The relationship between initial spinal construct stiffness and the stiffness of the resulting fusion mass was studied by performing standardized 10-segment posterior spinal fusions in goats. Animals were divided into 5 groups based on type of spinal construct, using rods of different diameters (3.2 mm, 4.8 mm, 6.4 mm) with or without rigid crosslinking to produce constructs of different stiffnesses. Stiffness data on 28 animals were obtained by removing the spines en bloc, at 6 or 12 weeks postoperatively, and performing load-deformation testing in axial and torsional loading to determine the stiffness of the fusion masses (rods removed). The initial construct stiffnesses were also compared by ex vivo testing on spine specimens to correlate initial construct stiffness with eventual fusion mass stiffness. In axial testing, results showed stiffer fusion masses from larger diameter rod constructs compared with smaller rod constructs. This was similar to results of control testing on spine specimens ex vivo. Rigid crosslinking did not produce stiffer fusions in axial testing, due to a technical limitation of the button-wire implants used to segmentally fix the rods at each vertebra. In torsional testing, stiffer fusion masses resulted from using larger rods, and rigid Crosslinking also produced the stiffest fusion masses, which was consistent with ex vivo testing. In general, larger diameter (stiffer) rods produced stiffer fusion masses, and no evidence of stress shielding was found.