The Champy-Maillet OsKI reaction has been used upon Golgi complexes to show two kinds of staining. It stains material being processed as it passes along the secretory pathway of the rough endoplasmic reticulum (RER) and Golgi cisternae (GC) up to crystallization in secretory vesicles. It also stains separately the environment within parts of the GC. This GC staining may occur in all compartments (transition vesicles, saccules, condensing vacuoles), but it is characteristically missing from any one of them. The unstained cisternae may be explained if outer saccules are made from either stained or unstained transition vesicles, both of which occur. The presence of empty, unstained transition vesicles is dictated by the surface to volume ratios of microvesicles in relation to saccules. Most transition vesicles must return their membrane to the endoplasmic reticulum, but from time to time it is presumed that they fuse to make a saccule. Saccules, stained and unstained, then mature through the stack. OsKI reactions with tissues and test molecules suggest that in the RER and GC the stain detects labile--S . S--bridges before they lock the tertiary configuration of proteins.