In order to test doses of metribu zin (0,0; 0,35; 0,70 and 1,05 kg a.i.ha ¹) in weed control and yield of soybe an (Glycine max (L.) Merri ll, Uberaba), an experiment was conducted under field conditions in a clayey red yellow podzolic soil with 2,9% organic matter at Viçosa, MG, during the 1985 /86 growing season. Most of the monocotyledqns species present in the experimental area consisted of Cyperus rotundus L., Brachiaria plantaginea (Link) Hitch and Cynodon dactylon (L.) Pers. Reduction observed was only for Brachiaria plantaginea as metribuzin doses increased, the sa me occuring with dycot iledons species existent in the area except for Oxalis oxypt era Prop. wich was not controlled by the doses using. The avera ge of the tota l weed density minus Cyper us rotundus, Oxalis oxyptera and Cynodon da+-ylon was 141, 124, 62 and 59 plants.m² for 0,0; 0,35; 0,70 and 1,05 kg a.i.ha-1 metribuzin, respectively. The lowest dose of metri buzin (0,35 kg a.i.ha-1) was as sufficient to promote reduction on weed shoot biomass at the highest did (1,05 kg a.i.ha-1). However, the mean of the total weed density was drastically recuded in the doses of 0,70 and 1,05 kg a.i.ha-1. The metribuzin caused damage to soybean plants at highest dose (1,05 kg a.i.ha-1) employed with a slight leaf injury '(chlorosis) until 25 days after emergence. After this period it was observed a total recovery of the plants treat ed with this dose. The yield and the harvest index were not affected by doses of metribuzin.
Herbicide; weed population; phytotoxicity