Taxonomy is a complex, sometimes arcane subject that is marked by two distinctive characteristics: its historical perspective and its openness to all practitioners. The complexity of such historical taxonomic information as synonymies cries out for the development of electronic, relational databases. The need for openness implies that databases must be freely available.In the flurry of activity to develop paleontological databases, insufficient attention has been paid to security. Three points underscore our concerns in this regard. First, data have value. Anything of value is subject to theft, misuse, or vandalism. Second, paleontologists who develop databases have a right to the first use of their data and a correlative obligation to share those data with the scientific community. These two aspects of the ownership of data-one a right and the other a responsibility-may lead to conflict. Third, sharing of information and the security of files may be incompatible. No system will ensure the security of data that are freely shared.Ensuring the security of databases involves attention to ten kinds of risks: power outages and surges, natural disasters, turnover of staff, the year-2000 problem, buggy software, computer viruses, hackers and vandals, access by unwelcome users, and problems associated with copyright laws and abbreviations, acronyms, and other jargon. Effective security requires a team effort that involves developers, programmers, contributors, users, and security specialists associated with the central computer facility. The strategy should weigh security against performance, productivity, and accessibility. The goal of any security system should be to improve access to while maintaining the integrity of those data.