The sample distribution of oviposition scar counts of the carrot weevil, Listronotus oregonensis (LeConte) was best fit by the Poisson distribution when the mean number of carrot weevil oviposition scars per plant was ⩽0.047. When the mean was ⩾0.06 oviposition scars per plant, the distribution of oviposition scar counts was best fit by the negative binomial distribution. Seventeen spatial patterns of parsley plants (Petroselinum crispum) infested with carrot weevil oviposition scars were obtained from sampling four commercial fields weekly for 4–5 wk on a 10 × 10 grid pattern. Using a 95% confidence level, 148 parsley plants having no oviposition scars indicates that the percentage of infested plants is <2%. Conversely, if one infested plant is found before 148 plants have been examined, then the desired level of confidence has not been achieved, and a treatment is indicated. Sampling with this decision rule and a variety of sampling transects was simulated using the data from commercial fields. Sampling in X- or V-shaped patterns resulted in the most correct decisions and the fewest incorrect decisions overall. The X-shaped pattern led to the fewest cases of failing to correctly classify an infestation that required control. It had one of the largest average sample sizes when the infestation was below the action threshold, and one of the smallest when the infestation was above the action threshold. Results of this study can assist parsley growers with better timing of pesticide applications or biological controls.