In Kenya, 70 long-term fertilizer trials were established in which the response of the major annual crops to nitrogen, phosphorus and farmyard manure was tested. As the sites all represent well-defined agro-ecological units, the trial results can be used as a basis for area-specific fertilizer recommendations. Results of 4 years (1987–1990) of fertilizer and manure application to maize at three sites were evaluated. The selected sites are at altitudes of 2020 m (Nitisol), 1160 m (Phaeozem), and 130 m (Alisol). All the sites are in maize-growing environments, but farmers use different hybrids. Analysis of variance revealed that maize on the Nitisol responded vigorously to phosphorus and manure, with even a significant interaction. On the Phaeozem, it was solely nitrogen that limited yields, and on the Alisol there was a response to both nitrogen and phosphorus. At a relatively low level of fertilizer input (Sh. 825–1125), farmers can, in 3 out 4 years, earn an extra Sh. 3000 from the Nitisol (value/cost 4.5), Sh. 4000 from the Phaeozem (value/cost 4.2), but only Sh. 425 from the Alisol (value/cost 1.5). The study clearly shows the need to (i) recommend fertilizers according to the agro-ecological diversity of agricultural land, and (ii) support systems of integrated nutrient management, particularly in areas of low soil fertility, where the farmer has too few economic incentives to rely solely on mineral fertilizers.