This study determined the biomechanical differences between anterior and posterior lumbar interbody fusion (ALIF and PLIF). Ten cadaveric spines were tested. Five specimens had ALIF and five had PLIF at L4-L5. Stabilization was performed with pedicle screws and rods (Cotrel-Dubboset, Sofamor-Danek, Memphis, TN, U.S.A.). Angular motion was measured in flexion, extension, bending, and torsion on the intact, instrumented, and "fused" specimens. Instrumentation alone caused a significant decrease in segmental motion in all loading modes (p < 0.01). After the simulated fusion procedures, all specimens were most stable in flexion, and significantly less stable in extension (p = 0.04). Comparing directly, ALIF was significant more stable in left torsion (p = 0.03) with trends in left bending (p = 0.08) and right torsion (p = 0.07). Thus, from a purely biomechanical perspective, ALIF appears to be slightly superior to PLIF.