Among the purposes of classification are association and relocation of data, expression of degrees of similarity, prediction of unobserved data, and expression of functional relationships. Classification must conform to the characteristics of the organism classified and also to the thought processes of the classifier. Analysis reveals hidden patterns and acts as a check on intuition, which though valuable is of uneven reliability.
The Linnaean system is well adapted to modern needs, though we have had to modify its theoretical basis. Modern information may force us to a fundamentally different system at levels at and below that of the Linnaean species.The data of conventional taxonomy are voluminous, but the obvious need for using more characters greatly increases the problem of recording and processing information. All characters, morphological or otherwise, of a species have potential taxonomic value. The latter are often less convenient but may yield special information.
Functional relationships must be distinguished from non-morphological characters. There is a problem of harmonizing classifications based on functional relationships with those based on similarity and difference of attributes. This problem takes different forms at the species level and at the level of higher categories.
The strategy of systematic investigation is influenced by the purpose. For identification a simple array of characters, in a dichotomous table, may be adequate. Such tables are well adapted to digital-computer operations. Where large arrays of characters are to be studied, as where functional relationships, variation or capabilities of organisms are important, time becomes critical, and priority should be given to characters that can be observed quickly or that are known to be specially informative. Methods of multivariate analysis are being applied increasingly in systematics. Methods using unweighted arrays of homologous characters are likely to be supplanted by those that permit controlled weighting, use of nonhomologous characters, and incorporation of sequential and functional relationships. Electronic computers will permit study of problems that would formerly have been too large or complicated. Trained observers will have to provide data, but special types of information, such as chromatograms, electrophoretic and serological observations, and biogeographic data, might be read automatically into a computer. With increased complexity and automation of computer operation, sound initial planning and checking will become more important. The whole subject of systematist-computer relationships needs experience, careful thought, and perhaps a reorientation of attitudes.
Systematics is an approach to biology rather than a department of it; it is intimately related to evolutionary genetic studies, and is in this generation uniting with them in a general science of population biology. Perhaps in the next generation population biology and individual biology will be united in an integrated body of biological science.
