Résumé
Soil degradation is a major challenge in Sub-Saharan Africa, where integrated soil fertility management has been promoted to restore productivity. A long-term experiment (1972–1992) run in Benin consisted of two phases: a depletion phase (1972–1980) with varying levels of mineral and organic fertilisation, and a regeneration phase (1981–1992) where all plots received full fertilisation and organic matter additions. Soils were sampled at 0–20 cm depth in 1973, 1974, 1982, and 1989 to assess fertility changes. Mineral fertilisation (N, P, K) and plant biomass management (crop residue retention and biomass additions) significantly influenced seed cotton and maize grain yields during the depletion phase. Soil organic carbon declined consistently in all treatments during depletion but remained stable during regeneration. The long-term effect was evident only in seed cotton yield during depletion. In contrast, due to high variability, maize grain yield showed no consistent trend. The combined use of organic resources and mineral fertilisers helped maintain crop productivity but led to declining soil chemical properties in this Ferralsol. The analysis of this outdated yet unpublished dataset shed light on how long-term soil depletion effects persist over time, even when soil fertility management is restored, indicating a sort of 'soil memory'. The persistence of these effect suggests that regenerative interventions must begin before critical thresholds of degradation are crossed. Future research should focus on alternative measures to restore/maintain soil fertility not evaluated in this experiment, such as conservation tillage or legume integration, to provide long-term benefits for smallholder farmers facing soil fertility challenges.