Crack formation and propagation is a significant element of the degeneration process in articular cartilage. In order to understand this process, and separate the relative importance of structural overload and material failure, methods for measuring the fracture toughness of cartilage are needed. In this paper, two such methods are described and used to measure fracture properties of cartilage from the canine patella. A modified single edge notch (MSEN) specimen was used to measure J, and a trouser tear test was used to measure T, both measures of fracture toughness with units of kN/m. A pseudo-elastic modulus was also obtained from the MSEN test. Several potential error sources were examined, and results for the MSEN test compared with another method for measuring the fracture parameter for urethane rubber. Good agreement was found. The two test methods were used to measure properties of cartilage from the patellae of 12 canines: 4-9 specimens from each of 12 patellae, with 5 right-left pairs were tested. Values of J ranged from 0.14-1.2 kN/m. J values correlated with T and were an average of 1.7 times larger than T. A variety of failure responses was seen in the MSEN tests, consequently a grade of 0 to 3 was assigned to each test, where 0 represented a brittle-like crack with minimal opening and 3 represented plastic flow with no crack formation. The initial cracks in 12/82 specimens did not propagate and were assigned to grade 3. The method for reducing data in the MSEN test assumed pseudo-elastic response and could not be used for the grade 3 specimens. Stiffness did not correlate with J. Neither J nor T was statistically different between right-left pairs, but varied between animals. The test methods appear useful for providing a quantitative measure of fracture toughness for cartilage and other soft materials.