The number of online resources for biodiversity information is growing. Names of organisms underpin access to information but present a number of unique problems when used as search terms. We examine these problems and assert that a taxonomic name-server or thesaurus is necessary to enable optimal retrieval of records from multiple datasets. A simple solution is presented, based upon our experience working with "real-world" data in the National Biodiversity Network (NBN) in the United Kingdom. The NBN provides access to over 18 million observational records and incorporates a nomenclator covering 198,000 names.