Samples of a sandy loam soil were supplied with normal, 10-fold and 100-fold rates of ioxynil, dalapon, mecoprop, dichlorprop, MCPA + dichlorprop, picloram, and amitrole-T and incubated at 29°C at 65 per cent of the waterholding capacity. Treated soil samples were compared with untreated samples. Samples supplied with (NH4)2SO4 and herbicides were used to investigate the effect of the herbicides on the rate of nitrification and the production of nitrite. In several cases higher numbers of bacteria were found for a longer or shorter period in soil treated with herbicides. There was some evidence that certain groups of bacteria had adapted to ioxynil in a soil sample supplied with the 100-fold rate of this chemical. After 2 or 4 weeks lower numbers ofAzotobacter chroococcum were found at the normal rate of ioxynil, dalapon, mecoprop, and dichlorprop. At the 100-fold rate of application the numbers of Azotobacter were unfavourably affected by all herbicides. The production of mineral nitrogen was hardly affected by the normal and 10-fold rates of application. In the first week the rate of nitrification was slightly depressed in soil samples treated with the normal rates of dichlorprop and amitrole-T and with the 10-fold rates of dalapon, mecoprop, and MCPA + dichlorprop. Strong inhibition of the nitrification for at least 7 weeks was shown by the 10-fold rate of amitrole-T. At the 100-fold rate all herbicides, with the exception of picloram, depressed the rate of nitrification for a longer or shorter period. During the second week a very small increase of nitrite was found in the samples treated with the normal and 10-fold rates of dalapon, mecoprop, dichlorprop, and amitrole-T. A small increase of nitrite was noted for 26 weeks in samples treated with the 100-fold rates of amitrole-T. A highly significant depression of CO2 evolution was found in the first week in samples treated with the normal rates of ioxynil, dalapon, mecroprop, dichlorprop, and amitrole-T, also in samples treated with the 10-fold rates of dalapon, mecoprop, dichlorprop, MCPA + dichlorprop, picloram, and amitrole-T. A highly significant depression of CO2 production was found after 8 weeks in all the samples treated with herbicides at the 100-fold rate with the exception of the sample treated with picloram. The decomposition of the herbicides was studied in soil samples treated with the 100-fold rates of herbicides. Only traces of dalapon and mecoprop were found after 9 months, but 7.2% ioxynil, 29.8% dichlorprop, 39% (MCPA + dichlorprop), 52.1% picloram and 52.2% amitrole-T were still present in active form.