The streaming potential response of cartilage in the confined compression creep configuration was assessed theoretically and measured experimentally in normal and proteoglycan-depleted tissue. The analytical solution, using the linear biphasic continuum model including electrokinetics and assuming homogeneous material properties, predicted that: (i) the peak streaming potential is delta V = ke x delta sigma, where ke is the electrokinetic coefficient and delta sigma is the change in compressive stress; (ii) the potential is maintained at 95 to 100% of the peak value for 0 < t < 0.10 tau, where tau is the gel diffusion time constant; and (iii) during short times, 0 < t < 0.01 tau, 90% of the peak streaming potential occurs over a region extending 23% into the tissue sample. Experimentally, adult bovine cartilage disks, 0.5 mm thick, were subjected to step changes of compressive stress. The measured changes in potential indicated a linear response for changes in stress up to 0.10 MPa. The ke of normal cartilage, estimated from the short time (0 < t < 2 sec) change in potential, was -1.65 +/- 1.25 mV/MPa. Digestion of cartilage by chondroitinase ABC resulted in an increased (less negative) ke of -0.75 +/- 0.70 mV/MPa and a 33 +/- 29% depletion of anionic glycosaminoglycan, whereas digestion with trypsin resulted in a further increase in ke to +1.74 +/- 0.95 mV/MPa and a 98 +/- 1 % depletion of glycosaminoglycan. The streaming potential measurement may be a useful addition to the widely used confined compression creep test to assess cartilage material properties.