Quantitative understanding of the parameters which control composite fracture is imperative to the implementation of fail safe design and inspection of critical load bearing structures. For isotropic materials, fracture is essentially controlled by a single parameter, e.g., the fracture toughness or the stress-intensity factor. This one dimensional nature lends itself to experimental quantification. However, for anisotropic composites there are at least seven primary controlling parameters: 1) crack length; 2) crack orientation with respect to material axis of anisotropy; 3) nature of applied combined stresses; 4) lamination geometry; 5) deformational and strength responses of the constituent lamina; 6) three kinematically admissible modes of crack extension and 7) crack trajectory. Because of this large number of parameters, experimental quantification by system~tic permutation of the parameters must be realistically viewed as intractable. This paper presents an analytical method of reducing these parameters from seven to two and furnishes experimental observations which lend support to the theoretical model. An experimental p~gram is conducted on fiberglass reinforced epoxy where a centrally notched-crack is subjected to combined loading. Several lamination geometries are tested and by varying the external combined loading, different crack trajectories are predicted by the theoretical model. These predicted trajectories agree well with the experimental observed fracture mode. Such agreement suggests that with further refinement, the general condition of laminated fracture can be characterized.