|
Résumé :
|
Binomial sampling plans were developed for Therioaphis trifolii (Monnell) (Homoptera; Aphididae), one of the most damaging insect pests of alfalfa crops in Argentina. During 1995, 50 stems from each of six alfalfa fields were sampled on each of seven to ten occasions during the season. The mean (m), variance (S-2) and the proportion of stems (P-T) with no more than (T=) 0, 1..5 aphids were calculated. Aphid mean densities ranged from 0.02 to 82 individuals per stem. For each T, Ln(m) was regressed on Ln[-Ln(P-T)], with R-2 ranging from 0.87 to 0.94. Rearranging these equations, the empirical relationships between m and P-T were obtained. The variance associated with the prediction of m from P-T was incorporated in the development of binomial sampling plans for each T and in the calculation of sampling precision. At very low densities large sample sizes are required to estimate mean density with a given precison level. The minimum sample size decreases as mean density increases. However, above certain population levels, the number of samples required for securing predetermined confidence levels increases again. For each mean density there is a tally threshold for which the minimum sampling size is the lowest. Requirements of additional information for selecting the optimal sampling plan are discussed.
|
