Pairwise co-existence of Bismarck and Solomon landbird species

Published

Journal Article

Questions: Can the difference between chance and pattern be determined by the composition of species across islands in an archipelago? In particular, will one find 'checkerboards' - a pattern of mutual exclusivity that is the simplest pattern that might occur under competitive exclusion? Organisms: 150 and 141 species of land birds inhabiting 41 and 142 islands of the Bismarck and the Solomon Archipelagos, respectively. (See http://evolutionary-ecology.com/data/ 2447-Supplement.pdf) Analytical methods: For each pair of species within each archipelago, the observed number of co-occurrences is compared to the distribution of the number of co-occurrences derived from a collection of 106 representative unique random, or null, communities. Those species pairs actually co-occurring less often than they do in 5% of those nulls are 'unusually negative' pairs; those co-occurring more often than they do in 95% of those nulls are 'unusually positive' pairs. Islands are ranked from those with the smallest number of species to the largest. A species incidence is the span from the smallest to the largest number of species on islands on which it is found. Results: In each archipelago, proportionately more congeneric species pairs than non-congeneric species pairs are unusually negative pairs. This holds even for species pairs that overlap in their incidences. Among congeneric species pairs found in both archipelagos, a pair that is unusual in one archipelago generally proves to be unusual in the other archipelago as well and to belong to a genus segregating ecologically by means of spatial niche differences. Conclusions: Diamond (1975) suggested that island bird communities were structured by assembly rules that could be deduced by observation of which species did or did not co-occur on particular islands. Critics countered with analyses arguing that co-occurrence patterns in several ecological communities did not differ from random expectations. We conclude that the difference between chance and pattern can be unequivocally determined. © 2009 Stuart L. Pimm.

Duke Authors

Cited Authors

  • Sanderson, JG; Diamond, JM; Pimm, SL

Published Date

  • July 1, 2009

Published In

Volume / Issue

  • 11 / 5

Start / End Page

  • 771 - 786

International Standard Serial Number (ISSN)

  • 1522-0613

Citation Source

  • Scopus