afXer the fourth, sixth and seventh application was generally similar to the untreated control. In these samples, nitraic content appeared to be correlated negatively with remaining weed number; the conirol thus contained less niiraie than efficient herbicidal treatments. Soil samples taken after the seventh application of bromacil, diuron, fluometuron, neburon and simazine. which contained appreciable rcsiduul concentrations, did not show signilicant dilTerences from control, in an in viiro nitrification test.Ten herbicides, bromacil, chlorthal-dimethyl, dlphenaniid, diuron, fluometuron, neburon, promelryne, pyrazon, simazine aod trifluralin at two doses were repeatedly sprayed, in autumn and in spring, for 4 consecutive years on noncultivated, sprinkler-irrigated field plots. Herbicidal effect was assessed at 1-2 month intervals on the natural weed population and after each observation a paraquat-I diquat spray destroyed emerged weeds. The response of various weed species to herbicides varied markedly but a herbicideinduced shift in the composition of weed population did not occur, presumably because of the paraquat treatment.The overall phytotoxicity to weeds present was, in decreasing order: diuron, bromacil, simazine, triflitralin, prometryne, neburon, fluometuron, pyrazon, diphenamid, chlortlial-dimethyl. Persistence of herbicides was in decreasing order: diuron = bromacil, simazine, neburon (at higher rate), fluometuron, trifluralin. prometryne. Control produced by pyrazon improved with the number of applications, but that of diphenamid and chlorthat-dimethyl remained weak and short. After repeated applications, the activity of these herbicides increased or remained at similar level, but in no case decreased.Soil samples were taken 5 months after each application and bioassayed. Phytoto.Kic residues were detected beneath the disturbed top-soil from bromacil, diuron, fluometuron and simazine after the first application, and from neburon after the second application; residues from trifluralin were found in the top soil only after the fifth application. After the seventh spraying, residues of bromacil were found in the 45-60-cm soil layer.Ammonia content in soil samples taken from treated plots * Contributioti from The Volcani Center, Agricultural Research Organization, Bet Dagan, Israel. 1973 Series, No. 181-E. Effects d'applications repit4es de di.K herbicides agissant par le sol sur une population de mauvaises herbes, i^accumulation des risidus et la nitrification.