Tribalism and Detribalizatioe in Southernand Central Africa*D. H. ReaderDepartment of Sociology, University College of Rhodesia,Salisbury.*Ł *ŁThe literature for Southern and Central Africaon these related concepts is now fairly extensive.A review of some of it, such as is proposed here,indicates that tribalism and detribalization shift inmeaning according to context. Tribalism in thebroadest sense means the condition of living intribes, found widely in Africa when the whiteman first came. The logical opposite of tribalismis non-tribalism; but since the transition is usuallyseen as a gradual and generally incomplete pro-cess, the name detribalization has been applied.It cannot, however, be reduced to any one con-tinuum, for the demographic, social and psycho-logical factors involved may cut across oneanother. Moreover, tribalism can occur in town,detribalization in the country. Such difficultiesrequire care in the definition and use of theseconcepts and suggest an analysis on semantic lines.To take tribalism first, four senses seem usefulaccording to context:1. The state of living in tribes: a generic sense,to which the name tribalism may be given.The quality of living in tribes is tribal,2. Remaining in, keeping in contact with, orreturning to settle in tribal areas. This* This paper was originally written as one of a serieson African Urbanization in a project for the SouthAfrican National Institute for Personnel Research. Iam indebted to former directors, Dr. S. Biesheuvel andDr. D. J. Gouws, for their help and encouragement,and to the South African Council for Scientific andIndustrial Research for permission to publish.sociographic sense, often in the context ofmigration, may be called ruralism. Theadjective would be rural-tied, the transitionruralized.3. Adherence to tribal values and behaviourpatterns. This psycho-sociological dimen-sion may be relevant in town or country,and in view of common usage will be called"tribalism" (in quotation marks). The trans-sitional condition is "tribalization," theadjective "tribalized."4. A category of interaction among Africansin towns. Research in Central Africa (e.g.,Epstein, 1956, 724: Mitchell 1956a, p. 18)has appropriated the generic term tribalismto this contest. Actually a role-category isinvolved which will be designated here astribe. The adjective, if needed, is tribe-categorized.In general terms, tribalism covers all theseaspects, which need only be distinguished by theirspecial names when the context demands it. This,however, is almost invariably the case whenspecific questions about tribalism are asked; andindeed each sense generates its own type of ques-tion:Ruralism: What measurable contacts do urban-dwelling Africans maintain with rural areas?"Tribalism": How far do relationships andloyalties based on tribal ties persist amongAfricans in urban areas?55Tribe: In what social contexts does tribe as acategory of belonging operate in urbanAfrican role-playing?The main purpose of this review paper is tosummarise the locally available literature on suchissues, to underline conclusions and to suggestwhere further research might profitably be done.TRIBALISMRuralismWhile it seems self-evident that contacts main-tained by urban-dwelling Africans with theirrural areas of origin would be a good indicatorof tribalism, the information availabie for South-ern and Central Africa is on the whole sparse.G. B. Wilson in his study of Broken Hill inZambia indicated that the great majority of tem-porarily urbanized African workers there paidoccasional short visits to their rural homes, andthat people from home came to visit them. Anaverage of 30 per cent of the men visited homeeach year (1941, pp. 46-47). For Gwelo in Rho-desia. Schwab contends that even Africans whohave been working for many years in the townhave close links with a reserve and make lengthyvisits to it (1961, 132).For a full account, intervisiting in both direc-tions, urban and rural, should surely be con-sidered. This has been done for South Africa ina study of Baumannville location, Durban, in1955. It was found that during the previoustwelve months more visits had been made fromthe country to this relatively desirable and con-venient city location than had been made to thecountry from it: 246 against 206 visits in anenumeration of 113 households. Frequency ofvisiting in both directions was proportional to thedistance of travel involved. The pattern of inter-visiting was not substantially different for bloodrelatives compared with affinals, nor did increas-ing residence in Baumannville disturb it. Thereasons given for visiting indicated that 79 percent of the visitors to. and 81 per cent of thosefrom, the location had as their objective themaintenance of family ties or the seeking ofsocial contact with kin. Thus rural-tied relation-ships were maintained in a community whoseaverage length of residence in the location wasmore than ten years (Reader and Mann 1955,98-99).Later, also in South Africa, the Border Re-gional Survey of Rhodes University included astudy of periodicity of home visiting in a localmigrant labour situation in East London, CapeProvince. It was shown on a sample basis(N=263) that frequency of home visiting wasdirectly proportional to the distance of travelinvolved, and that three in four of the migrantswere in contact with their rural family and homeonce a month or more (Reader 1961, p. 56).Other evidence of ruralism besides the fre-quency of rural ties has been put forward bydifferent writers. G. B. Wilson gave figures forthe transfer of cash and clothes to the countryby African workers in fulfilment of their familyobligations (1941, p. 40), Van Velsen states thatsuch contributions are regarded by the Tongamigrant from Malawi as a kind of insurancepremium (1961, 237). It might further be main-tained that the proportion of rural to urbanmarriages in a community, for example, 54 percent in Langa location, Cape Town (Levin 1947,p. 9 n.), was broadly diagnostic. The proportionof married workers with their wives in the countrycould also constitute a sociographic indicator ofruralism. In the investigation by the Universityof Natal of Dunlop workers, only 6 per cent ofthe African sample had their wives in Durban,while 14 per cent indicated that they would doso if housing and other facilities existed. But76 per cent were against making their home intown, and 80 per cent said that they would ratherwork in a factory in a reserve (1950, p. 220).The interpretative reliability of such figures,however, depends on the socio-economic context.As G. B. Wilson pointed out, it was the possi-bility of bringing wives to town which mostdifferentiated the situation in Zambia from thaton the Witwatersrand. Of the men over twentyyears of age at Broken Hill, the majority hadtheir wives with them, for it was the policy ofthe copper mines to accept married status, butnot to accommodate any issue of the union (1941,pp. 25, 47). In the Republic of South Africa,on the other hand, stringent conditions of influxcontrol are applied, and migrant labour ratherthan urban family life is encouraged (FaganCommission 1948, p. 33), It would evidently beinvalid to compare figures on wives in townderived from these two situations.Another possible factor for determining rural-ism is the incidence of sending African childrenfrom town to the rural areas. This too, however,is complicated by motivation and context. Hunterin an early work on the Pondo and Xhosa de-scribed the process in both directions. There wasa great deal of "borrowing" and "lending" ofchildren between town and country. Town child-ren were sent to relatives in the country becauseit was cheaper or more healthy for them to livethere. Farm children were sent to relatives in56-> f town to go to school (1932, 682). In a laterstudy of the East London locations by the BorderRegional Survey, it was found that children upto the age of fifteen years constituted two inŁ three of all persons away from the locations.Numbers of children away increased as this agewas approached, when they were considered tobe most susceptible to juvenile delinquency(Reader 1961, p. 46). Among the Marnbwe ofZambia a tribesman goes to the Copperbelt aloneto work and sends back money for his wife to, join him when he is established; but she bringswith her only a child in arms who cannot be leftbehind. Older children usually remain at homein the care of kinsfolk (Watson 1958, pp. 63-64).Finally, the factor of retaining land rights in arural area has received attention from those con-cerned with ruralism. Watson points out that int Zambia land is still the final tribal security. Aslong as rights to the use of land continue todepend on kinship affiliations, kin relationshipswill retain their importance (1958, p. 158).Gluckman similarly states that those who stayat home hold the land as security for support inmoney from those who go out to work (1960,68). Against this, however, is the fact that inthe South African context at least, not everytribesman can obtain land rights. In an analysisof 1,200 African operatives engaged in secondaryindustry and building, it was found that of therural-based part of the sample (52 per cent),40 ner cent had no iand rights in the present orforeseeable future, while an additional 20 per centonly anticipated them (Glass and Biesheuvel1961, p. 6). Clearly in circumstances of landshortage a younger son, for instance, might befully rural-tied, desire land, but fail to obtain it.Yet for a rural population as a whole, retaining> land rights might still be a reliable index ofruralism.It is against this background that Mayer's state-ment for migrants in East London, Cape Pro-vince, must be appraised. He says that themigrants who remain susceptible to the rural pullare the ones who make the most of their oppor-tunities for home visiting, and also for associatingwith fellow "exiles" from the same home-placeduring their stay in town (1962, 584). In effect,this is a two-stage hypothesis. First, it suggeststwo indicators of ruralism. The one of themMayer calls "home-visiting" (1962, 590), and theother "the principle of common home ties" (1964,30). Second, these two indicators together arepostulated as rural-tied criteria which can be usedto determine how "tribalized" (sense 3) a manis: how strong is his adherence to tribal valuesand behaviour patterns.The second stage remains a plausible hypothesisuntil confirmed by other and external criteria of"tribalism" (see below). The first stage, how-ever, since the sociographic factors involved arequantifiable, can be tested by an experimentaldesign. This may be applied to the seven poten-tial variates which have just been reviewed, ofwhich Mayer's are two, namely: (1) home-visit-ing; (2) rural transfer of cash and goods; (3)rural or urban marriage; (4) wife in country ortown; (5) children in country or town; (6) re-tention of rural land rights; and (7) commonhome ties in town.The hypothesis here is that all these variatesload highly on one sociographic dimension-ruralism. We wish to know whether this is true,and if so what the loadings are, so that in futurefield-work questions on the less heavily-loadingvariates may be omitted in favour of the morediagnostic ones. The technique which suggestsitself for this purpose is factor analysis. In orderto avoid the difficulty of uncontrolled backgroundfactors such as family size and distance awayof home area, however, it seems best to seekonly dichotomous responses for each variateŠmarriage rural/non-rural, wife in country/not incountryŠinstead of using continuous distributionswhich would invoke the background factors.The conventional factor-analytic approach withdichotomous distributions would be to make useof derived functions of joint frequencies, i.e.,fourfold point or tetrachoric correlations. Slaterhas suggested another method which operatesdirectly on joint frequencies such as would beobtained by taking every possible combination ofdichotomous responses to seven variates in a seriesof 2 x 2 tables (Slater 1947, 114-127). Apartfrom the saving of labour in this case, any ofthese methods would probably yield similar re-sults, and they could be used to cross-check oneanother (see Appendix I).The outcome, it Is to be hoped, would be ahigh loading for most of the variates on the onedimension, ruralism. A second or further dimen-sion with variates heavily weighted on it wouldbe a disturbing but important find: it wouldindicate that what we were measuring was nothomogeneous. In any event, some answer wouldbe given to the original questionŠwhat measur-able contacts with rural areas urban-dwellingAfricans maintainŠfor the particular populationfrom which the sample was drawn. Starting fromsuch a population, with background factors like57influx control policy held constant, the investiga-tion could be extended to other and differentpopulations. The problem of correlating the basicindicators or ruralism found with the psycho -sociological dimension of "tribalism" will be dis-cussed shortly."Tribalism""How far do relationships and loyalties basedon tribal ties persist among Africans in urbanareas?"Tribal Cohesion The strength with whichAfricans adhere to "tribalism" is usually assessedin the notion of tribal cohesion, a construct in-duced from their behaviour in and out of tribalsituations.By all accounts, tribal cohesion in West andCentral Africa is still strong in spite of somedegree of urbanisation. Indeed, in at least onecase, migration (one of the urbanising media)has led to an intensification of "tribalism": inother words, to "tribalization". This phenomenon,which Rouch calls "super-tribalization," he foundin a study of migrants from the former FrenchWest African territories into Kumasi, Accra andother towns of the Gold Coast. Forde in sum-marising his work has said that entry into urbanlife, "far from weakening the links between ethnicgroups, greatly strengthened them. Solidarity ofpeople of a single tribe and even of a singlevillage increases among these recent migrants, andthe groups so formed become more and moreclosed under their own chiefs. Associations andcults strongly express this solidarity, which isassociated with competition for employment ortrade under conditions of a comparatively freemarket" (Forde 1956, 38-39).In a summary of Busia's work in Sekondi-Takoradi, on the Ghana coast, it is stated thatalthough kinship loyalties are tending to breakdown, town-dwellers generally retain intenseloyalty to home and tribe, as evinced in the popu-larity of tribal associations. This hinders thegrowth of civic, as distinct from tribal, respon-sibility (McCulloch 1956, 86). For CentralAfrica, too, Epstein has said that ". . . examina-tion of the internal organisation of bodies suchas the trade unions and the African NationalCongress would suggest that "tribalism" still re-mains a most important factor; cleavages withinthese bodies are frequently expressed in tribalterms" (1956, 724).In East and South Africa there is a differentemphasis. In Kenya, Parker says that "... thereare definite signs of the passing of tribal cohesion.Prostitution, the large number of juveniles whohave from the outset of urban developmentsought freedom from parental and tribal controlin the towns, and a widening cleavage betweenold and young, provide additional evidence. . . .New forms of social discipline are in a measurereplacing tribal sanctions. Native tribunals in thetowns are of necessity intertribal in constitution. . . and [appear to be working well]." (Mc-Culloch 1956, 130-131). Speaking of SouthAfrica, Hellman says: "Every African knows towhich tribe he belongs, but the strength of triballoyalty varies. ... It is the Government's inten-tion to reinforce tribal loyalties. . . . ArticulateAfrican opinion [however] has condemned . . .[such] measures. . . . The functions of a chiefhave diminished greatly in the country. In a townthey have no relevance. Unlike the situation inBritish West Africa, the tribe has not become thebasis for present-day African political organisa-tions such as the African National Congress. Onthe contrary, the desire of chiefs to retain theirtraditional privileges does . . . conflict with theaims of the African political organisations tosecure the extension to Africans of the democraticrights accorded to Europeans, and tribal alle-giances impede the development of Africannational consciousness (Hellman 1956, 738-739).A comparison of this position with the vigorous"tribalism" analysed by Mayer for the Ciskei,South Africa, indicates that not two differentsenses of "tribalism" but two different contextsare involved, which need to be specified in eachcase. In East London, as Mayer puts it, tradeunions do not transcend tribes: there are nofundamentally different tribes to be transcended.Social opposition appears between those Africansin East London who regard themselves and areregarded as townsmen, and those who are con-sidered as in town but of the country. In thelatter category again, there is still a sharper oppo-sition between the so-called "red" and "school"categories. This reflects a bitter conflict over thedesirability of adopting white people's ways whichhas split the Xhosa for generations (Mayer 1961,p. xiv).Mayer's analysis of the African townsmen inEast London connects with the "juveniles" ofParker and Hellmann's "articulate Africanopinion" respectively for East and South Africa.These Africans are the town-rooted, who can becharacterised as "urban" or in the process ofurbanization. Their main social ties are all con-tained within the town, and in Mayer's words,"the urban traditions are the only ones thatcount for them" (1961, p. xv). The "red" and58If Ł"school" people who are in town but of thecountry are on the contrary migrants, and it isin the different context of migrancy that "tribal-ism" again assumes its full importance.Competent studies in Central Africa suggestnot only that migrants are similarly "incapsu-lated" by tribalism in towns there, but that allthe main participants in the migrant laboursystem have an interest in maintaining tribalcohesion. Watson in his analysis of the Mambweof Zambia says that tribal cohesion and migrantlabour are essential to one another (1958, p.188). Van Velsen for the Tonga of Malawi listsvarious factors working towards the continuanceif not reinforcement of tribal integration. Theseare (1) the basic assumption in the industrialeconomy that the average African is not in anyreal sense a wage-earner but a labour migrantwho has his tribal village to fall back on; (2) thevital interest based on land which Tonga abroadhave in maintaining their position in the com-munity and economy of Tongaland; and (3) theprinciple of "tribal integrity" on which the ad-ministration of the territory is based. Suchconsiderations lead van Velsen to suggest thatall three categories of personsŠindustrial em-ployers, the Administration and the Tonga them-selvesŠhave under present conditions an interestin maintaining tribal cohesion among the Tonga(1961, 239-240). Gluckman makes much thesame point in more general terms when he says:"We see, in short, that tribalism persists in therural areas because of Government support andbecause the tie to tribal land is of the utmostimportance to a man" (1960, 68).These structural data, however, although offer-ing external criteria of tribalism against whichvalidation might be sought, do not provide quan-tifiable material for intercorrelation with thesociographic data of ruralism. Such materialmight be obtained from the psycho-sociologicaldimension of "tribal" social values which onlyMayer so far appears to have analysed. For the"red" Xhosa of East London he posits conscious-ness of a common past (largely expressed in termsof descent from common ancestors) and also apresent duty to maintain the distinctive "tribal-ized" way of life which history and the ancestorshave sanctioned. This preservation is threatenedwhenever Xhosa people take over elements fromthe way of foreign peoples. "Thus the 'tribalism'of the Red Xhosa takes the form of opposingany blurring of intertribal boundaries on thecultural level; any participation in the institutionsof non-Xhosa or sharing of common interestswith them. This opposition is moral in that itis referred to the ancestors. There results theideal of exact coincidence between the boundsof group, culture and history, which may becalled essentially 'tribal'" (Mayer 1961, pp. 40-41).A sequence of dichotomous attitude questionsto test this area can be devised for intercorrela-tion with a sequence on the sociographic dimen-sion of ruralism. A simple questionnaire on theselines for administration to an urban Africanmigrant sample is shown in Appendix 4. Re-sponses on seven variates of "tribalism" wouldthus be returned, which could again be tested forunidimensionality using the first stage of a factoranalysis as suggested by Slater (1947). Assumingthat the one dimension of "tribalism" emerged,a number of statistical techniques could then beused to demonstrate what correlation existed, ifany, between the dimensions of "tribalism" andruralism (see Appendix 2). A suitable measureof association between individual items on thetwo scales might assist in deciding which questionsto eliminate in future field investigations. Finally,partial validation of both dimensions against theexternal structural criteria previously mentionedmight be achieved. These criteria may be tabu-lated against ruralism and "tribalism" as follows:Ruralism1. home visiting2. rural transfer of cash/goods3. rural or urban marriage4. wife in country or town5. children in country or town6. retention of rural land rights7. common home ties in townTable IDIMENSIONS OF TRIBALISMStructural Criteria (validation)1. degree of integration of migrant'swork with home economy2. support given by family in the indus-trial situation3. migrant's participation in politicalactivities at home"Tribalism" (Appendix 2)1. tribal recognition2. common ancestors3. cultural inheritance4. pro-rural way of life5. anti-education6. anti-church membership7. anti-acculturation59For validation purposes the sample populationmight be divided into an upper (more) and lower(less) tribal group on the basis of the structuralcriteria. Each group could then be scored onruralism and "tribalism" respectively, and dif-ferences between the mean scores on each ofthese dimensions could be tested for significanceusing Student's t-test (McNemar 1962, pp. 102-108).Consistently high correlations between ruralismand "tribalism" on different African populationsand in different social contexts could mean thateventually psycho-sociological tribal orientationmight safely be predictable on the basis of a fewsociographic criteria. Such criteria, as Mayersuggests, might reduce to home visiting and com-mon home ties in town (1962, 1964); or theymight not (see Appendix 3)."Tribalism" and Tribe It is not generallyknown that tribe as a category of belonging,although of great significance in Zambia, haslittle meaning in Rhodesia. The indigenous popu-lation of this country consists of two majorgroups, the Ndebele and Shona speaking peoples,the latter being divided into five main clusters ofdialects. The members of each of these groups,although coming from different parts of thecountry between the Limpopo and the Zambezirivers, have lived together for over one hundredyears and so developed their own customs anddialect. Consequently members of a given dialectgroup are not necessarily related by clan ties.For social purposes a tribesman is therefore morelikely to distinguish himself in terms of the terri-torially-defined chiefdom to which he belongs orthe dialect that he uses rather than in terms ofhis tribe.The distinction between the psycho-sociologicaland categorical dimensions of tribalism has beendiscussed more than once in the Southern Africanliterature. Gluckman points out that rural mem-bership of a tribe involves participation in aworking political system and sharing domesticlife with kinsfolk; and this continued participa-tion is based on present social and economicneeds, and not only on conservatism. Tribalismin towns, in Zambia at least, is primarily a meansof mutual classification for the multitude ofheterogeneous Africans who live there (1960, 55).Mitchell and Epstein have underlined the im-portance of the study of "para-tribal" systemsamong Africans: those which exist together with,but outside traditional tribal systems, forexample, in industrial centres (1957, 13).Mitchell's distinction between tribal structure(rural) and tribalism (urban) in another work(1956a, p. 30) suggests that there is no necessarycorrelation between rural "tribalism" and thepara-tribal category of belonging in towns. Ep-stein also indicates that the concept of tribalismat large has two distinct points of reference. Oneis intratribal and refers to the persistence of, orcontinued attachment to, tribal custom. The otherstems from a particular form of social organisa-tion, but operates today in a social system muchwider than that of the tribe. These two aspectsmust be carefully distinguished, says Epstein, sinceit is clear that there may be "revolutionarychanges in custom" while the tribe itself remainsan important category of interaction (1958a, p.231).The question of correlating rural "tribalism"with the paratribal category of belonging in townsseems to be a pseudo-problem. It disappears ifmigrant participation in town and in country lifeis regarded, as it surely should be, in terms ofrole-playing. The reasons why two apparentlydisconnected senses of tribalism are needed todescribe the migrant's behaviour in town andcountry seem to be (a) that this behaviour takesplace in two relatively disconnected social con-texts, urban and rural; (b) that such contexts,quite different culturally, invoke highly differen-tiated role-behaviour appropriate to each, and(c) that nevertheless, many migrants acting outtheir roles in towns often continue to remain, insense 3 of the term, fundamentally "tribalized,"and therefore to use "tribal" modes of role-play-ing where the urban context does not dictateotherwise. This interpretation in terms of roletheory will be used in the analysis which follows.Tribe"In what social contexts does tribe as a cate-gory of belonging operate in urban Africanrole-playing?"In practice the role-category of tribe is signifi-cant both in the migrant and in the settledAfrican townsman context. Mayer points outthat the differences between (in his case) "school"and "red" people must not be overstressed to theextent of suggesting that "school" people haveno particular "tribal" loyalties. Even the town-rooted, if plunged into a multi-tribal setting, mightstill regard some tribal categories as fundament-ally important (Mayer 1961, p. 41). There isevidence from Central Africa too that not onlyin the urban courts do urban Africans regulatemuch of their social behaviour in terms of tribalnorms and values (Epstein 1958a, p. 231), butthat cleavages in urban associations are on occa-60sion ascribed by Africans themselves to tribalantagonism (Epstein 1958b, 104). It is never-theless among temporary urban migrants that thecategory of tribe comes into highest relief, forpurposes of role-identification (Mitchell 1956a,pp. 21-22). Here, too, the problems of mutuallyincompatible role-performances are most acute.Mitchell distinguishes with relevance here be-tween three social contexts on the Copperbeltin which tribal membership has, or had, signi-ficance for the African in town. The first, whichmight be called intra-tribal, involves recognitionamong strangers in an urban location that theycome from the same rural locality or chiefdom.Such men are then linked by a set of roles andrelationships imported with them from their ruralhome. This in-group identification provides abasis for the second type of context, which willbe called inter-tribal. Here ethnic origin as dis-played by language, ties of association and wayof life enables members of other tribal groups tofit their acquaintances into role-categories whichthen determine behaviour. A third context, tobe called extra-tribal, involves tribal membershipas an approach to authority. This led on theCopperbelt to the tribal elder system, which forsome years served to present the African resi-dents' point of view to the mines and to locationsuperintendents (Mitchell 1956a, p. 32).In the intra-tribal context, Mayer indicates thatif a migrant's close friends in town are them-selves country-oriented migrants from the sametribe, his urban relationships are not really"urban" at all, but displaced tribal (1964, 28).This phenomenon of incapsulation implies thatthe migrant continues to play rural roles in townwhenever possible. The principle of commonhome ties which Mayer uses to define this situa-tion is in fact a role-category: one which bearssome resemblance to the "home-fellow" categoryset out by Epstein for Ndola in Zambia (1961.p. 51).From Epstein's work it appears that the home-fellow role category is only one of a number ofcriteria which Africans use to structure their re-lationships in towns. As Mayer has done (1964,30), Epstein classifies "home-fellow" (involvingmutual aid and implicit home-orientation)separately from the more general category of"tribe," which merely role-identifies acquaint-ances on the basis of the "diacritical indices" ofdress, custom, etc. (Epstein 1961, pp. 49-51).Other role-criteria in Ndola are neighbourhoodand locality, kinship and prestige, the latter pro-viding a working model for the incipient classstructure (Epstein 1961, 53).The classical exposition of tribe at the inter-tribal level is still that of Mitchell (1956a; 1956b)."Because tribal characteristics are so easily dis-played in dress, behaviour, and in particular inspeech," he says, "tribalism becomes the mostimportant means whereby day-to-day relation-ships on the Copperbelt are organised" (1956b,p. 6). This is reminiscent of SouthalPs class of"categorical relationships," based on Mitchell(1959, 100-101), in which persons meet in aninformal context without yet knowing one anothervery wel!, and thus intuitively assign one anotherto various type (i.e. role) categories as an em-pirical approach towards appropriate behaviour(Southall 1961, 29). Mitchell emphasises thattribalism (i.e., tribe) does not form the basisfor the organisation of corporate groups. It re-mains essentially a category of interaction incasual social intercourse. Together with theprestige ranking system, it operates to mediatesocial relationships in what is predominantly atransient society. One cannot, moreover, general-ize about the operation of these two "principlesof association" without reference to the specificsocial situation (i.e. context) in which the inter-action takes place (Mitchell 1956a, pp. 42-43).What Mitchell is analysing here is surely roleinteraction in informal urban contexts.More than this, however, by the use of a socialdistance scale Mitchell discovered that the manytribes in town were grouped into a limited numberof categories by town-dwelling Africans, and thatcertain institutionalized behaviours had de-veloped between the various categories (1956a,pp. 22-28). This seems to represent a furtherstage of role-behaviour in which specific stereo-types occur in the role-expectations assigned toeach role. The situation then comes to the fringeof "egocentric relations," in which people knowone another well enough to base their mutualexpectations of behaviour (i.e. role-expectations)on this personal or stereotyped knowledge(Mitchell 1959, 100-101; Southall 1961, 29).The extra-tribal context of tribe can be viewedin two main aspects: the political, in relation tothe white man or other racial group at large;and the industrial, vis-a-vis the employer. Muchhas already been made of the ambivalent politicalrole-playing of the modern chief or headman,representing his people's aspirations on the onepart and the policy implementations of the Ad-ministration on the other (e.g., Epstein 1958a,61p. 65, with refs.; Watson 1958, pp. 188-189).Role expectations for tribal authorities in towncan easily be mistaken by an urban administra-tion, as happened on the Copperbelt mines. Theinstitution of tribal elders, set up by the minesadministration, rested on the assumption, soonout of step with the times, that work roles be-tween urban Africans were also tribal ones (Ep-stein 1958b, 100),It is clear from the literature that in the worksphere in towns, a role-system is involved whichis different both from casual non-work relationsin town and from social relations in the country.The different systems sometimes overlap, as inMayer's contention that participation in thelarger-scale economy as worker involves for therural migrant a simultaneous discharge of a morevalued role as provider (1961, p. 146). But workand rural roles are usually at least disparate.Gluckman quotes the case in Zambia of a chiefvisting town who was treated by African tradeunion leaders with respect, until he tried to inter-vene in an industrial dispute (1960, 69). Mayerstates that the three reference groups which heenumerates for East LondonŠreal townsmen,"school" migrants and "red" migrantsŠhavelittle or no significance at work. White employersexpect all types of Xhosa to play, and the Xhosado play, similar roles in the urban economy(1962, 587). The behaviour of black workers inSouth Africa, once they are industrialized, showsno obvious differences whether their origin ororientation is urban or rural. IndustrializedAfricans of all types in town form a sociallyhomogeneous work group (Y. Glass in Reader1963a).All this must not obscure the fact that thereis an eclectic element in role-selection whichsometimes allows Africans to combine in theirurban role-playing both foreign and tribal ele-ments. At the individual, self-employed level,Mitchell's instance of William, the urban divinerand medicine man, is a case in point. His wearingof a white coat and his case-cards are based onthe high prestige and appearance of Westernmedicine. Yet his practice does not subscribe toWestern scientific principles, but is grounded intribal magic (1960a, p. 18). At the group level,Southall mentions that trade unions in Mombasa,East Africa, were able to use the appeal oftribalism in gaining African support. Here tribewas invoked as a basis for common action inthe work situation. But other special interestsin Mombasa have crossed tribal and racial boun-daries. The printers' union links African andAsian members. The ratepayers' associationbrings together Europeans, Asians, Arabs andColoureds (Southall 1961, 39-40).In sum, tribe is a category of belonging intowns for which the role-playing context mustalways be stated, since Africans can alternate orcombine rural and urban roles. The principleinvolved has been called situational selection(Mayer 1962, 580 and refs.), although role-selection might have been a better name. Thisis determined not only for each social context,but as Mitchell remarks, "the total set of externalimperatives is probably unique for each town"(1960b, 171).To say that more work needs to be done onthe problem of role-playing in tribal and extra-tribal contexts in Central and Southern Arricawould perhaps be presumptuous. At the intra-tribal level in a South African town, Mayer hasdone much of this work under the approach ofmigrancy and town and country rootedness. Hiswork, with or without a role-orientation, couldobviously be replicated to advantage elsewhere.In Central Africa, Gluckman, Mitchell, Epsteinand their associates have explored the inter-triballevel of urban relationships in a way which iseasily translateable into role and context theory.It may, however, be at the level of extra-tribalrole-playing that the richest field of research lies.Africans who today reject their European rulersare not necessarily rejecting the techniques androles of production and government which theserulers introduced (c.f. Mair 1960); yet neitherare they altogether free of the tribalism whichtheir rural background imparts.DETRIBALI ZATIO NIn the same way as with tribalism, there aredifferent dimensions of detribalization as a pro-cess which need to be separated for analysis.While these are broadly the logical opposites oftribalism, two possible senses of detribalizationfall away for practical purposes. The genericterm detribalism would be the state of living outof tribal life. While relevant in some rural areas(e.g., on European farms) as well as being con-comitant with full urbanization in towns, thiscondition is still largely in process of becoming,and is then covered by detribalization. Again,detribalism is not, so far as is known, a categoryof urban belonging. This leaves two dimensionswhich form the logical opposites of ruralism and"tribalism" as previously defined, namely:Immigration: moving from tribal areas, witha lapse of ties with people left behind.62"Detribalization": the dropping or rejection oftribal values and behaviour patterns.These are adaptations of the two meaningsascribed to detribalization by Mitchell (1956c,694-695). In his usage "the lapse of social rela-tionships with people living in tribal areas" isclassed under the psycho-sociological dimensionof "detribalization" together with "the droppingor the rejection of tribal modes of behaviour."This might be justified if sociological analysis ofthe lapsed social relations were intended. Inpractice they have only been used as sociographicindicators of the place of the individual on the" tribalized "/" detribalized " psycho-sociologicalscale. Mayer means by the process of urbanizationamong migrants "a shift in the balance betweenwithin-town ties and extra-town ties . , , [suchthat the] extra-town ties have collectively shrunkto negligible proportions . , ." (1962, 580). Thissurely amounts to a plain count of relationships,and can at best be an index of detribalization,not a definition of urbanization. Another purelysociographic use of rural ties is in their factor-analytic treatment as a variate of rural ism. sug-gested earlier.The dropping of rural ties can also be usedas one of the bases for distinguishing betweenimmigration and migration, as Mayer has done(1961, pp. 5-6). The contrast should strictly bebetween emigration and migration, the formerbeing more appropriate from a detribal stand-point (L. emigrate, to wander forth). However,this term may well be needed in other contextsconnected with urbanization. Hence the wordimmigration, with the selected meaning of "pass-ing into a new habitat" (Shorter Oxford EnglishDictionary), with the implication of permanence,will probably be more generally suitable, and isretained here. Labour migration is seen as themovement from tribal to other places of work,and perhaps back again. With the question ofthe gradual lapse of tribal ties left open, it seemsbest at present to dea! with migration as a specialcase of immigration. It will soon appear thatmigration is the prime means of leaving tribalareas in Central and Southern Africa, and is anecessary condition both for the undeniable over-all increase of urban African populations duringthis century, and for detribalization itself.Using the two dimensions of detribalization inthe present way, the merits and demerits of Gluek-man's well-known statement of detribalizationseems to become plain. In his view, the momentan African crosses his tribal boundary he isdetribalized, outside the tribe, though not outsideits influence (Gluckman, 1960, 58). Such a manis indeed detribalized. in the sociographic sensethat he is a migrant; and so, sociographically, hewill remain until, as Gluckman says (ibid.), hereturns to the political area of his tribe, whenhe becomes (sociographically) tribalized again.But this leaves open the extent of his "detribaliza-tion," if any, on the psycho-sociological dimen-sion. For this, further investigation is needed,as Mayer has suggested (1961), of the degreeto which the migrant has dropped or maintainedhome ties, consorted with new city associates asagainst home-fellows in the urban area, andneglected or arranged to perform certain crucialrural roles in absentia.As with tribalism, each of the two senses ofdetribalization generates its own type of question:Immigration: To what extent do Africansmigrate to urban areas (migration)? Howfar do their ties with rural areas lapse (im-migration)?"Detribalization": What is the process wherebyAfricans change their orientation from thetribal to another way of life?The available Central and Southern Africanliterature on these issues will again be reviewed.ImmigrationThe question "to what extent do Africansmigrate" is deliberately broad, since "to whatextent" is meant to cover the questions "how inview of influx control, and at what rate?"How in view of influx control? Official policyin South Africa for many years has been toadmit Africans into an urban area when there iswork for them; to exclude them when there isnot (Reader, 1961, p. 53). In addition, Havemannhas contended that influx control is also resortedto as a means of reducing or slowing down thedemand for African housing (1951, 7). This inturn might be reduced to a desire to keep outpeople who would be a drain on local municipalresources. Havemann points out that such a wishis not directed against Africans alone. When the"civilised labour" policy (employment of poorwhites rather than blacks) and the 1930 depres-sion jointly produced an influx of poor whitesinto South African urban areas, some munici-palities protested vigorously, and would certainlyhave controlled the influx had they been able todo so. Nor is influx control, says Havemann, apeculiarly South African expedient. It has closeparallels in the Settlement Law of seventeenth-century England, and in the anti-settlement lawsof twentieth-century America (eventually de-clared ultra vires the Constitution). In both these63precedents the reasons for migration and themotives for keeping out the migrants were muchas in the South African case (Havemann 1951, 8).Southall gives a useful brief review of influxcontrol in Central Africa (1961, 10). In Rhodesia,as in South Africa, legislation restricts Africanresidence in towns to land set aside for thepurpose, and, for the migrant, to the actual dura-tion of employment. Gussman's case-study ofBulawayo (1952, p. 24) and comparable evidencefrom Salisbury show that municipalities have onthe whole been unable to meet their Africanhousing obligations. Africans are forced to con-travene the law by the mere fact of their resi-dence, working women are obliged to findaccommodation in married or single men'squarters, and married couples have to share withsingle men. Similar regulations obtained inZambia, but were ineffective because of theabsence of physical barriers to movement. Ac-commodation on the Copperbelt mines wasusually adequate, but in towns like Lusaka, illicitcompounds run by European land-holders re-lieved the sheer physical pressure on accommo-dation. Botswana according to Schapera did notpractise influx control (1947, p. 85). It has sincedone so, but rather with a view to discouragingsettlement or transit on political grounds thanfrom any desire to control work-seekers.* Furtherafield, Southall reports that Kenya, Tanzania andthe Congo have all had unsatisfactory experienceswith influx control, largely on grounds of econo-mic pressure, administrative cost, or the impossi-bility of preventing evasion (Southail 1961, 10).The total picture is one of irresistible pressureby African labour upon urban centres in Centraland Southern Africa. From the African's view-point, as Havemann says, influx control means"sticking to his job lest he be expelled from townwhen he is unemployed, and staying in town lesthe be debarred from entry if he should go homefor any length of time. In these circumstances,the African worker naturally brings his familyto town, or as close to it as he can get. . . .To the extent that the influx has been stemmedat the city borders, it has produced great blackbelts outside . . ." (Havemann 1951, 7). In so faras basically rural-oriented workers have broughttheir families with them, this suggests caution inthe use of "family urban-dwelling" and "persist-ence in same job" as sociographic criteria ofpsycho-sociological detribalization.* African Immigration Proclamation, No. 14 of 1958;this required that each African entrant into Botswanashould obtain a visitor's or residential permit.At what rate do Africans migrate? The geo-graphers Green and Fair (1962, p. 80) foundthat the shifts of the African population on thewhole continent were not suitable for mappingbecause of the general unreliability of the de-mographic data and the absence of a full censusfor Africans in the Federation of Rhodesia andNyasaland. However, in the Republic of SouthAfrica the shift of Africans from rural to urbanoccupations "has been generally more spectacularthan even that of the Europeans" (Green andFair 1962, p. 81). In 1946 there were 1,856,028Bantu in South African urban areas. The increaseup to 2,328,534 in 1951 was attributed to aninflux from the European farms and other ruralareas (40 per cent), the Native Reserves (8 percent) and neighbouring territories (23 per cent),and natural increase (29 per cent). By 1960,although only 30 per cent of the total Africanpopulation were in urban areas, this represented3,192,130 Bantu compared with 2,461,162 urbanEuropeans. It has been calculated that if thetempo of African demographic urbanization ex-perienced during the period from 1946 to 1951is maintained to the close of the century, morethan 10,000,000 out of a total Bantu populationof 21,000,000 will then be established in urbanareas (Tomlinson Commission 1955, p. 29).In Rhodesia, the African population in 1956was 2,550,000 against 177,124 whites. Between1936 and 1956 the total increase in the numberof Africans in urban employment was 355,656.Of these, 44 per cent were accounted for in sixmain urban centres. In Zambia, the Africans in1956 totalled 2,110,000 against 65,277 Europeans.In that country there was a total increase ofurban-employed Africans amounting to 122,356between 1946 and 1956. Of these, 98 per centwere found on the Copperbelt and along the lineof rail extending southwards to Livingstone.Green and Fair conclude from their data that byfar the most significant factor in the distributionof population in Southern Africa today is themajor urban centre, and its surrounding clusterof lesser urban areas and comparatively denserural population (1962, p. 81).The general answer to the original question-to what extent do Africans migrateŠis thus thatthey migrate to urban centres overwhelminglyand in spite of all efforts to stop them. Thisleaves quite open, however, the second socio-graphic question: How far do their ties withrural areas lapse? (immigration). Hellmann(1948, p. 110), writing on a Johannesburg Africanslumyard, lists three criteria of detribalization64which bear directly on the present question ofimmigration:1. Permanent residence in an area other thanthat of the chief to whom a man wouldnormally pay allegiance.2. Complete severance of the relationship tothe chief.3. Independence of rural relatives both forsupport during periods of unemploymentand ill-health or for the performance ofceremonies connected with the major crisesof life.This list raises the important point that a mandoes not have to be sociog.raphica.ily urbanizedin order to be detribalized. On all three of thesecriteria African squatters or farm servants onEuropean farms in South Africa could be com-pletely detribalizedŠout of a tribal area andsevered from all tribal relationshipsŠwithout anydirect urban contacts at all. The farmer, to thegreater or less extent that his race attitudesallowed, would become their "father" and theirchief. In the Ciskei (Cape Province) of the1930s, Hunter reported that "Practically all those[Africans] resident on farms are servants per-manently employed and having no stake in anyreserve. . . . Families on farms keep their clanand tribal names, but are members of no tribalcommunity and under no chief. Many individualsdo not know the name of the reigning chief oftheir tribe" (1961, p. 505).The extent of contact with the reserves, shegoes on, naturally varies with the date at whichthe family left the reserve, and with the proximityof the farm to the reserve. Kinship bonds withthose outside the farm are, however, weakenedby lack of leisure, restrictions which hinder visit-ing and the summoning of relatives to ritualkillings of stock, and the fact that relatives canoften only get employment at a considerabledistance from one another (1961, p. 523). More-over, permanent immigration of farm workers tothe urban, areas has been the subject of complaintby farmers over the last twenty years at least.It is now virtually impossible for an Africanfarm worker to migrate legally to town with hisfamily (Roberts 1958, pp. 81-82). Yet, as pre-vious figures have shown, many do succeed inimmigrating.In town it is often hard to determine howimmigrant a man is in terms of Hellmann'scriteria. Hunter, working in the locations ofGrahamstown and East London, Cape Province,found that "about half the community are per-manent town-dwellers" (1934, 336). In Rooiyard,Hellmann considered that the great majority wereliving in town temporarily. She assessed this bythe fact that many left their cattle in the ruralareas or had from time to time sent articles offurniture home, "But it must be borne in mind,"she said, "that the Rooiyard populace is com-paratively young (few couples had adult child-ren) and it is possible that during the courseof years their attitude may change and they maycome to be permanent town-dwellers" (Hellmann1948, p. 111). This raises the difficulty of anyassessment of an urban African's intentions basedsolely on the length of time he has already spentin town. Mitchell's demographic notion of"stabilization" (1956c), while useful as a measureof urban exposure, has been criticised as a pre-dictor of future urban behaviour (Reader 1963b).The concept of "permanent residence" worksbest in conjunction with Hellmann's other criteria,which bear on it. She found that the relationshipto the chief varies from individual to individual.Some went readily with gifts to urban meetingsorganised. to welcome a visiting chief; othersmerely shrugged their shoulders. However indif-ferent they seemed in town, though, many Afri-cans visited their chief when in the rural area,as a mark of respect and allegiance. The de-pendence of land-owning Africans on the chiefrequires no emphasis (Hellmann 1948, p. 111).Among the Mambwe of Zambia, as Watsonpoints out, the chief retained his authority, withwhich land tenure and. tribal cohesion are inti-mately bound up, throughout the whole period ofdirect and indirect rule and the development ofindustry (Watson 1958, p. 5). Richards speakssimilarly of the Bemba (1935, p. 21), whileamong the Ngoni of Malawi, who have a highrate of labour migration (Read 1938, p. 56),and the Zambian Ngoni (Barnes 1954, p. 135)there is marked tribal cohesion and. a strongpolitical organisation.But it was in the dependence of the RooiyardAfrican on his rural relatives that his unbroken,although weakened connection with tribal lifebecame most apparent. Of the 40 families con-stituting the relatively settled population of thisnow defunct slumyard, only 17 had all theirchildren with them. Apart from one childlesscouple, all the others had one or more childrenexposed to tribal influence by their upbringingin the country under the care of rural relatives.These 22 families had necessarily to maintaincontact with their home and tribe of origin.Hence tribal child-rearing formed an effectivecheck to the total immigration either of the65Table IICRITERIA OF IMMIGRATIONimplies negative criteria of RuraiismImmigration1. permanent residence in an area other than the rural Šplace of origin2. severance of relationship with rural chief giving up rural land rights3. independence of rural relatives no home visitingno rural transfer of cash/goodswife in townchildren in town4. inler-tribal marriage Šchildren or their parents (Hellmann 1948, pp.111-112).Hellmann also suggests that the process ofimmigration is aided by inter-tribal marriage.Among 52 couples with husband and wife be-longing to different tribes, only one completefamily and eight wives went home. In 48 couplesmarried intra-tribally, 10 complete families and12 wives went home (presumably in the periodof investigation). It seemed that the men werebeing cut off from their tribal background in inter-tribal marriages more than the women, whotended to remain linked to their kin (Hellmann1948, p. 112).Four criteria of immigration thus emerge,which can be arranged in the above table.Whether to use a low score on ruraiism or ahigh score on immigration as a sociographic in-dicator of movement between the tribal-detribalpolarities is a matter for practical investigationin field situations. A priori, it might be better totake the positive variates of ruraiism than thenegative ones of immigration. There is also thepoint that "permanent residence" is hard todetermine. If scores on both these dimensionscould be taken, intercorrelation might be carriedout to determine their degree of overlap. Itcould then happen that a combined score wouldserve best for measuring sociographic detribali-zation (Appendix 4)."Detribalization""What is the process whereby Africans changetheir orientation from the tribal to another wayof life?"The definition of "detribalization" involves thedropping or rejection of tribal values and be-haviour patterns. This is meant, however, to heedthe caution once expressed by Fortes againstassociating the term with evaluations such as"pathological," "disintegrated" or "demoralized"(1938, 61). It does not suggest that tribal normsbecome unconsciously the standards againstwhich urban behaviour is assessed (Mitchell1959, 100-101). All that is implied is that (a)urban or non-tribal behaviour is contrasted withtribal behaviour; (b) in cases where apparentlypermanent behavioural changes take place, tribalbehaviour of some kind was the area of de-parture; and (c) if and when a psychologicalchance of orientation occurs, it is from a rurallife-outlook. All this may happen along econo-mic, sociological and psychological sub-dimen-sions of "detribalization," which, although theydo not warrant special names at present, will betreated separately in analysis.The economic sab-dimension. Migration cer-tainly does not imply "detribalization." But, aswill be seen below, it is usually migration forprimarily economic reasons which makes a manleave his tribal area at all, and thus exposes himto the non-tribal influences which make "de-tribalisation" a possible choice. Moreover, deci-sions for migration may trigger off those towardsimmigration and hence detribalisation,Mitchell has summarized along three axes thefactors motivating migration from rural areas(1958, 17):1. Through the economic system. All writerson the sociography of migration have em-phasized its economic basis.2. Through the normative system of a society.Migration becomes a habit, perhaps even arite de passage: the expected type of be-haviour for young men. An appropriaterole for a young man is thus to leave thetribal area and make his way into the outerindustrial world.3. Through the personal choice of individuals.Many reasons may be obtained from thestatements of informantsŠto escape quar-rels, witchcraft, avoid arduous duties, etc.As Mitchell says, however, "Several writershave pointed out that we should be cautiousof the technique which asks the direct ques-tion to the migrant [regarding the reasons66for migration]." He quotes Richards: ". . .a particularly angry scene with his [themigrant's] local chief may have becomedramatized in his mind for all time andquite obliterate consciousness of a longseries of economic frustrations and hard-ships which were equally 'motives fortravel'" (Mitchell 1958, 17-18).A review of sources other than those examinedby Mitchell shows that this tabulation is basicallycorrect, with the addition of "propaganda andadvertisement" which in the discussion followinghis paper, he said that he had not sufficientlyconsidered (1958, 22).PRESSURES TO1. ECONOMIC Š General:Inability of reserves to support populationEconomic necessity and pressure:Diverse money wants:Table HILEAVE RURAL AREAS TO WORKG. B. Wilson 1941, p. 55; South African Institute of RaceRelations 1947, 65; Reader 1961, p. 36.Phillips 1938, p. 55; Schapera 1947, p. 121; Houghton1958,40.Janisch 1942, p. 23; Schapera 1947, p. 141; South AfricanInstitute of Race Relations 1947, 65; Eberhardt 1949, p. 50;Mitchell 1954, p. 10; Watson 1958, p. 39; Gluckman1960, 65; Mayer 1961, pp. 91, 94.Special Wants:Cash for taxes and levies:Marriage payments:Clothes:Trader debts:Court fines:To build homestead:2. NORMATIVEExpected behaviour:Initiation into manhood:For urban experience:3. PERSONALEscape from domestic problems,consequences of offences, etc:Desire for adventure and change,dullness of tribal life:Urban-experienced men preferred by girls:4. POLITICALPropaganda and political pressure:Janisch 1942, p. 23; Schapera 1947, p. 141; South AfricanInstitute of Race Relations 1947, 65; Watson 1958, p. 40.Janisch 1942, p. 23; Watson 1958, p. 40; Mayer 1961,p. 91.G. B. Wilson 1942, p. 18; Schapera 1947. p. 142; Watson1958, p. 40.South African Institule of Race Relations 1947, 65;Watson 1958, p. 40.Watson 1958, p. 40.Mayer 1961, p. 91.Schapera 1947, p. 115: Watson 1958, p. 40; Mayer 1961,p. 91.Schapera 1947, p. 116; Mayer 1961, p. 91.Watson 1958, p. 43; Reader 1961, p. 36.Schapera 1947, pp. 118, 120; Watson 1958, p. 47.Schapera 1947, pp. 115, 117.Schapera 1947, p. 116.Schapera 1947, pp. 144-155.Mitchell concludes that economic factors, whilebeing a necessary condition of labour migration,may not also be sufficient. Additional "last-straw"causes (Gulliver 1955, p. 28) may be needed toprecipitate the migration. The rate of migrationmay be determined by economic conditions, suchas the degree of subsistence possible in the tribalarea and the level of wants created by westerncontact; but the incidence of migration may de-pend on psychological conditions (Mitchell 1958,18-19).In South Africa in 1958 a rural investigationwas conducted by the National Institute for Per-sonnel Research on migrant labour in three dist-ricts of the Transkei, Cape Province. Besides con-firming the inadequacy of local food production,and investigating other matters confidential to thesponsor, the Institute found that there were twomajor ultimate needs in migrancy: cattle andmoney. The cattle-need was fundamental andcould never be fully satisfied. Money needsvaried greatly according to whether the migrantwas, on his own assessment, "red" (blanketed)or "dressed". The former, in a gross majority inthis sample, were working primarily to maintaina standard of living dictated by the rural sub-sistence economy. Any surplus money left afterpaying debts and satisfying needs of the typeshown under Special Wants in Table III, theyput into stock.It was found, furthermore, that economic needsdriving the migrant fell into three broad classes,each with a different degree of urgency. The firstcentred about cattle, the second and third aboutmoney:1. The long-term desire for cattle, analogous tothe European's wish for wealth and security.Indeed, the African's cattle kraal is sometimesreferred to nowadays as his 'bank'.2. Fairly long-term accumulating needs, usuallycyclic in character. The main one of these isthe growing need for maize as the harvest iseaten off. The need for new blankets to wearas the old ones become ragged is anotherexample.3. Sudden, short-term needs, requiring immediatesatisfaction. The legal attachment of a beastat the dipping-tank when taxes have not beenpaid, or the sudden sickness of a family mem-ber, are examples. These may correspond onoccasion to Gulliver's "last-straw" causes(Gulliver 1955, p. 28).Economic needs can easily change in typeaccording to context. Cattle for lobolo is a goodexample. A young man first going to work, stillwith no permanent sweetheart in view, will see thisas the first type of need: something wanted in thefuture but not pressing. When he begins to "gosteady", however, the cycle of obtaining so manyhead of cattle comes before him as the secondtype of need. When, finally, the in-laws demandcertain beasts before they will allow the weddingto proceed, the man is faced with the third typeof need for which in the last resort he may haveto go out and work at once.If, as often happens, the "red" migrant experi-ences these work pressures at a time of social,economic or ritual activities at home, he will doeverything possible to put off going until it is moreconvenient. He may hold a beer-drink to raisemoney, sell a beast or small stock, or even borrowmoney from a man who has recently returnedfrom work. In the end, when he has nothing todispose of and needs press even harder, he maygo to the local trader for an advance or interest-bearing loan (Reader: unpublished work for theNational Institute for Personnel Research).By such devices, marked seasonal fluctuationsof migrant labour occur in South Africa, on thewhole to the advantage of the migrant. In theBorder Regional Survey of the Ciskei, Cape Pro-vince, it was clear that at least four major factorscontributed to migrant labour fluctuations,weekly, monthly and seasonally. These wereweekly or monthly commuting home, lengthierreturns to plough or to rest, seasonal fluctuationsof demand from various sectors of the economy,and the generally accepted holiday intermissions.But it was also found that the availability of mig-rant workers was tied to seasonal fluctuations intheir own rural economy. During the harvestperiod, April to October, for example, less labourwas available in the local city of East London. Itwas hard to resist the conclusion that althoughlabour was plentiful in the area, it was labour thatultimately called the tune (Reader 1961. pp.56-59).These are evidently some of the economic con-comitants of rural rootedness. In the Transkeiansample of the National Institute for PersonnelResearch, however, a notable though smallergroup were "dressed" instead of "red". To wearEuropean clothes alone is not, of course, neces-sarily to be town-rooted. The readiness of"dressed" people to accept new ideas does suggest,however, that they are more likely to becometown-rooted that "red" people are, and that wherechange is taking place, these people will be in thevanguard of it. Moreover, many of the "dressed"men were industrially rather than migrant-oriented, in the sense that they tended to comehome only after long periods of work, and thenonly during recognised factory recesses. It wasnoted of such workers that even to work in a citylike Cape Town had the effect of strengthening aman's urban outlook if he were already "dressed",or of inducing a "dressed" outlook if he were not(Reader: unpublished work for the NationalInstitute for Personnel Research).To pursue this line of research further requiresanalysis at the level of the sociological and psy-chological sub-dimension of "detribalization".While migration alone secures only sociographicdetribalization, its economic co-ordinates maywell be necessary conditions for "detribalization"to happen.The sociological sub-dimension. Watson statesthat the concept of "detribalization" impliesthat an African must choose between two systemsof social relations and values: one based onmodern industrial production, the other on tradi-tional subsistence production. The choice, how-ever, is not always final. An African can keepthese different spheres distinct and separate. Hemay move from his tribal area to town, engagein paid labour and take part in social and econo-mic activities there. On his return to the reserve,however, he will quickly resume tribal ways, forthe good reason that his claims to the use of landare bound up with the continued existence oftribal society (Watson 1958, p. 6). This is evi-dently the general position of the migrant labourerin Southern Africa.Nevertheless, translation of this situation intorole and context theory suggests that while true,something more is wanted for present purposes.There is more than one social context open to theman who leaves the tribal milieu. Through lackof land or the prospect of it, he may go as asquatter on European farms. This, as has beenindicated, implies at least a degree of detribaliza-tion on all counts, for he has burnt his boatsalmost as thoroughly as if he had immigrated totown. Only closeness to the tribal area of originmay encourage the retention of some rural ties.But the roles which the man and his family willplay are new: servant-roles towards the Europeanfarmer and his family and neighbour- and fellow-worker roles towards the other farm-servants.Even in town, however, there are a number ofdifferent contexts in which the migrant or immi-grant may role-play. On the mines the compound-context encourages "tribal" roles as far as isfeasible in an urban industrial context. The con-text in secondary industry, as has emerged,demands highly specific work-roles whose recipro-cal rights and obligations tend to converge forworkers everywhere. In East London, Cape Pro-vince, workers tended to act out rural residentialroles in shack accommodation, suburban roles inmunicipal houses (Reader 1961, p. 125 et passim).There are also the three sub-contexts describedunder tribe, and no doubt many others.The whole question in terms of "detribalization"is, under what conditions do immigrant Africanspermanently reject the rural context of origin,and associated contexts and roles, and choose toact out appropriate roles in urban or other non-tribal contexts?Mayer, who has preferred the "social network"model, has discussed this question for Xhosaimmigrants in East London (Mayer 1961, 1964).The answer seems to be largely one of finding outwhat factors overcome the natural pull of thecountry home. On the face of it, the rural pull haseverything in its favour. "In South Africa," asMayer says, "there is an obvious lack of securityin town . . . One has only to think of two pointsŠthe denial of freehold rights to Africans intowns and the regulations [of influx control]allowing them to be expelled if they do notremain in employment. In that situation . . .any South African migrant with land rights inthe country, however poor, has good reason tosee his rural home as a sheet anchor, a sole in-surance against the emergencies of want and oldage" (Mayer 1964, 23).Hellmann, however, does not believe that suchinsecurities and grievances "are instrumental incausing the Native to leave an urban area. Whenhe does return to the country he returns with hiseyes opened to a stern reality, but he does notreturn because of that reality" (1948, p. 113).In terms of "network" theory, Mayer statesthat if a man's new ties in town come to havesufficient moral content, he will have becomepersonally rooted in town and will think he is"at home" there (1964, 24). Mayer's meaningof "moral," however, is only clear for the rural-rooted, where it implies "referred to the an-cestors" (1961, p. 41). What is the moralreference for the town-rooted? Whatever it is,it is apparently expressed through the socio-graphic shift in balance from extra-town towithin-town ties which Mayer calls the processof urbanization. But how is this shift in turn tobe evaluated? Is a man with only two remaininghome ties more detribalized than a man with six?This evidently depends upon (a) the connected-ness (Bott 1957, p. 59) of the network of whicheach is part; (b) the intensity of the ties, theirstrength, and the willingness with which theparties are prepared to forego other considera-tions in carrying out the obligations associatedwith these ties; (c) the regularity with whichcontacts person-to-person are made; and (d) therelative statuses of Ego and his interactors. Notsociographic description and balance, but full-scale sociological analysis would be required inevaluating "detribalization" in this way.At the behavioural level, role and context theorywould seem to give an account of "detribaliza-tion" which is more clearly separated from ques-tions of motivation and orientation on thepsychological dimension. In role theory, a manis ruralized and "tribalized" if he persistently playsrural tribal roles in both tribal and non-tribalcontexts, unless obliged by external authority toplay other roles. He is to some degree detribalizedif in the same contexts he plays non-tribal roleswhen he might have played rural tribal ones.The verification here does not consist entirely inasking him about his extra- or within-town ties,although this will be valuable in understanding69his role-playing. It consists in observing andanalysing how he behaves in specific social con-texts.In this process the classification of contexts isobviously important, and a knowledge of thesocial structures of communities will assist. Totake the institution of marriage as an example,in exogamous corporate kin groups such as arefound in tribal areas, the conjugal partner whomarries into the group is in a special positioncompared with individualized marriage in urbanareas. Politically, jurally and ritually, she (or hein a matrilineal system) remains affiliated withthe natal corporate kin group, although assimi-lated to the affinal group by marriage. Segrega-tion of conjugal roles is then likely to be evenmore marked than the most extreme casesdescribed for urban families with close-knit net-works (Bott 1957, pp. 99-100). Thus we havethe situation as analysed by Mayer for the Ciskei,in which "red" peasants do not think it right tobring their wives to town at all (1961, p. 210)and will spend months away from them withoutdistress. At a more "detribalized" level is Ep-stein's case of the African trade-union leader whomarried paying bride-wealth and still maintainstribal marital obligations, but requires a con-siderable re-ordering of his domestic life in townin order, inter alia, to entertain importantdelegates (1958a, p. 237). Still more "detribal-ized" are evidently marriages in town, where inmost cases "two of the most characteristic fea-tures of Bantu marriageŠthe gradual rapproche-ment of the families of contacting parties and theinvocation of the dead ancestors to sanction theunionŠare absent" (Hellman 1948, p. 84). Judg-ing by Bott's material on families in London,England, it would, however, be dangerous to gobeyond this and evaluate "detribalization" interms of degrees of marital co-operation, at leastwithout careful consideration of network andcontext.Role theory does not, then, attempt to answerthe question of why Africans permanently rejectrural contexts of origin. It only answers, in termsof behaviour, how and when they do so. Surelyno sociological theory can answer the why ques-tions except in terms of its behavioural end-effects. This question is necessarily relayed backto the psychological level for investigation.The psychological sub-dimension. As Kuperhas pointed out for West Africa (Kuper 1965.p. 19), we do not know the psychological con-comitants of African migration, with or withouturbanization. We should be prepared to find,however, that migrating Africans reject tribalism,not primarily because of complex re-orientationsof mental process, but because they like some-thing else better. Writing as a westernized Afri-can, Makulu speaks of the gloomy, monotonous,unattractive and desolate aspect of present-dayAfrican tribal life. Villages are depleted of able-bodied men away on migrant labour, leavingbehind them old men, women and children. "Inmany parts of Africa," he says, "disintegrationof rural communities starts at the point whereareas of industrialization are accessible to thevillages. Men leave their villages and go to theindustrial centres. At first they are merely"campers," as they still maintain their footholdin the village. They return home during timesof ploughing. At this stage the economy of therural communities is not terribly disrupted. Thesecond phase follows when men begin to bringtheir families [to town]. ... At first they con-tinue to keep contact with their villages by sendingwives home occasionally, but this phase soonpasses and home-returning becomes irregular.Families find their place in the new centres wheretempting town life offers easy life. Families growup in towns . . . and urbanization thereforebecomes the order of the day" (Makulu 1960.11).As suggested previously, one necessary condi-tion of "detribalization" may be migration.Another condition would seem to be westerniza-tion, whose values may be defined as to do andbe done unto a? white people do and are doneunto. Even on the farms, where the industrializa-tion of agriculture has to some extent also takenplace, African squatters are affected by westerni-zation. M. Wilson reports them as saying, "Wewho have lived among Europeans cannot standthe hardships of the Xhosa country" (1961, p.507), and this despite the fact that, "It is clearthat the "farm Native" has lost economically bycontact with Europeans. Working very muchharder than he did under tribal conditions, hehas no more nourishing or varied a diet thanthe rawest Pondo on the reserves . . ." (Wilson1961, p. 517). With "school" or "dressed" people,however, those who have opted for Westernvalues, the full process of detribalization stillrequires some additional commitment over andabove a desire to be westernized. Hence Mayer'sdistinction between "school" migrants, who re-turn to the reserves, and "real townsmen" whodo not (1961, p. 4).The extra commitment required has evidentlyto do with what Mayer calls town-rootedness.70By the town-rooted he means the core of (inthis case) East Londoners whose homes androots are in town only. Children who are bornand grow up in town are naturally town-rooted,but this is not the only way to qualify (Mayer1961, p. 5). Here emerges, however, an extremecase of the sociographic variate "permanent re-sidence in an area other than the rural place oforigin" which was used to determine detribaliza-tion. There must presumably be an equivalentterm "farm-rooted" to describe those born andbrought up in a family of farm servants, thoughhere the situation may be against their free voli-tion. But what is other-than-tribal-rootednessover and above sociographic factors such as"town-born" which exhibit it? "Country-bornpeople can become town-rooted too," says Mayer,"by deciding to stay on permanently and becom-ing incorporated in the town community. Theyare then immigrants rather than migrants" (1961,p. 5).The investigation of this factor of decision-making should not be confused with an apparentlysimilar issue previously raised in connection withMitchell's use of the term "stabilization" (Mitchell1956c; Reader 1963b). There the question waswhether a sociographic index could predict afuture intention at the psychological level to re-main in town. Here too the issue at first seemsto be whether, at the psychological level, what aman says now regarding his non-tribal intentionscan be held to apply or to bind him in the yearsto come, especially after he retires from work.This, however, is surely a pseudo-problem inrelation to "detribaiization." If an African statesthat he has decided to remain in town for certainreasons, and this is confirmed by his havingrenounced tribal land-rights, discontinued hiscontacts with relatives in the rural area of origin,and so on, then surely to all intents and purposeshe is detribalized, and in this case town-rootedtoo. The further question of what rural valuesand behaviour elements he retains in a state ofdetribalization would have to be decided by care-ful questioning and examination of the way inwhich he plays his various urban roles. Thepossibility that in the future he may becometired of or disgusted with town life, begin nego-tiations for a rural plot of land and retire tothe country, merely indicates that he could become"tribaiized" again. To regard the tribal-detribalpolarities at this stage as joined by an irreversiblecontinuum is surely unjustifiably to limit theirusefulness and application.If these points are accepted, the way to investi-gate other-than-tribal-rooted decision-making be-comes fairly evident. A series of focusedcase-histories might be taken, stratified on age,sex, education, tribe and any other factors heldto be relevant to the decision-making function.The focus would be upon (a) the broad periodover which the critical decision, or decision-sequence, was taken, in relation to the presentage of the respondent; and (b) the reasons andpredisposing factors leading up to the criticaldecision. We should not anticipate that the rea-sons given will be other than at the level of broadcommonsense, except for maladjusted individuals.These reasons may even be downright hedonistic.The decision to remain in town for a tribalAfrican may be at the level of any comprehensivepreference the world over.To sum up this review of tribalism and de-tribalization in Centra] and Southern Africa asa whole, it would seem that there has been toomuch reliance so far on sociographic factors ofruralism and immigration alone to indicate howmatters stand on the psycho-sociological dimen-sion. These sociographic variates need to bedrawn together compactly by factor analysis, andthe non-significant among them discarded. Atthe sociological level the problem to solve is nolonger that of migration, which has been wellcovered, but that of (permanent) immigrationwhich has not. Here more work in terms of"network" and of role and context theory mighthelp to show how a man handles urban situationsand makes decisions in them. Co-operation withsocial psychologists is greatly to be recommendedin this field. Research along the lines of theattitude questionnaire in Appendix 4 and thecase-history approach suggested above mightindicate how and why non-tribal decisions aremade. It should be pointed out, however, thata low score on the "tribalism" part of the ques-tionnaire will not by itself mark "detribalization."It will show only a readiness to accept non-tribalinfluences which is characteristic of the "school"or "dressed" person. The factor of decision fornon-tribal rootedness has to be separately assessedthrough the individual case-history.It is to be hoped that sufficient emphasis hasbeen laid in this account on detribalization as aprocess. As Mayer says, "The valid objectionto detribalization as a working concept, after all,is not that it implies change, but that it tends toimply synchronized change of the whole man.Provided we . . . allow for his perhaps remaining'tribal' in some situations while becoming 'urban'in others . . . there is no a priori objection to'Ł i71using the idea of change. Indeed, the difficultyof accommodating movement, or process, hasalready been noticed as a main limitation of anymodel that works solely in terms of situationalselection" (1962, 589). It seems clear thatAfricans in Central and Southern Africa, apartfrom the degree to which they are being urban-ized and westernized, are also undergoing ageneral process of detribalization. The "alterna-tion" model, as Mayer calls it (1962, 579), hasbeen valuable for initial differentiation of thedouble roles of migrants in town and country,at the sociographic level. Used in isolation fromthe psycho-sociological dimension of "detribaliza-tion" and the case-history of decision-making andchoice, it could prove a barrier to further under-standing.APPENDICES1. Following Slater, the procedure suggested is:(a) Joint frequencies of response on the variates canbe studied by means of Chi-squared tests on 2 x 2tables for each variate cross-tabulated with everyother variate.(b) To avoid the difficulty of attaching undue valueto extreme chi-squares, an overall test of non-association can be obtained by summing the indi-vidual chi-squares and testing with N degrees offreedom, where N is the number of 2 x 2 tables.(c) If the overall test proves significant, a test ofwhether the items cluster can be obtained bycarrying out the first step (i.e., the test for acommon factor) of the method suggested bySlater for factor analysis of a set of 2 x 2 tables.I am obliged to Mr. R. S. Hall and Prof. J. M,Schepers of the National Institute for Personnel Researchfor these suggestions and interpretations.2. A suitable procedure might be:(i) Complete the first-stage factor analysis of variateson "tribalism" according to Slater (1947).(ii) Rank-order the items in ascending order of mag-nitude, rejecting those of loading less than, say,.3, where N>100.(Hi) For each subject calculate a total score (0/1)based on the items retained in each of the twodimensions ruralism and "tribalism." Thus eachS will have two total scores.(iv) Using these scores, calculate the product momentcorrelation between ruralism and "tribalism."With so few items it will probably be necessaryto correct for coarse grouping, using Sheppard'scorrection (vide McNemar 1962, p. 24).(v) To achieve the best measure of association be-tween individual items on the two scales, thebiserial phi technique may be used. This allowsof correlation between truly dichotomous re-sponses, as on the ruralism scale, and maskedcontinuous ones, as on the "tribalism" scale (videThorndike 1949, pp. 168-169).Prof. J. M. Schepers and Mr. A. O. H. Roberts of theNational Institute for Personnel Research kindly madethese suggestions.3. Since completing this article I have received a copyof Grant, G. C. 1969 The Urban-Rural Scale: ASocio-Cultural Measure of Individual Urbanization,In this work, initially undertaken quite independentlyof mine, Grant has assembled a pool of 52 itemshypothesized as measuring urbanization. The dataobtained by administering these items to a sampleof 100 Zulu industrial workers were subjected to twoforms of analysis: an item analysis for reliability ofthe instrument and a principal factor analysis. Thelatter yielded six factors, which were rotated by thevarimax procedure. Factors U, V, X and Y (1, 2, 4and 5) clearly imply various aspects of urbanization.But factors W and Z seem to be more associated with"tribalism," while factors U and V have overtonesof detribalization. Place of birth, ownership of ruralland and place of residence of parents, and allegianceto chief, loaded particularly heavily on the principalfactor, U.Questionnaire on Tribalism4. Ruralism (sociographic)1. Home visitingŠ1.1 In the last year did you visit relatives in therural areas? YES/NO1.2 Did relatives from the rural areas visit you?YES/NO2. Rural transfer of cash I goodsŠ2.1 During last year did you send cash or goods tothe rural areas? YES/NO(N.B.ŠFor the above questions "last year" inthe case of fairly recent arrivals may betaken as the year going back from thedate of interview. For very recent arri-vals to whom this cannot apply, the abovequestions should be marked N/AŠnotapplicable.)3. Rural or urban marriageŠ3.1 Were you married in the country or in an urbanarea? RURAL/URBAN/N/A4. Wife in country or townŠ4.1 Is your wife now living in the country or withyou in town? COUNTRY /TOWNNot married/Elsewhere (Not applicable)5. Children in country or townŠ5.1 How many living children have you?No children (not applicable)5.2 How many are at your country home?5.3 How many are living with you in town?COUNTRY/TOWN(constructed)5.4 How many are living elsewhere?6. Retention of rural land rightsŠ6.1 Do you have land in the rural area? YES/NO6.2 Do you expect to get land in the rural area inthe future? YES/NO6.3 Do your parents have land in the rural area?YES/NO7. Common home ties in townŠ7.1 When not at work, do you try to meet only withhome-fellows in town? YES/NO72"Tribalism" (psycho-sociological)8. Common way of lifeŠ8.1 Do you think that Africans should be distinguishedby their tribes/places of birthŠ8.1.1. in the rural areas? YES/NO8.1.2 in towns? YES/NO8.2 Do you feel close to other members of yourtribe because you have common ancestors withthem? YES/NO8.3 Do you say that the ancestors have handed downcustoms of vour tribe which you must preserve?YES/NO8.4 Would you oppose for yourself any customs whichdid not agree with your tribal way of life?YES/NO9. ConservatismŠ9.1 Are you willing to be educated?9.2 Are you willing to be receivedYES/NOinto churchmembership? YES/NODo educated men of different tribes have any-thing in common? (Probe if necessary.)YES/NOREFERENCESBARNES, J. A. 1954 Politics in a Changing Society: a Political History of the Fort Jameson Ngoni, London, OxfordUniversity Press.BOTT, E. 1957 Family and Social Network; Roles, Norms, and External Relationships in Ordinary Urban Families.London, Tavistock Publications.BUSIA, K. A. 1950 Report on a Social Survey of Sekondi-Takoradi. London, Crown Agents. Quoted by M. McCul-loch in: INTERNATIONAL AFRICAN INSTITUTE, LONDON. 1956 Social Implications of Industrializationand Urbanization in Africa South of the Sahara. Paris, UNESCO.EBERHARDT, J. L. 1949 "Survey of housing and family conditions (Orlando Township). . . ." Unpublished M.A.Thesis, University of the Witwatersrand, Johannesburg.ECONOMIC COMMISSION FOR AFRICA. 1962 Report of the Workshop on Urbanization in Africa. Addis Ababa.EPSTEIN, A. L. 1956 An Outline of the Political Structure of an African Urban Community on the Copperbelt ofNorthern Rhodesia. In: INTERNATIONAL AFRICAN INSTITUTE, LONDON, Social Implications of Indus-trialization and Urbanization in Africa South of the Sahara. Paris, UNESCO.1958a Politics in an Urban African Community. Manchester, Manchester University Press forRhodes-Livingstone Institute.1958b Tribal Elders to Trade Unions (Northern Rhodesia). In: SMITH, P. ed., Africa in Transi-tion. London, Reinhardt.1961 The Network and Urban Social Organisation. Human Problems in British Central Africa,Rhodes-Livingstone J., 29, 29-62.FAGAN COMMISSION; see SOUTH AFRICA 1948.FORDE, D. 1956 Social Aspects of Urbanization and Industrialization in Africa: a Genera! Review. In: INTERNA-NATIONAL AFRICAN INSTITUTE, LONDON, Social Implications of Industrialization in Africa South of theSahara. Paris, UNESCO.FORTES, M. 1938 Culture Contact as a Dynamic Process: an Investigation in the Northern Territories of the GoldCoast. In: INTERNATIONAL INSTITUTE OF AFRICAN LANGUAGES AND CULTURES, MEMORAN-DUM 15. Methods of Study of Culture Contact in Africa. London, Oxford University Press.GLASS, Y. Quoted in: READER, D. H. 1963 African and Afro-European Research: a Summary of Previously Unpub-lished Findings in the National Institute for Personnel Research. Psychol. Afr., 10, 1-18.GLASS, Y. and BIESHEUVEL, S. 1961 "Urbanization among an Industrial Labour Force." Paper read at CCTA/CSA Conference on Urbanization, Abidjan.GLUCKMAN, M. 1960 Tribalism in Modern British Central Africa. Cah. Etud. afr., 1, 55-70.GRANT, G. V. 1969 The Urban-rural Scale: A Socio-cultural Measure of Individual Urbanization. Johannesburg,National Institute for Personnel Research, South African Council for Scientific and Industrial Research, Pers. 118.GREEN, L. P. and FAIR, T. J. 1962 Development in Africa: a Study in Regional Analysis with Special Reference toSouthern Africa. Johannesburg, Witwatersrand University Press.GULLIVER, P. H. 1955 Labour Migration in a Rural Economy: a Study of the Ngoni and Ndendeuli of SouthernTanganyika. Kampala, East African Institute of Social Research, East African Studies No. 6.GUSSMAN, B. W. 1952 African Life in an Urban Area: a Study of the African Population of Bulawayo, part I.Bulawayo, The Federation of African Welfare Societies in Southern Rhodesia.HAVEMANN, E, A. E. 1951 The Principles Underlying Urban Native Administration in the Union. Race Relat., 18,2-15.HELLMAN, E. 1948 Rooiyard: a Sociological Survey of an Urban Native Slumyard. Cape Town, Oxford Uni-versity Press for Rhodes-Livingstone Institute, Paper No. 13.1956 The Development of Social Groupings among Urban Africans in the Union of South Africa. In:INTERNATIONAL AFRICAN INSTITUTE, LONDON, Social Implications of Industrialization and Urbanizationin Africa South of the Sahara. Paris, UNESCO.HOUGHTON, D. H. 1958 Migrant Labour. In: SMITH P. ed., Africa in Transition. London, Reinhardt.HUNTER, M.: see WILSON, M.73JANISCH, M. 1942 A Study of African Incomes and Expenditure in 987 Families in Johannesburg. Johannesburg,Native Affairs Department.KUPER, H. ed. 1965 Urbanization and Migration in West Africa. Berkeley, University of California Press for Afri-can Studies Center, University of California, Introduction.LEVIN, R. J947 Marriage in Langa Native Location. Cape Town, University of Cape Town, Communications fromthe School of African Studies, New Series No. 17.McCULLOCH, M. 1956 Survey of Recent and Current Field Studies on the Social Effects of Economic Developmentin Inter-tropical Africa. In: INTERNATIONAL AFRICAN INSTITUTE, LONDON Social Implications ofIndustrialization and Urbanization in Africa South of the Sahara. Paris, UNESCO.McNEMAR, Q. 1962 Psychological Statistics, 3rd edit. New York, Wiley.MAIR, L. P. 1960 Social Change in Africa. Int. Affairs, 36, 447456.MAKULU, H. F. 1960 "Pattern of rural-urban development in Africa." In: WORLD COUNCIL OF CHURCHES,"Background information for church and society." Geneva, Mimeographed Circular, 25, 1.MAYER, U. P. 1961 Townsmen or Tribesmen: Conservatism and the Process of Urbanization in a South African City.Cape Town, Oxford University Press for Rhodes University, The Institute of Social and Economic Research.MAYER, U. P. 1962 Migrancy and the Study of Africans in Towns. Am. Anthrop., 64, 576-592.1964 Labour Migrancy and the Social Network. In: HOLLEMAN, J. F. et al. eds., Problems ofTransition: Proceedings of the Social Sciences Research Conference 1962. Pietermaritzburg, Natal University Pressfor Institute for Social Research, University of Natal, Durban.MITCHELL, J. C. 1954 African Urbanization in Ndola and Luanshya. Lusaka, Rhodes-Livingstone Institute, Com-munication No. 6.1956a The Kalela Dance, . . . Manchester, Manchester University Press for Rhodes-Livingstone In-stitute, Paper No. 27.1956b Africans in Industrial Towns in Northern Rhodesia. Oxford University Press. H.R.H. TheDuke of Edinburgh's Study Conference No. I.1956c Urbanization, Detribalization and Stabilization in Southern Africa: a Problem of Definitionand Measurement. In: INTERNATIONAL AFRICAN INSTITUTE, LONDON Social Implications of Industriali-zation and Urbanization in Africa South of the Sahara. Paris, UNESCO.1958 Factors Motivating Migration from Rural Areas. In: APTHORPE, R. J. ed., Present Inter-relations in Central African Rural and Urban Life: Proceedings of the Eleventh Conference of the Rhodes-Living-stone Institute for Social Research, Lusaka, 1958. Lusaka, Rhodes-Livingstone Institute.1959 The Study of African Urban Social Structure. In: COMMISSION FOR TECHNICAL CO-OPERATION IN AFRICA SOUTH OF THE SAHARA Housing and Urbanization: Second Session of the Inter-African Conference, 1959, Nairobi. London, CCTA, Publication No. 47.1960a Tribalism and the Plural Society. Inaugural Lecture, Salisbury, University College of Rho-desia and Nyasaland. London, Oxford University Press.1960b The Anthropological Study of Urban Communities. Afr. Stud., 19, 169-172.MITCHELL. J. C. and EPSTEIN, A. L. 1957 Power and Prestige among Africans in Northern Rhodesia. Proc.Trans. Rhod. scient. Ass., 45, 13-26.PARKER, M. "Political and social aspects of the development of municipal government in Kenya, with special refer-ence to Nairobi." Quoted by M. McCulloch, in: INTERNATIONAL AFRICAN INSTITUTE, LONDON 1956Social Implications of Industrialization and Urbanization in Africa South of the Sahara. Paris, UNESCO.PHILLIPS, R. E. 1938 The Bantu in the City: a Study of Cultural Adjustment on the Witwatersrand. [Lovedaie.C.P.] Lovedale Press.READ, M. 1938 Native Standards of Living and African Culture Change. . . . London, Oxford University Press forInternational Institute of African Languages and Cultures.READER, D. H. 1961 The Black Man's Portion: History, Demography and Living Conditions in the Native Loca-tions of East London, Cape Province. Cape Town, Oxford University Press for Rhodes University, The Instituteof Social and Economic Research.1963a African and Afro-European Research: a Summary of Previously Unpublished Findings in theNational Institute for Personnel Research. Psychol. Afr., 10, 1-18.1963b Demographic Stabilization in the East London Locations, Cape Province. S. Afr. J. Sci.,59, 269-272.READER, D. H. and MANN, J. W. 1955 Marriage, Family Life and Children. In: HALLENBECK, W. C. ed., TheBaumannvitte Community: a Study of the First African Family Location in Durban. Durban, University of Natal.Institute for Social Research.RICHARDS, A. I. 1935 Tribal Government in Transition: the Balemba of North-Eastern Rhodesia. J. Afr. Soc, 34,Supplement to No. 137, 3-26.ROBERTS, M. 1958 Labour in the Farm Economy. Johannesburg, South African Institute of Race Relations.ROUGH, J. "Unpublished findings on migrant peoples in Kumasi, Accra and other towns of the Gold Coast, reportedto Unesco Conference at Abidjan 1954." Quoted by D. Forde, in: INTERNATIONAL AFRICAN INSTITUTE,LONDON 1956 Social Implications of Industrialization and Urbanization in Africa South of the Sahara. ParisUNESCO.74SCHAPERA, I. 1947 Migrant Labour and Tribal Life: a Study of Conditions in the Bechuanaland Protectorate, Lon-don, Oxford University Press.SCHWAB, W, B. 1961 Social Stratification in Gwelo. In: SOUTHALL, A. ed., Social Change in Modern Africa,London, Oxford University Press for International African Institute.SLATER, P. 1947 The Factor Analysis of a Matrix of 2 x 2 Tables. Roy. Stat. Soc. h, 9, Supplement, 114-127.SOUTH AFRICA 1948 Report of the Native Laws Commission, 1946-1948. Pretoria, Government Printer for De-partment of Native Affairs, U.G. 28-1948.1955 Summary of the Report of the Commission for the Socio-economic Development of the Bantuwithin the Union of South Africa. Pretoria, Government Printer, U.G. 61-1955.SOUTHALL, A. ed. 1961 Social Change in Modern Africa. London, Oxford University Press, Introductory Summary.SOUTH AFRICAN INSTITUTE OF RACE RELATIONS 1947 Evidence Submitted to the Native Laws Commissionof Enquiry: Survey of Urban and Rural Areas, Parts I and II. Race Relat., 14, 25-78.THORNDIKE, R. L. 1949 Personnel Selection; Test and Measurement Techniques. New York, Wiley.TOMLINSON COMMISSION; see SOUTH AFRICA 1955.UNIVERSITY OF NATAL, DEPARTMENT OF ECONOMICS 1950 The African Factory Worker: a Sample Studyof the Life and Labour of the Urban African Worker. Cape Town, Oxford University Press, Durban EconomicResearch Committee, Report No. 2.VAN DEN BERGHE, P. L. 1964 Caneville: the Social Structure of a South African Town. Middleton (Conn.),Wesleyan University Press.VAN VELSEN, J. 1961 Labour Migration as a Positive Factor in the Continuity of Tonga Tribal Society. In:SOUTHALL, A. ed., Social Change in Modern Africa. London, Oxford University Press for International AfricanInstitute.WATSON, W. 1958 Tribal Cohesion in a Money Economy: a Study of the Mambwe People of Northern Rhodesia,Manchester, Manchester University Press for Rhodes-Livingstone Institute.WILSON, G. B. 1941 An Essay on the Economics of Detribalization in Northern Rhodesia, Part 1. Livingstone,Rhodes-Livingstone Institute, Paper No. 5.1942 Part H. Livingstone, Rhodes-Livingstone Institute, Paper No. 6.WILSON (nee HUNTER), M. 1932 Results of Culture Contact on the Pondo and Xhosa Family. S. Afr, 1. Set., 29,681-686.1934 Methods of Study of Culture Contact. Africa, 7, 335-350.1961 Reaction to Conquest: Effects of Contact with Europeans on the Pondo of SouthAfrica, 2nd edit. London, Oxford University Press for International African Institute.75