Born-digital material such as archived Web sites provides unique challenges in ensuring access and preservation. This article examines some of the technical challenges involved in harvesting and managing Web archives as well as metadata strategies to provide descriptive, technical, and preservation related information about archived Web sites, with the Library of Congress' Minerva project as an example. It explores a possible data model for archived Web sites, including discussion of a proposed METS profile for Web sites to enable the packaging of different forms of metadata. In addition it explores how the work on metadata for long-term preservation undertaken by the PREMIS working group can be applied to Web archiving. (Contains 52 notes and 4 figures.)