This chapter was written as the introduction to a French edition of Meggitt's Lineage system of the Mae-Enga (1965a), an edition that never eventuated. I had become concerned with the lack of attention paid to demographic factors in the classic models of African lineage systems proposed by Evans-Pritchard and Fortes and their pupils (e.g. Fortes 1953). The continually segmenting lineages they postulated seem to me to entail a population that doubled or trebled in every generation, yet most of the other aspects of their models of African societies appeared to be based on the assumption of a steady state. I thought that a model of lineage structure was required that was more plausibly diachronic, that allowed for the disappearance of segments as well as their formation, that took account of the availability of land, and that was linked to whatever historical evidence was available. Peters’ analysis of the genealogies of Cyrenaican Bedouin and Lebanese Muslims (Peters i960, 1963) provided empirical evidence for the kind of model I had in mind. I made a preliminary sally in that direction in a review of Meggitt's book (Barnes 1967a), for both the ethnographic data presented by Meggitt and his analysis of them seemed to fit well with Peters’ work. I saw the invitation to write an introduction to the book as an opportunity to develop a more fully specified model, taking account of a very perceptive paper by Kelly (1968) which placed the Mae Enga in the broader context of other Highland societies.