An agronomic and economic evaluation of weed control systems in the Pernambuco agreste

Abstract

Two identical experiments were conducted in corn (Zea mays, Var. Azteca) and beans (Phaseolus vulgaris, var. Costa Rica) in Caruaru, Pernambuco State, Brazil, during 1974, to evaluate the effectiveness and relative profitability of 12 different manual, mechanical, chemical and integrated weed control treatments. Control plots averaged only 5 and 26% of the production achieved on plots receiving complete weed control for corn and beans, respectively. Partial control systems - mechanical cultivation between rows only, or herbicides within rows only - produced yields averaging 64 and 46% of corn and beans yields obtained with complete control systems such as cultivations between and within rows. Differences in yields among systems belonging to the same level of central generally were not statistically significant, indicating perfect technical substitutability among manual, mechanical and chemical techniques as long as weed control coverage was complete. The most efficient system for monoculture corn under local conditions and 1974 prices was application of the preemergence herbicide simazine at 1.5 kg active ingredient per ha which yielded a benefit/cost ratio of 6.03. However if only one rather than two manual or mechanical cultivations was sufficient to give complete control, thon one cultivation would be more efficient than herbicide application. The most efficient system for beans in these experiments was two animal traction hoeings within rows, which yielded a benefit/cost ratio of 3.40. Evidence from other sources, however, indicated that the wider row spacing required for animal traction or micro-trator cultivators can reduce yields which whould make cultivators inefficient in monoculture beans. If cultivator use was precluded, traditional hoeing remained the most efficient weed control method for beans.