|
Résumé :
|
Based on field sampling of the green peach aphid, Myzus persicae (Homoptera: Aphididae) on Brassica campestris ssp. chinensis in the suburbs of Hangzhou, Zhejiang Province of the People's Republic of China during 1990-1992, a data set consisting of 90 estimates of mean aphid density (m, number of aphids per plant), variance (s(2)) and the proportion of plants (P-T) with no more than T(= 0, 5, 10, 20, 30 and 40, respectively; defined as tally threshold) aphids was generated. Empirical relationships between m and P-T were developed for each T value by using the parameters from the linear regression of In(m) on In[-1 n(P-T)]. The importance of the T values in reduction of sampling errors and their application to binomial sampling plans were discussed. The results indicate that small T values, particularly 'empty' plants (T = 0, conventional binomial sample), could lead to spurious estimates of m from P-T with small determination coefficients, large prediction variance and sampling errors. Binomial sample sizes based on the T value of 30 were found to be associated with a relatively high determination coefficient (0.9348), small prediction variance and small sampling error, and thus are recommended for use in estimating the mean population level of green peach aphid on B. campestris ssp, chinensis in the field at population densities above 10 aphids per plant.
|
