of educatedparents from advantaged home backgrounds,who have access to books, television and par-ental assistance, have a marked advantageover those who have to finish farm tasks be-fore tackling homework by inadequate lightand whose parents cannot provide educationalguidance. This advantage is, however, indirectand mediated by the child's own level of abi-lity and ambition, and there are thus amongchildren no significant differences in educa-tional attainments which are directly depen-dent on the levels of their fathers' schooling.CONCLUSIONIn the light of the widespread opportuni-ties for primary education (there are sevenprimary schools to serve a population of some5 500 in Msengezi), of the competitivescramble for secondary school places beyondthe grade 7 bottleneck (at present only 37,5per cent of grade 7 school-leavers proceed tosecondary school), and of the fact that the bu-reaucratic selection system is not open to pri-vate manipulation, it seems likely that theformation of a stable class structure will bedelayed among black Rhodesians. The educa-tion system will continue to produce a smallelite drawn from widely differing backgrounds,because this system militates against the de-velopment of class 'fixation techniques' on any58significant scale. It therefore seems unlikelythat the proto-classes identified earlier in thispaper will harden into a rigid structure so longas the present educational system is in opera-tion. It is more probable that the elite (socio-economic category A) will continue to be arelatively exclusive social clique in Msengezi,although the clique personnel will alter, beingdrawn from different families at differenttimes. The present patterns of extensive occu-pational differences between adjacent genera-tions may be expected to continue for as longas the educational system retains its presentcontrolled, exclusive character at the secon-dary school level and for as long as new occu-pational opportunities become available toAfricans at a fairly rapid rate. Only when theeducational system at the secondary and tech-nical levels is open and the complete spectrumof occupational opportunities is available toAfricans, does the emergence of a stable classstructure appear feasible in Rhodesian societyas a whole, in rural as well as urban areas.Opportunities for the investigation of socio-economic mobility under present circumstan-ces should, therefore, be taken while thesecircumstances prevail.ACKNOWLEDGEMENTAn earlier version of this article was presented to the Third Rhodesian Science Congressheld in Salisbury in September 1974.The data used in this article were gathered during my tenure as Social Studies Faculty Re-search Fellow in the University of Rhodesia during 1973-4. I am grateful to the Research Boardof the University of Rhodesia for financial assistance towards the costs of the research.REFERENCESCHEATER, A. P. 1974 'A marginal elite? African Registered Nurses in Durban, South Africa', African Studies, 33,143-58.HUGHES, A. J. B. 1974 Development in Rhodesian Tribal Areas: An Overview, Salisbury, Tribal Areas of Rho-desia Research Foundation.HUNT, A. F. 1960 'Native Purchase Area Farmers: An Economic Appraisal', Salisbury, Dept of Native Economicsand Markets (mimeo).LLOYD, P. C. 1966 'Introduction' in The New Elites of Tropical Africa, ed. P. C. Lloyd, London, Oxford Univ.Press.SCHAPERA, I. 1947 Migrant Labour and Tribal Life, London, Oxford Univ. Press.SOROKIN, P. 1959 Social and Cultural Mobility, Glencoe, Free Press.TUDEN, A. and PLOTNICOV, L. 1970 'Introduction' in Social Stratification in Africa, ed. A. Tuden and L.Plotnicov, New York, Collier MacMillan.WATSON, W. 1958 Tribal Cohesion in a Money Economy, Manchester, Manchester Univ. Press.WINTER, E. H. 1955 Bwamba Economy, Kampala, East African Institute of Social Research.5960