Taxonomy of the tribe Triticeae (Poaceae) using various numerical techniques. I. Historical perspectives, data accumulation, and character analysis

Abstract

A taxonomic study of the Triticeae at the generic level is needed for both practical and theoretical purposes. Many genera in this tribe contain species of prime importance to agriculture. It appears from many previous studies in many different disciplines, that the genera are very close to each other so that there are difficulties in properly defining them; but this very fact brings all the members of that group up to a high degree of importance for gene transfer purposes because of their close evolutionary relationships.The first paper in this series deals with the definition of 45 characters and their states, and evaluates their interrelationships and relative degree of importance for classificatory and diagnostic purposes. This is done through examination of conventional and new characters, and scrutinous verification of characters used hitherto by different authors in different areas at various times in the last 100 years. Analysis of the data was done using Estabrook's information theoretic model of character analysis and various other techniques, such as multidimensional scaling, principal component analysis, and clustering. It is concluded that no necessary correlation exists between the usefulness of a character for discriminating a particular genus and the value of the same character for distinguishing genera in the tribe and classifying them.