Relative values of physiological parameters of P response of different genotypes can be measured in experiments with only two P treatments

Improved methods for selecting cultivars for their ability to grow well on low P soils will improve productivity and minimise pollution. Estimates are required of the relative values for different genotypes of physiologically meaningful parameters that define yield response to increasing levels of plant available soil P. The object of this study was to devise ways of obtaining these estimates from conventional trials that include numerous genotypes but only two P treatments. The parameters were estimated as coefficients of a modified Michaelis-Menten equation, namely A, the maximum yield that can be obtained with ample P, Km, the concentration of plant available P in the rooting medium at which yield is half the maximum, and B, the gradient of yield against available P as it approaches the origin. Two novel methods of estimation were devised. They both require inputs of yield and its variance at each of two concentrations of available P, but these must be within certain limits. Estimates of parameter values for 12 Brassica oleracea genotypes made by these methods from two P-levels in each of two experiments with six P levels were compared with the values obtained by conventional fitting to the data from all six P levels. According to regression and ranking analyses, the relative values of A obtained for the different genotypes by both two P-level methods were similar to those from the six P-level method and both were reproducible. None of the methods however detected any reproducible differences in Km. Application of one of the two P-level methods to published data for genotypes of other species indicated that there was little intra-species variation in Km within some data sets on maize, beans and leguminous cover crops but much within other data sets. Sensitivity and algebraic analyses are presented to define the experimental conditions required for successful use of two P-level methods. It is concluded that, provided these experimental conditions are met, measurements at two P levels can give almost as much information about intra-species differences in P response parameters as measurements at six P levels. The procedure may also be useful for interpreting other responses of a "diminishing returns" type such as those to other nutrients or plant spacing. © Springer 2006