The objective of this study was to investigate the effects of different light-curing methods on microleakage, marginal adaptation, and microhardness of composite restorations. Slot-type preparations were made in bovine teeth, with gingival margins on dentin. Specimens were divided into 12 groups (n = 12) according to composite-light-curing unit (LCU) combinations. Three composites were used: Filtek Supreme, Herculite XRV, and Heliomolar. All restorations were placed using the same adhesive. Four LCUs were used: a quartz-tungsten-halogen (QTH) LCU (Optilux 501), a first-generation light-emitting diode (LED) LCU (FreeLight 1), and two second-generation LED LCUs (FreeLight 2 and Translux Power Blue). After finishing and polishing, specimens were subjected to mechanical load cycling (100,000 cycles). Gingival margin adaptation was determined as a function of gap formation using epoxy replicas. Microleakage was evaluated by measuring dye penetration across the gingival wall in cross-sectioned specimens. Microhardness was measured as Knoop Hardness number (KHN) at different occluso-gingival locations in cross-sectioned specimens. Data were analyzed for statistical significance (p = 0.05) using appropriate statistical tests. Marginal adaptation was affected by load-cycling in most specimens, but no significant differences were observed among composites and LCUs. Microleakage was not affected by LCU, except for Heliomolar specimens which when cured with Optilux 501 resulted in higher microleakage scores than those obtained with the other LCUs. For microhardness, Translux Power Blue generally produced the highest values and the FreeLight 1 produced the lowest. The performance of the second-generation LED LCUs generally was similar to that of the QTH control, and better than that of the first-generation LED unit. (c) 2006 Wiley Periodicals, Inc.78B2302311