The influence of specimen thickness on Mode I interlaminar fracture behavior was investigated with laminates made from ICI APC-2 prepregs (AS4/PEEK) and from Toray P2212 prepregs (T800/epoxy). The mechanisms of the influence of the thickness effect were discussed on the basis of the direct measurement of the fiber bridging. Double cantilever beam (DCB) specimens 3, 4, 5, and 8 mm thick were prepared from the same unidirectional laminate of 8 mm in thickness in order to avoid molding effects. Fracture toughness, GIC, at the initiation of crack growth was carefully measured in order to clarify the increment of GIC from the initiation values with crack propagation. The initiation values of GIC were independent of the specimen thickness both for AS4/PEEK and T800/3631 laminates. The GIC values for AS4/PEEK laminates increased quickly from the initiation values and leveled off. The thicker specimen indicated slightly higher propagation values of GIC. The thickness effect obtained here was much smaller than that obtained from specimens of originally different thickness. For the case of T800/epoxy laminates, two series of tests from different panels indicated different propagation values of GIC. For one panel, the values of GIC increased continuously with crack length, whereas for the other panel, the GIC values were almost constant, and agreed with the initial values. The influence of different panels and the scatter of data was much larger than the influence of specimen thickness. This suggests that the propagation values of GIC can be used as an indication of fiber alignment of laminates which cannot be detected by the initiation values of GIC. Fiber-bridged zone length was measured in order to discuss the mechanism of the R-curve behavior. A weak correlation between the fiber-bridged zone length and the increase of GIC from the initiation value was observed.