The degradation of proteoglycan was examined in cultured slices of pig articular cartilage. Pig leucocyte catabolin (10 ng/ml) was used to stimulate the chondrocytes and induce a 4-fold increase in the rate of proteoglycan loss from the matrix for 4 days. Material in the medium of both control and depleted cultures was mostly a degradation product of the aggregating proteoglycan. It was recovered as a very large molecule slightly smaller than the monomers extracted with 4M-guanidinium chloride and lacked a functional hyaluronate binding region. The size and charge were consistent with a very limited cleavage or conformational change of the core protein near the hyaluronate binding region releasing the C-terminal portion of the molecule intact from the aggregate. The 'clipped' monomer diffuses very rapidly through the matrix into the medium. The amount of proteoglycan extracted with 4M-guanidinium chloride decreased during culture from both the controls and depleted cartilage, and the average size of the molecules initially remained the same. However, the proportion of molecules with a smaller average size increased with time and was predominant in explants that had lost more than 70% of their proteoglycan. All of this material was able to form aggregates when mixed with hyaluronate, and glycosaminoglycans were the same size and charge as normal, indicating either that the core protein had been cleaved in many places or that larger molecules were preferentially released. A large proportion of the easily extracted and non-extractable proteoglycan remained in the partially depleted cartilage and the molecules were the same size and charge as those found in the controls. There was no evidence of detectable glycosidase activity and only very limited sulphatase activity. A similar rate of breakdown and final distribution pattern was found for newly synthesized proteoglycan. Increased amounts of latent neutral metalloproteinases and acid proteinase activities were present in the medium of depleted cartilage. These were not thought to be involved in the breakdown of proteoglycan. Increased release of proteoglycan ceased within 24h of removal of the catabolin, indicating that the effect was reversible and persisted only while the stimulus was present.