The incidence of eggs of Bruchus pisorum (L.) was estimated on the edge and 10 m from the edge of field pea, Pisum sativum L., plots in Victoria, Australia between 1986 and 1988, to examine sampling statistics and develop a cost-effective sampling plan. Mean number of eggs per pod varied from 0.19 to 5.31, was generally higher on the plot edge than at 10 m, and decreased progressively with later sowing times. The variancemean (log10[S2]versus log10[m]) and mean-incidence loge[m] versus loge[−logep^o]; p^o = proportion of pods without eggs) relationships were determined for data pooled across all plots. The minimum number of pods (nmin) needed to estimate the mean egg density was calculated for three levels of precision. For incidence counts, a large variance associated with prediction from the regression meant the highest level of precision possible for nmin was SEM/m = 0.38. Calculation of the variance about the regression of log10(S2) on log10(m) allowed computation of 95% confidence intervals for 52 and then for nmin. The confidence intervals showed that the variance about the regression had a noticeable effect on the precision of nmin. A more certain estimate of m could be obtained by using the upper confidence interval. A sequential sampling plan with fixed-precision-stop-lines was constructed for direct counts because it was cost-effective at most egg densities and precision levels.
No takes yet. Share an insight, caveat, or question.
Smith et al. (1992) studied this question.