The first sustained effort to develop a ley farming system (a pasture legume rotated with a grain crop) for the Australian semi-arid tropics began in the late 1970s at Katherine, Northern Territory, where various strategies were identified and implemented. It was soon discovered that a main constraint to success was the invasion of the legume ley by grass weeds. This occurred despite the replacement of Townsville stylo (Stylosanthes humilis), which provided the base for the early work, with more competitive legumes such as Verano (Stylosanthes hamata) and Centurion centro (Centrosema pascuorum). Early weed control work focused on the use of chemicals, but later competition and population dynamics were studied in S. humilis pastures. The ley farming system comprised a number of essential elements, each of which offered opportunities for weed control First, a legume ley was rotated with grain crops. Past work concentrated on legume leys, but nitrogen (N)-fertilised grass leys may be successful if the N was economically supplied. Grasses can effectively suppress weeds. Rotation of herbicides is possible. Sicklepod (Senna obtusifolia) can be controlled with atrazine in the sorghum phase, and emergent Sida spp. can be controlled in the pasture phase using a herbicide roller. Second, cattle graze the legume ley and crop residues in the dry season. Cattle can be used to suppress grass weeds in the wet season. The major use of the legume ley, however, was to provide high quality food in the dry season when there was little other available food. Future research should investigate the economics of balancing the need for grass weed control with food provision in the dry season. Third, crops sown directly into a ley killed with a knockdown herbicide. Although the need for improved soil surface management was recognised early, development of reduced tillage was delayed until the availability of glyphosate. Initial work concentrated on improving plant establishment under mulch. Recent work has shown that mulches can effectively control weeds in crops without the use of herbicides; best weed control occurred when glyphosate was applied as a single application at sowing. Fourth, ley regenerated as an intercrop in the grain crop. There is no doubt that the presence of an intercrop reduces the yield of the grain crop. Future research should investigate the economics and risks associated with intercrop-induced changes in yield and herbicide use patterns. The extreme climate and sandy soils pose problems for herbicide use. There is marked seasonal variation in effectiveness of pre-emergence herbicides, ranging from zero to marked crop phytotoxicity. The effectiveness of knockdown herbicides is reduced by stressed target plants and rain soon after spraying. Mulch dynamics, grazing, competition and interference, and herbicide interactions with target plants and the environment, were identified as key features requiring attention in a ley farming system to achieve practical weed management.

Field experiments were conducted at Berrimah, Douglas Daly and Katherine in the Northern Territory (NT) during the 1987-88 and 1988-89 wet seasons to obtain yield data for kenaf (Hibiscus cannabinzis L. cv. Guatemala 4) grown under rainfed and irrigated conditions. Under rainfed conditions, maximum stem yield was obtained from sowings early in the wet season. Yield decreased with delay in sowing until the late-December-January period. The maximum rainfed stem yield at Katherine in an above-average rainfall season was 18 400 kg/ha. The maximum yield in a below average rainfall season was 11 700 kg/ha at Katherine, 9200 kg/ha at Douglas Daly and 9400 kg/ha at Berrimah. The applicability to the NT of growth and yield relationships established for irrigated kenaf in the Ord Irrigation Area (OIA) was assessed. The yield potential under irrigated conditions in the NT (21 600 kg/ha at 131 days after sowing) was higher than that reported elsewhere in Australia for the same growth period, but similar to that reported elsewhere for longer growth duration (180-300 days). In the NT, in contrast to the OIA, stem yield showed little or no response to N fertilisation. Stem yield was not related to N uptake, and at high levels of N application, there was marked N accumulation in the stem. Kenaf was able to accumulate up to 110 kg N/ha from the soil reserve where no N was applied. The yield response to plant density varied with the yield level and was similar to that in the OIA. Bark and core yield could be estimated directly from biomass, and indirectly from stem length and plant density, over a wide range of yield levels and cultural conditions. It was concluded that data relating to yield potential and response to N fertilisation cannot be transferred directly from the OIA to the NT.

No abstract