This paper compares area-average and unloading compliance crack-growth values with experimental crack-front shapes recorded at various stages of crack growth during fracture tests conducted on 2024-T351 aluminum alloy plate. Crack-front shapes were determined by fracturing the specimen up to a predetermined amount of crack growth and fatigue cycling the specimen for about 4000 cycles at a high stress ratio to mark the crack-front location. For each shape, the area-average and unloading compliance crack lengths were determined. Boundary collocation results provide an approximation to the δ₅ unloading compliance crack length. The crack tunneling results show that the area-average technique produces crack-length measurements more representative of the straight-front finite element analyses than optical based surface measurements. The δ₅ technique is significantly more sensitive to tunneling than the CMOD technique and is easier to apply than the area-average technique.