TESTING GYPSUM REQUIREMENTS FOR DRYLAND CROPPING ON A RED-BROWN EARTH

Gypsum requirements of a Red-Brown Earth used for dryland cropping wer e determined by 2 methods: those of the State Chemistry Laboratory (SC L) and of the Institute for Sustainable Irrigated Agriculture (ISIA). These are based, respectively, on exchangeable cations of soil by the Tucker method, and on water-soluble cations in a 1:5 water extract of soil. Information was also gained on longevity of gypsum effects on so il. Gypsum was applied to plots at rates 0.5, 1, and 2 times that pred icted by ISIA, whereas the SCL method predicted that no gypsum was req uired. After 2 years, plots were split and the 3 rates of gypsum were either applied to soil previously untreated with gypsum, or re-applied to the soil treated 2 years before.According to the SCL test, the 0-1 0 cm soil depth of the Red-Brown Earth was not sodic (ESP <6), it was slightly magnesic (EMgP >25), and required no gypsum. However, it was found that gypsum lowered ESP, EMgP, and clay dispersion, with some ef fects extending into the 10-20 cm soil depth. With the ISIA method, th e 0-10 cm soil depth was classed as low-sodic, but potentially dispers ive; it required 2.5 t/ha of gypsum if soil was cultivated, but no gyp sum if it was direct-drilled or was under pasture. One year after appl ication, only the highest rate of gypsum (5 t/ha) significantly (P < 0 .001) raised the electrolyte concentration of the soil suspension, alt hough all rates reduced (P < 0.001) SAR, and the 2 higher rates reduce d (P < 0.01) clay dispersion. However, these effects had disappeared a fter Year 3. There were significant increases in crop yields due to gy psum treatment in Year 1 (0.5-0.9 t/ha, P < 0.01) and Year 2 (0.3 t/ha , P < 0.001), but no response in Year 3. Fungal diseases seemed to red uce wheat responses, and the highest rate of gypsum caused chlorosis o f lupins. Judging by crop performance, the ISIA method predicted an op timal rate of gypsum for the 0-10 cm layer of this soil type. The SCL prediction was also only for the 0-10 cm layer, but had it been used f or deeper layers in the profile (the original intention for the techni que), it would have given a recommendation not very different from the ISIA method.