Animal models for spinal fusion are essential for preclinical testing of new fusion methods and adjuncts. They allow for control of individual variables and quantification of outcome measures. Model characteristics are considered. Preclinical experiments to evaluate proof of concept, feasibility, and efficacy are generally studied in an orderly progression from smaller to larger animal models with an evolving cascade of evidence which has become known as the "burden of proof". Methods of fusion analysis include manual palpation, radiographs, computed tomography, histology, biomechanical testing, and molecular analysis. Models which have been developed in specific species are reviewed. This sets the stage for the interpretation of studies evaluating bone graft materials such as allograft, demineralized bone matrices, bone morphogenetic proteins, ceramics, and others with consideration of the variables affecting their success. As evidence accumulates, clinical trials and applications are defined.