The short-time of six pure herbicides (atrazine, terbuthylazine, rimsulfuron, primisulfuron-methyl, glyphosate and gluphosinate-ammonium) with respect to the corresponding commercial formulations on microbial activity and biomass of sandy loam soil were investigated. Application rates were: agricultural rate, 20 and 200 mug a.i. g(-1) soil. Application at normal agricultural rates did not lead to significant effects on soil microbial activity, whereas soil microbial activity was markedly stimulated when pure and commercial formulations of the six herbicides were applied at 20 mug a.i. g(-1) soil. The addition of 200 mug a.i. g(-1) soil of four pure herbicides (atrazine, terbuthylazine, rimsulfuron, primisulfuron-methjyl) led to a significant decrease of soil microbial activity. Commercial formulations characterized by a higher relative a.i. concentration (atrazine and primisulfuron-methyl) approximately determined the same decreasing effect of the pure compound, whereas herbicide formulations with a lower relative a.i. concentration (terbuthylazine and rimsulfuron) produced a significant increase in soil microbial activity.