Nested nets of three different mesh apertures were used to study mesh-size effects on drift collected in a small mountain stream. The innermost, middle, and outermost nets had, respectively, 425 pm, 209 pm and 106 pm openings, a design that reduced clogging while partitioning collections into three size groups. The open area of mesh in each net, from largest to smallest mesh opening, was 3.7, 5.7 and 8.0 times the area of the net mouth. Volumes of filtered water were determined with a flowmeter. The results are expressed as (1) drift retained by each net, (2) drift that would have been collected by a single net of given mesh size, and (3) the percentage of total drift (the sum of the catches from all three nets) that passed through the 425 pm and 209 pm nets. During a two day period in August 1986, Chironomidae larvae were dominant numerically in all 209 ,um and 106 pm samples and midday 425 pm samples. Large drifters (Ephemerellidae) occurred only in 425 pm or 209 pm nets, but the general pattern was an increase in abundance and number of taxa with decreasing mesh size. Relatively more individuals occurred in the larger mesh nets at night than during the day. The two larger mesh sizes retained 70% of the total sediment/detritus in the drift collections, and this decreased the rate of clogging of the 106 pm net. If an objective of a sampling program is to compare drift density or drift rate between areas or sampling dates, the same mesh size should be used for all sample collection and processing. The mesh aperture used for drift collection should retain all species and life stages of significance in a study. The nested net design enables an investigator to test the adequacy of drift samples.