Ecological correlates of plant range size: taxonomies and phylogenies in the study of plant commonness and rarity in Great Britain

Abstract

Recent simulations have shown that the lack of resolution to be found in taxonomic relative to phylogenetic classification systems only slightly inflates the probability of Type 1 error (inappropriately rejecting the null hypothesis) in evolutionary comparative studies that take relatedness into account. Thus, on this basis, taxonomies may be expected to organize the data adequately for such studies when phylogenies are not available. However, phylogenies instigate particular rearrangements in plant classification that are not easily addressed by simulation, and an analysis of specific variables is used here to gain insight into the consistency of results when a large dataset is organized prior to analysis first with a taxonomy, and then with a phylogeny. This study examines commonness and rarity as it is measured by plant range size of species drawn from the floras of the islands of Great Britain and Crete. The data were restricted to native species of the main island of each group. Using a taxonomic classification of the data, both floras were investigated for the effect of range size on woodiness versus non-woodiness, and tree versus shrub life form. The flora of Great Britain was further examined for possible differences in range size due to overall plant size, and pollination and dispersal types. Using a phylogeny to organize the data, all of the same comparisons of growth form and mutualistic interactions were made for the flora of Great Britain, plus comparisons of cloning versus non-cloning, vining versus non-vining and reproductive schedule (annual, biennial and perennial). Although a phylogeny is the preferred tool for organizing that data, the results suggest that analyses using a taxonomy can lead to conclusions reliable across classification schemes. With either organizational system, trees had larger range sizes than shrubs, and wind-pollinated species had larger range sizes than related non-wind-pollinated species. Additional analyses also indicated that non-clonal plants are more widespread than related clonal species. Changes in the relative position and number of contrasts within comparisons indicate that the exact effect of using a phylogeny versus a taxonomy will depend on the distribution within the phylogeny of the trait being studied.