The automatic synthesis of descriptive data using the taxonomic hierarchy

Descriptive data about taxa at one level of the taxonomic hierarchy may be inferred from representative data referable to another level. The process is bi-directional. Data agglomeration enables taxon characteristics to be derived from the characteristics of a subordinate level and data propagation enables taxon characteristics to be derived from the characteristics of a superior taxonomic level. This applies to any form of descriptive taxonomic data (e.g. morphological, cytological, distributional, uses, etc.) and at all levels of the taxonomic hierarchy. Examples taken from a revision for the production of DELTA format descriptions and distribution data are discussed. The major advantages of implementing these processes in a database system are: products and queries are based directly on representative data; they can be derived automatically from the data set without the necessary intervention of a taxonomist; less work is required to synthesize products or answer queries; there is greater guarantee of database integrity; data redundancy is reduced and thus storage space is conserved. The use of these processes should play a central part in any development of taxonomically intelligent database management systems.