Taylor's power law analysis was performed on among- and within-subsample mean and variance estimates of the number ofUrophora affinisFrauenfeld andUrophora quadrifasciata(Meigen) galls per spotted knapweed seed head. Highly significant correlation coefficients suggested that Taylor's power relationship fits the data well. A comparison of thebparameters from Taylor's analyses suggested that the spatial dispersion ofU. affiniswas more clumped than that ofU. quadrifasciata. The spatial dispersion ofU. quadrifasciataappeared to become more uniform in the presence ofU. affinis. Means and estimated variances (among and within subsamples) were used to develop numerical sampling equations for the two fly species. A cost (work effort) analysis was performed on the among- and within-subsample units. Optimal subsample sizes calculated for overall mean gall densities ofU. affinisandU. quadrifasciata(averaged over all sampling sites) were 18.95 and 21.86 seed heads per subsample, respectively. To protect against sampling low population sites, it is recommended that 15 subsamples of 25 seed heads each be collected per sampling site and used to assess population levels of the two fly species. Information presented in this paper should improve accuracy and efficiency in future sampling ofU. affinisandU. quadrifasciataand aid in determining those sites warranting future releases of the flies.