The inverse relationship described in an earlier publication between phosphate response and the degree of base saturation has been confirmed with three further groups of field experiments. As an alternative to the degree of base saturation soil pH may be employed.

The discrepancies sometimes found with the more acid base-unsaturated soils, between actual phosphate responses and those expected from the degree of base saturation were found to be related to the control yields. In general, the higher the control yield on a distinctly acid soil the more the percentage response to phosphate fell short of that expected, and vice versa. By forming multiple regressions of percentage phosphate response on both control yield and the percentage saturation of the B.E.C. a more accurate assessment of phosphate response is possible than by using the simple regression of response on the percentage saturation of the B.E.C. A measurable soil characteristic that could be used in the multiple regressions instead of the control yield was not found.

Very significant and inverse relationships were established between percentage phosphate response and the amount of water-soluble or citric acidsoluble silica. These silica contents were also found to be significantly and directly related to the percentage saturation of the B.E.C. It appears that measurements of pH, silica and base saturation function similarly in classifying the soils, distinguishing between the almost neutral soils retaining phosphate in an available form associated with exchangeable bases, acid soils with relatively unavailable phosphate associated with iron and aluminium compounds, and soils intermediate between these.