Many of the key debates in archaeology hinge on the chronology and interpretation of data gathered from cave and rockshelter stratigraphies, especially those in karstic limestone environments which are selectively targeted by archaeologists because of their superior preservation characteristics. It has long been recognized that such sites often contain a variety of cemented deposits including cave breccias and that some breccias contain anthropogenic inclusions such as stone artefacts, shell and burnt animal bones. Cementation enhances the survival through time of such brecciated deposits. This can result in chrono-stratigraphic intervals surviving on cave walls and speleothems that are no longer represented in the stratigraphy of cave floors. This has important implications for understanding apparent presence/ absence of human occupation and cultural continuity as seen in archaeo-stratigraphy in caves and rockshelters, especially in relation to human migration in the humid tropics in SE Asia and the Pacific, and over Pleistocene to Holocene timescales. Here we discuss localized breccia formation, the erosional processes that leave remnant deposits adhering to walls and speleothems at heights well above current cave floors, and the possible significance of local and regional processes, especially changing base levels, in triggering gutting out phases impacting cave floor sediment architectures. Equally significant in terms of chronological completeness, representativeness and bias is the contribution made by cultural materials encased in older breccias as they erode and are (re-)incorporated into younger accumulating cultural deposits. Case studies from cave sites in Papua New Guinea and Timor Leste are used to illustrate these issues.