It has been hypothesized that applied mechanical or osmotic loads which decrease cartilage volume by 5% or more are sufficient to relieve all collagen tensile forces, and that further changes in the applied load are completely supported by changes in proteoglycan osmotic pressure. In thisview, cartilage should behave mechanically like a concentrated solution of proteoglycans. We tested this hypothesis by measuring the equilibrium axial and radial stresses in bovine articular cartilage during uniaxial confined compression. If the hypothesis is correct, the observed changes in the radial and axial stresses in confined compression should be equal for compression greater than 5%. However, the observed change in axial stress was always substantially greater than the change in radial stress over the range of strains (5–26%) and saline concentrations (0.05–0.15 M) tested. This indicates that the mechanical behavior of cartilage in confined compression cannot solely be explained by changes in proteoglycan osmotic pressure even for strains as large as 26%. A linear isotropic model was found to describe the observed equilibrium behavior adequately. In addition, the inferred shear modulus was found to be independent of saline concentration and similar to measurements by others of the flow-independent shear modulus. Our results have implications regarding the relative contribution of the proteoglycans and collagen to the mechanical properties of the tissue in compression, and suggest that tensile forces in the collagen network may play an important role in determining tissue behavior in confined compression even for relatively large volume changes.