An in vivo model was developed to compare the biomechanical stability, incidence of radiographic union, bone formation rate, and bone graft remodeling parameters of anterior interbody fusions. Eighteen 1-year-old beagles underwent anterior and posterior spinal destabilization procedures at L5-L6 to produce a reproducible amount of spinal instability--resection of the anterior longitudinal ligament, L5-L6 intervertebral disk, L5 and L6 lamina, spinous processes, zygoopophyseal joints, and ligamentum flavum. Group I (N = 6) were surgically destabilized controls; Group II (N = 6) underwent anterior L5-L6 interbody fusion with iliac crest bone graft; and Group III (N = 6) underwent anterior stabilization with a longitudinal fibular strut graft in addition to the same operative procedure as Group II. Six months postoperatively the group with the highest incidence of successful radiographic L5-L6 arthrodesis was Group III, anterior interbody fusion and fibular stabilization (p less than .10). The rank order of biomechanical stability was the same for the three groups for both torsional and axial compressive stiffness, with Group I (destabilized controls) being the least rigid, then Group II (anterior fusion with iliac crest grafting only), and the most rigid to both torsion and axial compressive loading was Group III (anterior fusion with fibular stabilization and iliac crest bone graft). The bone formation rate [mm3/(mm3 x year) x 10(3)], which was derived from the distance between fluorochrome markers, revealed that the more stable the individual spinal construct, the lower the bone formation rate. In summary, the beagle provided a successful model for studying in vivo the response of anterior bone grafts over a 6-month interval and provided comparative biomechanical and histomorphometric data on spinal interbody fusion techniques.