It is necessary to assess the responses of microbial communities and respiration to the revegetation of eroded soils for understanding the dynamics of soil carbon (C) pools and fluxes. In this study, three typical abandoned croplands (CL1, CL2 and CL3) and three secondary grasslands planted with Coronilla varia (GL1, GL2 and GL3)
on the Loess Plateau of China were selected for sampling, and quantitative polymerase chain reaction (qPCR) and high-throughput sequencing were applied to intuitively discern differences in the soil bacteria and fungi. Our results showed that bacterial abundance in the abandoned croplands was 57 times higher than that of the secondary grasslands (P < 0.05), but no obvious changes (P > 0.05) in fungal abundance and microbial diversity were observed after 31 years of revegetation. We observed positive responses in Actinobacteria, Firmicutes, Zygomycota and Ciliophora and negative responses in Bacteroidetes and Planctomycetes to revegetation. In addition, the maximum soil microbial respiration was observed in the GL3 site (20.86 ± 0.69 mg CO2-C kg-1 soil d-1) followed by the GL1 site (19.97 ± 0.65 mg CO2-C kg-1 soil d-1), so revegetation significantly improved (P < 0.05) soil microbial respiration. Multiple stepwise regression analysis showed that dissolved organic carbon (DOC) explained up to 68.5% of the variation in soil microbial respiration, which indicated that the effects of changes in microbial properties in response to revegetation on soil microbial respiration were likely to be smaller than the potential effects of changes in the quality of organic matter. Labile organic matter is the primary rate-limiting factor for soil microbial respiration.