Investigation of the principles of rapid soil tests for available potash and other plant nutrients of Virginia soils: III. selection, development and calibration of rapid chemical soil tests for available potassium, phosphorus, calcium and magnesium

Four rapid soil test extracting procedures for potassium, phosphorus, calcium and magnesium, and one absorption procedure for potassium and phosphorus were compared. The soils studied represented important agricultural soils in the Coastal Plains, Piedmont and the Limestone Valley areas of Virginia. Extracting and testing procedures were evaluated on the basis of laboratory performance by correlation with supplemental yield data from 49 field experiments. Four systems of soil test interpretation were applied in correlating soil tests with crop response to current fertilizer applications.

It appeared that only an empirical correlation could be made between crop response and a soil testing procedure which attempts to simulate seasonal crop removal of soil nutrients. A more basic, more tenable correlation appeared possible with soi1 testing procedures which approach total measurement of exchangeable K₂O, CaO and MgO, and adsorbed and acid soluble P₂O₅. With certain modifications, Bray's sodium perchlorate solution was the most promising of the extractants tested for measuring total exchangeable K₂O, CaO and MgO. Neither the acetate or the chloride ion adequately measured adsorbed P₂O₅. No other anions were tested in the phosphorus extractions. It appeared questionable that absorption techniques could be adapted to a correlation with the response of general field crops to fertilizer applications.

The chemical testing procedures used for determining K₂O, P₂O₅ and MgO in the soil extracts were found quite satisfactory. Calcium tests on soil extracts were erratic. The variability of calcium readings appeared to result from pH relationships involved both in the extraction and in the developed test solutions.

The "high-low-medium" interpretation and the interpretation based upon simple response or non-response permitted only a general correlation to be made. Their potential value appeared to be chiefly that of giving support to fertilizer recommendations arrived at on the basis of more specific information regarding the soil type, soil pH, the crop to be grown, the location, and the cropping and fertilization history of the field.

The balance sheet interpretation must, of necessity, be based upon fractional extraction. Such a correlation must be strictly empirical. A balance sheet correlation, using any the extracting procedures that were tested, did not appear to be possible for corn on the heavy soils of the Limestone Valley.

The growth curve interpretation as developed by Bray in Illinois appeared moat promising as a means for resolving fertilizer practice based upon rapid soil tests to a simple formula which takes into consideration current soU productivity and the specific behavior of specific crops to soil and fertilizer nutrients in specific soils. The accurate determination of the constants involved will permit a quantitative relationship between soil teats and fertilizer recommendations.