Résumé
Cover crop are frequently used to control the spontaneous species (Lu et al. 2000; Snapp et al. 2005). Soil tillage is another mean to control the spontaneous species by destroying seedlings and decreasing the seedbank (false seedbed technique) (Cordeau et al. 2017; Kanatas et al. 2020), while improving cover crop rate of germination. However, the effects of the combination of these practices on the spontaneous community are poorly documented. Our aim was to understand how different modalities of soil tillage and species of cover crops affect the density and richness of the spontaneous community in a fallow of banana cropping system after five months. The experiment was carried out in Guadeloupe Island, in a banana cropping system during the fallow period preceding banana replantation. Two factors were crossed in a split-plot design composed of five blocks and eight treatments repeated in each block. The main factor was the modality of soil tillage before sowing of the cover crop (M1: coarse tillage, M2: coarse tillage followed by false seedbed technique). The nested factor was the species of cover crop (B: a mixt of Brachiaria decumbens+Brachiaria ruziziensis, C: Crotalaria spectabilis, P: Pueraria phaseoloïdes, T: a reference of spontaneous vegetation). Five months after sowing, in three quadrats per elementary plot, all spontaneous species were identified, the number of individuals of each species was counted and the specific richness calculated. The effects of the treatments on these variables were assessed with nested anovas. Five months after the sowing of the cover crops, we observed that the eight treatments provided communities with different species composition and densities. Densities and specific richness of communities associated with C and T were significantly higher than the ones associated with B and P. This suggests that Brachiaria decumbens+Brachiaria ruziziensis and Pueraria phaseoloïdes are more competitive than Crotalaria spectabilis. In C and T treatments, specific richness and densities were significantly higher in M2 than in M1. This suggests i) that the soil tillage modalities impacts the spontaneous community only when poor competitive cover crops are used, ii) and that the false seedbed technique has promoted the germination of the spontaneous species seeds without efficiently reduce the seedbank, contrary to what had been expected. We concluded that Brachiarias and Pueraria phaseoloïdes were highly efficient to control spontaneous species after five months of growth and that the effects of the modalities of soil tillage depended on the species of cover crop sown.