The problem of weeds under continuous cropping systems in the scrub and thicket vegetation belt of the Central Region, Ghana

Abstract

The long term effects of three cropping systems, monocropping, two-course rotation (maize/cowpea or sweet potato) and three-course rotation (maize/sweet potato/cassava), on the quantity and quality of both the above-ground weed flora, and the non-dormant portion of the soil's weed seed bank were studied at the Teaching and Research Farm of the School of Agriculture in Cape Coast. After 14 years of continuous cropping, it was realized that there were no significant differences in species composition of the weed flora among the three cropping systems whilst the rotations were significantly less weedy than monocropping. The intensity of rotation had some influence on the population of monocots which were more abundant in monocropping than in the two-course and three-course rotations, in the order of decreasing population. Fears of a shift from the easy-to-control annuals to the hard-to-control perennials were unfounded. The only problematic perennial found in appreciable numbers was Cyperus rotundus. Only seven out of the total of 32 species recorded in the weed flora were found in the non-dormant portion of the seed bank. The two-course rotation with relatively higher tillage frequency had the lowest number of non-dormant seeds followed by monocropping and three-course rotation, in order of increasing populations. An average of 82,539 non-dormant seeds were recorded in an area of 1 m2 in the top 20 cm of moist soil in the study. The bulk of the seeds was contributed by mostly the short cycle and tufted annual monocots including Brachiaria lata, Dactyloctenium aegyptium and Eleusine indica. It was estimated that only one out of a total of 48,821 non-dormant seeds of Dactyloctenium aegyptium could develop and reach maturity. Corresponding figures for Eleusine indica and Brachiaria lata were one out of 32,240 and 2,896 non-dormant seeds, respectively. Cyperus rotundus was the only species in which the quantity of tubers in the seed bank was related to and could be used to predict the level of shoot infestation, regardless of cropping system. (Ghana Journal of Agricultural Science, 1996, 29(2): 81-90)