At 25° C under aerobic conditions with or without gluamate 10% of the [1-(14)C]glycollate oxidised in spinach leaf peroxisomes was released as (14)CO2. Without glutamate only 5% of the glycollate was converted to glycine, but with it over 80% of the glycollate was metabolised to glycine. CO2 release was probably not due to glycine breakdown in these preparations since glycine decarboxylase activity was not detected. Addition of either unlabelled glycine or isonicotinyl hydrazide (INH) did not reduce (14)CO2 release from either [1-(14)C]glycollate or [1-(14)C]glyoxylate. Furthermore, the amount of "available H2O2" (Grodzinski and Butt, 1976) was sufficient to account for all of the CO2 release by breakdown of glyoxylate. Peroxisomal glycollate metabolism was unaffected by light and isolated leaf chloroplasts alone did not metabolise glycollate. However, in a mixture of peroxisomes and illuminated chloroplasts the rate of glycollate decarboxylation increased three fold while glycine synthesis was reduced by 40%. Although it was not possible to measure "available H2O2" directly, the data are best explained by glyoxylate decarboxylation. Catalase reduced CO2 release and enhanced glycine synthesis. In addition, when a model system in which an active preparation of purified glucose oxidase generating H2O2 at a known rate was used to replace the chloroplasts, similar rates of (14)CO2 release and [(14)C]glycine synthesis from [1-(14)C]glycollate were measured. It is argued that in vivo glyoxylate metabolism in leaf peroxisomes is a key branch point of the glycollate pathway and that a portion of the photorespired CO2 arises during glyoxylate decarboxylation under the action of H2O2. The possibility that peroxisomal catalase exerts a peroxidative function during this process is discussed.