Soil fertility in Africa is under pressure as an increasing number of fanners attempt to make a living based on what the land can offer to growing plants. Studies in Africa from about 1989 have focused on different spatial scales, i.e., subcontinental, subnational, and farm. This chapter reviews the results obtained at these three levels and compares methodologies and implications. For N, annual depletion was recorded at all levels at rates of 22 kg ha-1 (sub-Saharan Africa), 112 kg ha" 1 (Kisii District, Kenya), and 71 kg ha~' (average for 26 farms in Kisii, Kakamega, and Embu Districts). If the soil nutrient balance is to become a suitable land quality indicator for wider use as a policy instrument, increased sophistication is required, including data on soil nutrient stocks and availability. The advantage of the nutrient balance approach over traditional rate-response research on fertilizers is that it includes all possible nutrient flows at the spatial scales discussed. A drawback, however, is the lack of hard data on flows that are difficult to measure (leaching, gaseous losses, and erosion), and the fact that the balance comprises several inputs minus the sum of several outputs. Nonetheless, the message comes out clearly that improved soil nutrient management is crucial for maintaining and improving soil productivity in Africa. Soil fertility is not a static feature. On the contrary, it changes constantly and its direction (accumulation or depletion) is determined by the interplay between physical, chemical, biological, and anthropogenic processes. This dynamism also is reflected in terminology such as nutrient cycles, budgets, or balances, referring to inputs and outputs in natural ecosystems and managed agroecosystems, to which nutrients are added and from which nutrients are removed. As the world
