To improve C sequestration in no-till soils requires further development of crop rotations with high phytomass-C additions. The objectives of this study were (i) to assess long-term (17 years) contributions of cover crop- or forage-based no-till rotations and their related shoot and root additions to the accumulation of C in bulk and in physical fractions of a subtropical Ferralsol (20-cm depth); and (ii) infer if these rotations promote C sequestration and reach an eventual C saturation level in the soil. A wheat (Triticum aestivum L., winter crop)-soybean (Glycine max (L.) Merr, summer crop) succession was the baseline system. The soil under alfalfa (Medicago sativa L, hay forage) intercropped every three years with maize (Zea mays L., summer crop) had the highest C accumulation (0.44 Mg C ha(-1) year(-1)). The bi-annual rotation of ryegrass (Lolium multiflorum Lam., hay winter forage)-maize-ryegrass-soybean had a soil C sequestration of 0.32 Mg C ha(-1) year(-1). Among the two bi-annual cover crop-based rotations, the vetch (Vicia villosa Roth, winter cover crop)-maize-wheat-soybean rotation added 7.58 Mg C ha(-1) year(-1) as shoot plus root and sequestered 0.28 Mg C ha(-1) year(-1). The counterpart grass-based rotation of oat (Avena strigosa Schreb., winter cover crop)-maize-wheat-soybean sequestered only 0.16 Mg C ha(-1) year(-1), although adding 13% more C (8.56 Mg ha(-1) year(-1)). The vetch legume-based rotation, with a relative conversion factor (RCF) of 0.147, was more efficient in converting biomass C into sequestered soil C than oat grass-based rotation (RCF = 0.057). Soil C stocks showed a close relationship (R(2) = 0.72-0.98, P < 0.10) with root C addition, a poor relationship with total C addition and no relationship with shoot C addition. This suggests a more effective role of root than shoot additions in C accumulation in this no-till soil. Most of the C accumulation took place in the mineral-associated organic matter (71-95%, in the 0-5 cm layer) compared to the particulate organic matter. The asymptotic relationship between root C addition and C stocks in bulk soil and in mineral-associated fraction supports the idea of C saturation. In conclusion, forages or legume cover crops contribute to C sequestration in no-till tropical Ferrasols, and most of this contribution is from roots and stored in the mineral-associated fraction. This combination of soil and rotations can reach an eventual soil C saturation.
