Quantitative Studies of Inter-relationships Amongst the Liliatae

Abstract

There is an immense amount of taxonomic data and new information is accumulating rapidly, thereby making it increasingly difficult for taxonomists to comprehend more than a small amount of what is available. Hence it is essential that advantage be taken of modern computing facilities for storing and retrieving information but more especially for generating classifications. The introduction of such technology should lead to a full exposure of the data and methodology thereby enabling potential users of a classification to assess its value for the problem in hand.

As an example of the methodology a sample of eighty-eight liliate families has been classified using data pertaining to fifty-one attributes and using an intensely clustering sorting strategy. Of the four major groups of families emerging from the analysis two, the Zingiberales and a group of water plants corresponding closely with the Alismidae are well established in traditional classifications. The two remaining groups did not agree closely with any recognized grouping in established classifications but nevertheless showed a strong internal homogeneity in that one is comprised largely of wind-pollinated and the other mainly of insect-pollinated families.