AbstractRecent criticisms of demographic theory and methodology have pointed,inter alia, to the need to take on board advances in cognate disciplines. This case study of the Gwembe Valley Tonga of Zambia highlights the important methodological contribution that social anthropology can make to the sub-discipline of family demography. It provides evidence of the empirical invalidity of the ‘family nucleation’ paradigm, which holds that a shift towards conjugal marriage and nuclear household residence patterns is an inevitable consequence of globalisation, and a precursor of the social and economic changes necessary for the fertility transition. According to nucleation theory, evidence of increased conjugality will be reflected in the reduced symbolic importance of the lineage and ancestors; greater marriage stability; the demise of polygyny and widow inheritance; reduction in the size and significance of bridewealth payments; increasing age of first marriage for women, and decreasing age differentials between spouses. The case study therefore focuses on recent changes in the matrilineal kinship system, notably the emergence of localised lineages, and the extent to which these changes reflect family nucleation (largely but not exclusively in terms of increased conjugality). By analysing household structure and marital residence patterns, including the role played by the husband/father in family affairs, nuptiality (notably bridewealth, divorce and polygyny), inheritance and the role of ancestors, the case study demonstrates that changes in the family and kinship structure in response to local social and economic transformation can be equated not with nuclearisation but with the emergence of a modified form of family and kinship, quite distinct in structure and meaning from the nuclear family.