Sample size requirements to evaluate spore germination inhibition by compost extracts

We examined the effect of compost sample size (50 to 5000 g) on the precision of estimates of inhibition of Venturia inaequalis (Cke) Winter conidial germination induced by extracts of anaerobically-incubated composts. Composts were prepared from vegetable material; some were amended with manure. Variability due to sample size was considered as a function of extract efficacy (high, medium or low) and compost heterogeneity (high or low). Extract efficacy was characterized as ability to inhibit more than 75% of the conidia (high), 75 to 40% (medium), or fewer than 40% (low). Heterogeneity was characterized as visual or tactile presence (high) or absence (low) of parent material in the compost. To address the question of sample size-related variability we partitioned individual extract incubations (samples) into aliquots (sub-samples), and aliquots into microtiter plate wells (sub-sub-samples). For this nested design, the largest component of variation was consistently found to be that associated with wells; aliquots were a negligible source of variation. Sample size over the range examined was generally of small importance for extracts of medium and high efficacy, but not low efficacy, independent of compost heterogeneity. For previously untested composts, particularly those of high heterogeneity, statistical analyses of our data suggest that samples of at least 500 g circumvent potentially large errors and consequent difficulties to detect differences among composts or effects of experimental variables. Copyright (C) 1996 Elsevier Science Ltd