The recently elevated pest status of Lygus spp. across much of the U.S. cottonbelt has accentuated the need for improved understanding of commonly used sampling methods. A mark-release-recapture method was developed for use in sampling studies of adults of the western tarnished plant bug, Lygus hesperus (Knight), in Pima cotton (Gossypium barbadense L.). Adult bugs were marked with fingernail polish to facilitate their identification and prevent flight. Marked bugs released in sample rows at known population densities were then sampled with the standard 38-cm sweep net. Recovery of marked bugs from 1-m row sections suggested a large proportion (> 85%) of released bugs remained in sample rows. Based on two sets of pooled regressions relating numbers of collected adult bugs to expected numbers (assuming 100% collection efficiency), estimated collection efficiencies of the sweep net changed with cotton crop development from approximately 21.4% in smaller, less developed plants (plant heights < 50 cm on most dates) to 7.7% in more developed plants (plant heights > 52 cm). Increasing the sample unit size from 10 to 20 sweeps improved fit of regression models for both sets of samples, but fit of the model corresponding to less developed plants was still better than for more developed plants. These results illustrate the utility of the mark-release-recapture approach for sampling studies of adult lygus in cotton, and suggest opportunities for its use in quantifying the influence of factors such as time of day, plant development, and variation among samplers on population indices provided by the sweep net in cotton.