Analysis of sample size for variables related to plant, soil, and soil microbial respiration in a paddy rice field.

Pre-samplings for sample size determination are strongly recommended to assure the reliability of collected data. However, there is a certain dearth of references about sample size determination in field experiments. Seldom if ever, differences in sample size were identified under different management conditions, plant traits, varieties grown and crop age. In order to analyze any differences in sample size for some of the variables measurable in rice field experiments, the visual jackknife method was applied to pre-samples collected in a paddy rice field in Northern Italy, where a management typical for European rice was conducted. Sample sizes for 14 variables describing plant features (plant density,
spikelet sterility, biomass, carbon and nitrogen concentration for the different plant organs and for the whole plant) and for 12 variables describing physical and chemical soil features (texture, pH, water holding capacity, soil organic matter, total carbon and nitrogen concentration, mineral nitrogen concentration) and soilmicrobial activity were estimated. The elementary units of observation were a 3-plant sample and an aggregate sample of four 125 cm3 sub-samples respectively for plant- and soilrelated variables. Sample sizes ranged between 15 and 27 for plant-related variables and between 5 and 6 for soil variables. Relating to plant features, remarkable differences in sample size were observed in
carbon concentration values of different plant organs, probably due to maintenance respiration. Homogeneity among sample sizes for soil variables could be explained by the capability of aggregate samples in capturing a big part of the total variance. This study underlines importance of carrying out pre-samplings aiming at sample size determination for different variables describing the cropping system.