Effects of a three-year maize-soybean rotation on soil bacterial diversity and fumonisin-producing Fusarium spp. inoculum potential.

Maize-soybean succession dominates grain production in tropical Brazil, and rotation is often recommended to improve soil health. However, few field studies have followed both soil bacterial diversity and fumonisin-producing pathogens over several seasons in tropical no-till systems. We compared continuous maize, continuous soybean and maize-soybean rotation for three seasons (2016-2017 to 2018-2019). In a fourth season (2019-2020), maize was grown in all plots to test legacy effects. Bacterial communities were profiled by 16 S rRNA (V3-V4) amplicon sequencing. The fum1 gene of fumonisin-producing Fusarium spp. was quantified by qPCR in soil and in maize stalks left on the soil surface. In the first three seasons, rotation did not detectably increase bacterial diversity, although statistical power was low (n = 4 plots; post-hoc power 12-61%). Community composition differed among seasons and was associated with antecedent precipitation as Actinobacteria were most abundant in the driest season. In the fourth season, a continuous-soybean history was associated with higher bacterial richness and Shannon diversity in the maize rhizosphere at tasselling than a rotation history, whereas bulk-soil diversity did not differ among land uses. A rotation history was associated with lower fum1 abundance in bulk soil. In contrast, continuous maize was associated with lower fum1 in stalks (third season), less stalk rot and fewer infected grains. Soil and crop residues therefore responded differently to rotation. Given the single site and limited plot-level replication, these findings should be regarded as hypothesis-generating.