To evaluate extracellular hydrolytic enzymes in an in vivo system, plastic chambers were glued over rabbit dermal BCG lesions in various stages of development, after the central epithelium was removed with a scalpel. They were filled with tissue culture medium and left in place 2 days. The following enzymes in the fluid were assayed: collagenase (an enzyme secreted but not stored in macrophages); lysozyme (both secreted and stored); DNase and RNase (released on cell death and possibly regurgitated but not secreted); and, as a control, lactic dehydrogenase (released only on cell death). Tissue sections were prepared and studied histologically for the type of cell infiltrate, for beta-galactosidase (our marker enzyme for macrophage activation), and for necrosis. At 11 and 18 days of age the BCG lesions were largest and the number of activated macrophages in the chamber beds was highest. At this time the levels of the five enzymes assayed in the chamber fluids reached their peaks, tuberculin hypersensitivity was well developed, and the bacilli components would still be plentiful. In general, the chamber fluids from 11- and 18-day BCG lesions contained higher enzyme levels than chamber fluids from tuberculin reactions. Active collagenase was only detected in fluids from such BCG lesions. Evidently, the serum in the chamber fluids was sufficient to inhibit the lower amounts of collagenase probably released from smaller BCG lesions and tuberculin reactions (and from the 2-week polystyrene lesions that were also evaluated). These studies demonstrate that in chronic inflammatory reactions, both acid-acting and neutral-acting hydrolytic enzymes are released extracellularly. Tissue components would be hydrolyzed locally wherever the acid-acting hydrolytic enzymes encounter a drop in pH and wherever the concentration of neutral-acting hydrolytic enzymes exceeds the concentration of their inhibitors.