Abstract. Effective and efficient management and manipulation of XML documents requires stable decisions at the time a document enters the XML DBMS to provide for storage structures which are adjusted to the document characteristics and which reflect the future processing needs. While some of the critical parameters require a kind of pre-specification, others can be determined by pre-analysis or sampling of the incoming document or by just making experience-driven “educated guesses”. Important parameters are related to node labeling, path synopsis, document container layout, and indexing. In this paper, we discuss approaches to achieve an adaptive behavior of the storage manager to provide tailor-made native XML storage structures to the extent possible.