Objectives: Curing light sources propel the photopolymerization process. The effect of 3 curing units on polymerization shrinkage and depth of cure was investigated. Method and Materials: The curing lights were a conventional and a soft-start quartz-tungsten-halogen (QTH) light source and a light-emitting diode (LED) source. The soft-start QTH and LED intensity outputs were 9% and 17% less than the conventional QTH source, respectively. For a 40-second light cure, the light energy was 32% and 14% lower, respectively. The light sources were applied to 4 restorative composites (microfilled, 2 hybrids, and nanofilled). For each light unit–composite combination, the development of postgel shrinkage during polymerization was measured with strain gauges (n = 15), and the Knoop hardness was tested at 0.5-mm-depth increments to assess degree of cure 15 minutes after polymerization (n = 5). The results were statistically analyzed with 2-way ANOVA at .05 significance level, followed by pairwise comparisons. Results: Both factors, light source and composite, significantly affected postgel shrinkage and hardness (P < .05). The conventional QTH unit generally produced the highest shrinkage and hardness (at composite surface and 2-mm depth). The soft-start QTH unit generated the least shrinkage but achieved the lowest depth of cure. The resulting values for the LED unit were mostly in between the results of the other 2 units. Conclusion: Curing lights should provide sufficient light energy to thoroughly cure composite restorations, which might be achieved without compromising shrinkage stresses if initial intensity is reduced. (Quintessence Int 2009;40:203–214)