The objective of this study was to evaluate weed dynamics based on phytosociological parameters. The experiment was conducted with corn and bean crops cultivated in succession from November, 1998 to May, 2001 in a Cambic Yellow Red Argisol soil, terrace phase, in Viçosa, MG. The weed community consisted of Amaranthus deflexus, Brachiaria plantaginea, Cyperus rotundus, Galinsoga parviflora, Mucuna aterrima and Oxalis latifolia. The treatments consisted of two soil management systems (conventional and no-tillage) and corn use purposes (grain and silage) in blocks with four replicates. Under conventional tillage, the soil was moldboard plowed before sowing, and under no-tillage, weed desiccation was carried out with glyphosate and 2,4-D. Weed evaluations occurred at 20 and 55 days after emergence in 1999/00 and at 20 days after emergence in 2000/01. In both years, evaluations were accomplished immediately before or after application of selective herbicides. For the bean crop, evaluations were also carried out before and after fluazifop-p-butyl and fomesafen application in post emergence, at 20 and 40 days after emergence, respectively. Weed population dynamics was evaluated through phytosociological parameters based on density, biomass and frequency of the sampled species. The results showed that at 20 days after emergence before herbicide application, greater density and relative importance for dicot weeds were observed under no-tillage for both corn and bean crops. Under conventional tillage, there was greater density, dominance and relative importance of Cyperus rotundus. After herbicide application, Cyperus rotundus had the greatest relative importance in both systems evaluated. For both corn use purposes, Cyperus rotundus had its population reduced under no-tillage, compared to conventional tillage.
phytosocioloy; herbicides; no-tillage and conventional tillage; Cyperus rotundus
