A test fixture for testing a thick split cantilever beam for scissoring delamination (Mode III) fracture toughness was developed. A three-dimensional finite element analysis was conducted on the test specimen to determine the strain energy release rate, G, distribution along the delamination front. The virtual crack closure technique was used to calculate the G components resulting from interlaminar tension, GI, interlaminar sliding shear, GII, and interlaminar tearing shear, GIII. The finite element analysis showed that at the delamination front no GI component existed, but a GII component was present in addition to a GIII component. Furthermore, near the free edges, the GII component was significantly higher than the GIII component. The GII/GIII ratio was found to increase with delamination length but was insensitive to the beam depth. The presence of GII at the delamination front was verified experimentally by examination of the failure surfaces. At the center of the beam, where the failure was in Mode III, there was significant fiber bridging. However, at the edges of the beam where the failure was in Mode II, there was no fiber bridging and Mode II shear hackles were observed. Therefore, it was concluded that the split cantilever beam configuration does not represent a pure Mode III test. The experimental work showed that the Mode II fracture toughness, GIIc, must be less than the Mode III fracture toughness, GIIIc. Therefore, a conservative approach to characterizing Mode III delamination is to equate GIIIc to GIIc.