Field experiments were carried out in 2000 and 2001 in northern Italy, with the aim of developing a composite sampling strategy for estimating populations of Liriomyza huidobrensis (Blanchard) larvae and mines on lettuce. Larval parasitoid populations also were sampled to estimate the contribution of these beneficials to pest control. Covariance analysis and parallelism test indicated that there were not significant differences in the coefficient of Taylor’s power law between treatments (untreated versus treated) and between seasons (2000 versus 2001) for both pest and parasitoid stages. The slope of each regression was significantly >1 for L. huidobrensis mines and larvae, and endoparasitoid and ectoparasitoid larvae, indicating a clumped distribution of both pest and parasitoids. Constant precision level stoplines for mines and larvae of L. huidobrensis and for endoparasitoid and ectoparasitoid larvae were calculated, by using the common a and b derived from Taylor’s power law. Wilson and Room binomial sampling was used to estimate the mean number of L. huidobrensis mines or live larvae from the ratio of leaves with mines. In this way, it is possible to estimate the mean density of L. huidobrensis by counting the leaves with mines, an index parameter easy to obtain in field sampling. Validation of the presence-absence sampling plan by using field data collected in 2002 and 2003 seasons, showed that the Wilson and Room model fits very well empirical data. Statistical analysis does not show any significant difference between observed and predicted data.