Introduction Cattle are the basis of livelihood for pastoralists and agropastoralists in the northern parts of West Africa. They are used for milk, manure, meat and draught power, and they serve as savings and insurance. The composition of pastoral herds, particularly cattle, has often been analysed. Usually, the sex and age groups of individual herds, counted directly or assessed from interviews with the owner or herder, are recorded. Conclusions are then drawn from the herd structure on breeding and commercial strategies, as well as on the main goals of production and market integration. Results indicate that these parameters vary, depending on the specialisation of production in ecological niches or regions (Amanor 1995; FDLPCS 1992a, b; Sutter 1987; Vabi 1993). Cattle herds in the arid parts of West Africa are more specialised in the production of young stock, and their owners are mainly nomadic pastoralists. The herds contain a high percentage of reproductive females (over 45 percent), with a calving rate higher than 60 percent, and a commercial off-take of at least 11 percent. To lower mortality in a risky environment, these pastoralists sell young stock, particularly males, under the age of three. In contrast, herds in the semi-arid parts of West Africa belong to transhumant pastoralists or agropastoralists. Living in a more resource-endowed and less risky environment, such herds have a lower percentage of reproductive females (30-40 percent), a low calving rate, and a larger number of adult males over the age of six (24-63 percent). Besides milk, the production of mature males of heavy weight is a production goal that is related to favourable conditions of pasture and water and the farmers' demand for draught animals (Amanor 1995; Vabi 1993). Data on livestock numbers are generally hard to obtain in northern Nigeria. Collecting data on herd size by counting or interviews is problematic. First, pastoralists are reluctant to give information about their livestock numbers, partly due to cultural reasons and partly for fear of taxation (Gefu 1992; Mohammed and Bello 1994). Secondly, management practices to cope with risk in an ecologically and economically highly variable environment exacerbate the problems of getting accurate data on livestock ownership. Such practices lead to an overlapping of ownership and management patterns, hence, the term 'herd' applies to the unit of management, rather than to that of ownership. This is due to: * the widespread exchange of animals within a social network; * the splitting of herds into management units which might be herded far away from the owner's homestead; * the herding of animals of different (absentee) owners in one herd; * and the hired herder's reluctance to admit to being just a 'poor' employee who has to herd for others. This study attempts to describe size and composition of the herds counted along transects on natural range. Counting domestic livestock gives no information on ownership patterns, but provides information on herding management on a spatial scale and on the existing stocking density. The method of animal ground counts along transects had been established mainly in the field of wildlife studies for the assessment of game populations in inaccessible areas (Norton-Griffiths 1978). It is hoped that this provides some insights regarding herd management in a forest reserve in northern Nigeria. The Study Site The Zamfara Forest Reserve (6[degrees]30' to 7[degrees]15'E, 12[degrees]10' to 13[degrees]05'N) is located in the north of Zamfara State, which shares borders with the Republic of Niger to the north, Sokoto State to the west and Katsina State to the east (Figure 1). Annual rainfall ranges from 500 mm in the north to 850 mm in the south with considerable inter-annual variations (CV 22 to 32 percent). The vegetation is of a northern Sudan savannah type. [FIGURE 1 OMITTED] In 1919 the area was declared a forest reserve by the British, and it became unlawful to pasture domestic herbivores or cultivate the soil (Forestry Law of 1938, 1963 version). …