BACKGROUND Finishing of composites may adversely influence the sealing ability of direct adhesive restorations. OBJECTIVE This study was conducted to determine whether finishing time and techniques under wet or dry conditions affect the microleakage of 2 types of class V composite restorations. METHODS Eighty class V preparations were made in the buccal surfaces of noncarious molars with the occlusal margins ending in enamel and the gingival margins in dentin. Dental adhesive (Single Bond) was applied according to the manufacturer's instructions. Half of the preparations were restored with a microfilled composite (Silux Plus) and the other half with a hybrid composite (Filtek Z250). Within each major composite group (n = 40), 4 subgroups (n = 10) were established based on finishing technique (diamond finishing burs or aluminum oxide discs) and finishing time (immediate or delayed by 24 hours). The specimens were stored for 24 hours in 37 degrees C water, thermocycled for 250 cycles between 5 degrees C and 55 degrees C, immersed in 2% methylene blue solution for another 24 hours, and sectioned longitudinally. For both enamel and dentin margins, dye solution penetration at the tooth/composite interface was scored from 0 to 4 under stereoscopy at original magnification x 20. Data were analyzed with the Kruskal-Wallis test (P<.05). RESULTS Significantly lower leakage scores were recorded for teeth restored with microfilled composite and finished under wet conditions after 24 hours (P<.01). No significant differences were found among the hybrid composite groups. CONCLUSIONS For microfilled composite restorations on dentin margins, delayed wet finishing with diamond burs resulted in significantly lower microleakage scores than all other finishing protocols. Hybrid composite restorations had equivalent levels of microleakage regardless of the finishing method.