Fertilizing programs

Abstract

Long-term experiments designed to measure the effects of residues from phosphorus fertilizer applications on successive crops of wheat in Australia, and of soybeans in Brazil, showed that effects vary markedly in successive years because of uncontrolled variations in growing conditions. These variations constitute a serious problem in the prediction of residue effects from the results of fertilizer experiments. Given the time requirement and expense of long-term experiments, it does not seem practicable to attempt to solve the problem by carrying out a sufficient number of experiments to provide bases for statistical estimates, with allowance for variability in growing conditions during sequences of several years.

The possibility of transforming yield data to percentages of maximum attainable yields for the year of the experiments, to reduce or avoid the problem is examined. It is concluded that this simple and popular device is not satisfactory. Apart from the problems of estimating and predicting maximum yields, the experimental data show that yield response to fertilizers is not simply proportional to level of yield for individual years.

The usefulness of substitution rates, expressing the effectiveness of fertilizer residues as a proportion of equivalent fresh fertilizer applications, is also examined. Although this device may effectively reduce the problem of variation from that of sequences of years to that for the reference year for the fresh application, it is not considered potentially useful because rates vary with both the yield level nominated for the calcuation, and with combinations of fresh and previous fertilizer rates possible for a nominated yield.

Effects of fertilizer residues on soil composition were shown, using a standard soil test procedure for phosphorus fertilizer requirements. Since soil tests can be used to predict fertilizer yield functions, with statistical allowance for uncontrolled variations in growing conditions, it is suggested that residue effects are better measured by soil analyses than by fertilizer experiments.