"Band spraying and direct drilling", a technique in which bands are sprayed in a pasture with paraquat, followed by the direct drilling of seed into the centre of the bands, was investigated, with the aim of increasing the winter yield of pastures. The work was divided into two parts. In the first, a band sprayer was constructed and tested, and with it a pasture band sprayed and direct drilled in autumn. The resulting production was measured over the following winter period. The second part consisted of an investigation into the distribution of spray within the bands and spray bounce outside the bands, using the same nozzles and operating heights as in the earlier work. The band sprayer was constructed on a disc - drill so that bands could be sprayed and drilled in the same operation, the coulter spacing being 6 in. Measurements were taken of several performance characteristics of a variety of nozzles in order to select three sizes spraying 1 in., 2 in., and 3½in. bands at 30 gal. liquid per sprayed acre. Seed costing with bentonite with the aim of reducing paraquat damage (if this was a problem) was briefly examined, and abandoned after finding that the coat reduced seed germination considerably more than any paraquat damage that may have resulted. In the autumn - winter trial, the factors included were : 4 band widths ("blanket" plus those mentioned above); 3 paraquat application rates (1, 2, and 4 oz. a.i./acre); 2 varieties ("Grasslands Tama" Western Wolths ryegrass and rye-corn); and a nitrogen sub-plot treatment (each half of every plot had either 0 or 1 cwt. nitrolime/acre placed with the seed). Irrigation was carried out prior to spraying and drilling, and was followed by a dry spell of four weeks. This combination appeared to have a deletarious effect on the resulting establishment of the sown species, which together with the wet winter period were partially to blame for the poorer yields in all treated plots (compared to control plots). Measurements taken were mainly of soil moisture, seedling emergence, botanical composition and dry matter yields. Results were analysed by means of t-tests and analysis of variance where these tests were suitable. The results and literature suggested various hypotheses as to the fate of bands and the growth of plants within them and it would appear that defoliation frequency and intensity are important factors. The defoliation treatment in the trial (i.e. at 6-8 in. height; was considered inadequate. The investigations from the second part of the thesis work led to the use of two techniques which could have further use in spray distribution analysis. In the first, the spray liquid incorporated a metal salt (e.g. copper sulphate) in solution and was collected on narrow / blotting paper strips. The metal concentration was measured by an atomic absorption spectrophotometer. With this method, a graph could be drawn of the within band spray distribution, and with log. transformation of the results, spray bounce outside the edge of a band width pasture strip could be confidently measured to 3½ in. from the band. The total amount of spray outside the band was small however, (rarely above 10%), and largely within an inch to either side of the band with the nozzles used. In the other technique, the spray nozzle was photographed in action, the lighting and exposure methods used enabling the extent of spray splash to be observed.