Weed control represents a high percentage of the production costs in no-till systems in Brazil, and chemical control using herbicides is by far the most important method used. However, the weeds are not uniformly or randomly distributed but have a patchy distribution such that the broadcast application of herbicides can spray post-emergence herbicides in areas where there are no weeds. Therefore, this work had the objective of demonstrating the potential of saving of herbicides in the no-till production system of the Brazilian agriculture, based on weed seed bank and weed seedling maps. The density of several weeds was mapped using a backpack DGPS and laptop computer. Experiments were conducted in Sao Paulo [date not given] in a 17.7-ha field of no-till maize under centre pivot irrigation. Seed bank data was determined from soil cores collected from a depth of 0.05 m in the centre of a 20*20 m grid and emergence assessments in a greenhouse. On the same grid size, weed seedlings were counted in 0.25 m 2 quadrats. Resultant maps showed a high weed density in the seed bank over just 4.67 ha, which was only 26% of the field area. The seedling maps demonstrated that grasses and broadleaf weeds had different distributions with broadleaf weeds occupying 12.6% of the field and grasses 87.4%. The targeting of herbicide to weed patches using pre- and postemergent herbicides has the potential to reduce herbicide use compared to broadcast application giving both environmental and economic advantages.
