Introduction

The identification of organisms leads to a number of problems, which never arise when identifying technical material or chemical pure substances. Individual organisms are unique (except for, perhaps, some microbes and viruses). They differ not only in their phenotype (according to the environment), but usually also in their genotype. In contrast, technical objects can often be identified using only a few characters, like weight, colour and measurements. An unambiguous assignment may even be reached with only one character, which provides a unique assignment of all other problem-domain relevant characters. For example, when chemical substances are grouped into partial sets according to their structural formula, all other character states within the groups are homogeneous.

Problems in the Identification of Organisms

The identification of organisms is more complex than the identification of technical materials such as mentioned above, for the following reasons:

• Large object-dependent set of characters. The total set of characters is almost incalculable and strongly depends on the examined object. E.g. the set of characters needed for the identification of lichen is completely different from that needed for higher plants.
• Character set dependent on external parameters. Selecting the characters necessary for an identification from the almost unlimited total number of characters requires expert knowledge. Much of the variation in character states is due to growth conditions of the plant, or the character state vary within the examined taxonomic category, rendering the character useless for identification purposes.
• Almost all instances are unique. With respect to the entire character set, all organisms in a taxonomic category are unique. Normally, the process of classification aims at the establishment of homogeneous groups. This cannot be applied to organisms, however, because each group would consist of one organism only.

The taxonomic system attempts to establish a hierarchical classification of organisms based on phylogenetic criteria, independent of the phenotype. However, since our knowledge is incomplete this system remains somewhat arbitrary, it is frequently updated, and disagreement upon its form may persist (Berendsohn 1995).

Modelling Context of the Taxonomic Identification

Taxonomic identification is a component of the Information Model for Biological Collections developed by the CDEFD project ( Berendsohn & al., this volume) which defines a uniform and detailed data-model for biological collections. The identification of biological material establishes a link to the IOPI-Model (Berendsohn 1994), which treats taxonomic, nomenclatural and biographical information in detail. Both models are parts of a projected larger model giving a unified view of biological information. Entitytypes used in taxonomic identification, but not treated in the CDEFD model are Reference Title (a bibliographic or informal source citation), Person Team (see Elankovan & al., this volume), and Potential Taxon Name (a scientific plant name with a circumscription reference defining its usage).

The CDEFD-Model defines a unit as an object of a biological observation or in a collection. One of the principal properties of a unit is its homogeneity as to placement in a taxonomic category. The entitytype Taxon Identification Event describes the assignment of a Unit to a Potential Taxon Name or a Scientific Name. This assignment is the result of an identification by a Person Team consisting of one or more persons responsible for the event. The attributes of the entity describe only the event itself. For the detailed data structure and an Entity-Relationship-Diagram see Berendsohn & al. (this volume, diagrams 29 and 30).

From a purely semantic view it is possible to assign the data of related entities or constructs to a specific subsystem. For the purpose of this discussion, the subsystem Taxon Identification Event is considered to include information structures belonging to the Unit, the Person Team, the Scientific Name, and Bibliographic References.

Definitions: Terminology, Data Structure Diagrams, Entity Relation Diagrams
Next; Contents; Entity list; References; Mail to wgb@zedat.fu-Berlin.de