Within 1 to 3 weeks after the application of liquid livestock feces, the soil benzyloxycarbonyl‐l‐phenylalanyl‐l‐leucinase (z‐FLase) and caseinase activities in upland fields (120 t feces ha−1 year−1, 600 t feces ha−1 year−1) increased to the maximum levels (z‐FLase 562 pKat and 840 pKat g−1 dry soil, respectively; caseinase 1,018 pKat and 2,141 pKat g−1 dry soil, respectively), and decreased to the original level after 5 weeks. Numbers of culturable bacteria counted on albumin medium (AA), bacteria counted on peptone‐polymixin medium (PP), and vegetative cells of Bacillus spp. counted on BTV medium (BTV) also increased to the maximum levels after application of the feces. Although there was a significantly high correlation between bacterial number on AA and both the soil protease activities (z‐FLase, r=0.895; caseinase, r=0.935), more significant correlations were observed between bacterial number on AA and inorganic nitrogen contents (NH4+ content, r=0.902; NO3− content, r=0.981). Similarly, although a significant correlation was observed between the soil caseinase activity and bacterial number on PP (r=0.835) and BTV (r=0.780), and the number of fluorescent Pseudomonas spp. (P1) (r=0.867), more significant correlations were observed between the numbers of these culturable bacteria and the soil NH4+ content (PP, r=0.944; BTV, r=0.959; P1, r=0.933) and soil NO3− content (PP, r=0.982; BTV, r=0.961; P1, r=0.990). Therefore it was difficult to identify the bacterial group responsible for the soil protease production based on these correlation analyses. Rapid increase in the total bacterial number (AA) by the 600 t application was mainly due to the increased number of bacteria on PP medium, which were assumed to have been introduced from liquid livestock feces.