Design and analysis of mixture systems: Applications in hydroponic, plant nutrition research

Abstract

This study demonstrates that nutrient solutions can be defined as 'mixture systems'. A general methodology for design and analysis of mixture optimization experiments is developed. The emphasis is centered on multivariate investigation of the zone of optimal solution properties as a function of the ion composition and the total ionic strength of the solution. The study of the effects of ion interaction on well-defined solution properties is also possible by this multivariate approach. This work is a valuable tool in mineral nutritional research, because for the first time the chemical feasibility conditions of such solution, combined with additional chemical, physiological or economical constraints, form the foundation of the statistical experimental design theory, which makes the optimization of complex mixtures of ions in relation to well-defined response variables possible.

Contrasting Nutritional Acclimation of Sugar Maple (Acer saccharum Marsh.) and Red Maple (Acer rubrum L.) to Increasing Conifers and Soil Acidity as Demonstrated by Foliar Nutrient Balances

“…Second, nutrients in the leaf are not only related physiologically through interactions (Marschner, 2011), but they are also numerically related given that all nutrients are constrained to sum to 100%; when one nutrient increases, then the others or some of the other nutrients must decrease concomitantly, and vice versa (Aitchison, 1986). Third, a D-part composition (D is the number of nutrients that have been measured) has only D-1 degrees-of-freedom because one component can be calculated as the difference between the unit and the sum of the other components (Claringbold, 1955;Schrevens and Cornell, 1993;Aitchison and Greenacre, 2002). Using raw compositional data or dual ratios to conduct linear statistical analysis therefore leads to methodological biases and analytical incoherence caused by: (1) spurious correlations;…”