Library

Notes: Summary The yield of phosphorus in sorghum plants grown on samples of 22 different soils from the United States was determined in the greenhouse. Analyses of variance of the regressions of these values on the measurements of soil phosphorus by various laboratory methods were calculated as an aid in evaluating the methods. No general advantage was found from incubating the samples in moist condition for a week before analysis versus making the analysis on the initially-dry samples. In the order of decreasing precision of predicting the yields of phosphorus the methods are as follows: (1) anion-exchange resin method of Amer and coworkers, (2) 0.5M NaHCO3 method of Olsen and coworkers, (3) phosphate potential method of Schofield and Aslyng, (4) phosphorus concentration in the 0.01M CaCl2 extract of Schofield and Aslyng, and (5) 0.1N HCl, 0.03N NH4F method of Bray and Kurtz. After taking into account the additional variables, little or no improvement in precision of prediction was obtained when the phosphorussorbedvs. time curves for the anion-exchange-resin method were (a) divided into four segments on the basis of time, the quantities of phosphorus in the four segments being used as independent variables in a multiple-regression equation, or (b) broken down into a maximum of four components on the assumption that the overall curve represents the summation of a group of simultaneous first-order reactions, the quantities of phosphorus in the several components being used as independent variables in a multiple-regression equation. The precision of prediction was improved by using as the soil-phosphorus measurement the sum of the products of the rate of phosphorus extraction by the anion-exchange-resin method and the quantities of phosphorus extracted within individual time intervals. A logarithmic expression was used to fit the relationship, however, and it appeared that the greater precision of prediction resulted from the logarithmic transformation rather than the superiority of the method as such. The precision of prediction was improved also by using the H2PO 4 − concentration instead of the total-inorganic-phosphorus concentration as the independent variable in the 0.01M CaCl2 extracts of Schofield and Aslyng and by using the H2PO 4 − instead of the total inorganic phosphorus sorbed by the anion-exchange resin. This modification made the anion-exchange-resin method considerably better than the others tested.