Prediction of plant response to fertiliser by means of soil tests

摘要

Phosphate response curves were obtained on 48 soil samples belonging to the brown-grey earth and dry-subhygrous yellow-grey earth soil groups, by growing ryegrass in pots in the glasshouse. Phosphorus fractions and phosphorus extracted using modifications of several common extraction techniques were measured on the soil samples; many of the soil sets were high in apatite-type phosphate. Phosphorus uptake by control treatments and control yields as a percentage of maximum yield were used as measures of plant response for correlation with phosphorus fractions and soil test values, using linear regression for phosphorus uptake and an asymptotic regression for relative yields. The 16 h modification of the Olsen method gave the closest relationship with both uptake and relative yield. The relationship was almost as close for the 0.5 h Olsen extraction if soil volume was used rather than a weight, and there was little difference in correlation between modifications using 2.2 ml soil/100 ml solution or 5 ml/100 ml, although more phosphate was extracted by the wider ratio. The Bray No. 1 tests were almost as satisfactory as these. The Truog test was not reliable, because it dissolved part of the non-available apatite-P fraction. The MAF has now adopted the 0.5 h Olsen test modification using 5.0 ml soil/100 ml extractant as its standard test on all soils. On the soil groups studied, this new test is less satisfactory than the 16 h Olsen test, but the latter could not be fitted into the regular soil testing routine without expanding facilities. Field trials on a wide range of soils suggest that a level of 15 #x3BC;g/ml best discriminates responsive from non-responsive soils, but not necessarily for the soil groups studied. Further field trial calibration is essential.