Sample Number Effect on Precision of Measurements of Rice Canopy Photosynthesis and Leaf Area

Abstract

To effectively utilize resources in the measurement of canopy photosynthesis, acceptable statistical precision is necessary. Our objective was to determine the effect of sample number per plot on effective precision of canopy apparent photosynthesis (CAP) and leaf area index (LAI) measurements. A rice (Oryza sativa L.) field experiment with four N fertilizer treatments (0,67,134,201 kg ha−1) was sampled in 2 yr when rice was at the panicle differentiation stage of growth. Samples were taken at four locations in each plot for CAP and LAI. The CAP and LAI error mean squares were reduced by 32 to 81%, and 20 to 53%, respectively, by increasing numbers of samples from one to four. Given similar experimental conditions, between two and four samples per plot would be necessary to detect differences among the treatments for CAP. A single sample per plot would suffice to detect differences in LAI. To detect differences in canopy photosynthesis of flooded rice experiments when repeated samples of the same plots cannot be made, additional sampling of plots is likely to be necessary to detect treatment differences for CAP, but not for LAI.