A year-long case study of nutrient cycling within an agro-pastoral community of southern Mali revealed substantial differences in nutrient balances between households and between settlement sub regions of the study area. Inter-household differences should, in great part, bb attributed to differential access to livestock products such as manure and animal traction (both ox-ploughs and donkey carts). Exchanges of manure, carts, and ploughs between owners and non-owners of livestock account for the different levels of access. Exchanges within the settlement sub-regions were more numerous than those between them, had exchanges between e pastoral Fulani and the other communities were limited. The farming systems of the sub-regions presented three patterns of agricultural intensification: 1) cotton-centred rotations sing ploughs and carts to increase yields on perm - anently cultivated sites at the core of the village and in the adjacent hamlets,2 ) cotton-centred rotations on sandiers oils of the village periphery supported largely by intensive hand cultivation with the borrowing of carts to transport nutrients,and3 ) cereal-centred cultivation of small fields heavy manure by large cattle herds owned and managed by Fulani, with plough-ownerships seeming to ensure that household labour could remain available for herding. None of these models of intensification correspond fully with the owner/manager operation of a "mixed farm" and suggest that the potential for better crop-livestock integration in the sub-humidz one may be under-estimated and mis-represented by the "mixed farming" model. Although the overall nitrogen balance of the region after the 1996 cropping year was -8.2 kg N/ha, the different systems described above are associated with significantly different balances:the villa.g1 >. e core and hamlet cores were- 3.2 and- 4.7 kg N/ha, the village periphery- 21.4 kg N/ha, and the Fulani system+ 23.3k g N/ha. Phosphorus and potassium were in positive balance throughout the study area. 0 The different degrees of crop-livestock integration associated with each sub-region suggest that the importance of soil fertility problems must be understood with reference to local algro-pastoral practice. The growing livestock population and the presence of transhumant Fulani pastoralists in the sub-humid zone are key components of these practices and must not be excluded. The access of different actors to key resources such as manure, animal traction, pasture and crop-land deten-nine the ability to increase agricultural production and the degree to which agricultural intensification will mine the soil of nutrients. b ateg eneralisations about the extent of Regional or national studies that make aggregate nutrient loss misrepresent the soil fertility of the region, and risk feeding into an unrealistic" crisis narrative".