Evidence from negatively associated distributions

A number of studies have used patterns in distribution as evidence for the importance of interspecific competition. Foremost amongst these is Diamond's (1975) survey of the land birds living on the islands of the Bismarck Archipelago off the coast of New Guinea. The most striking evidence comes from distributions that Diamond refers to as 'checkerboard'. In these, two or more ecologically similar species (i.e. members of the same guild) have mutually exclusive but interdigitating distributions such that any one island supports only one of the species (or none at all). Figure 19.11 shows this for two small, ecologically similar cuckoo-dove species: Macropygia mackinlayi and M. nigrirostris.

A null model approach to the analysis of distributional differences involves comparing the pattern of species co-occurrences at a suite of locations with that which would be expected by chance. An excess of negative associations would then be consistent with a role for competition in determining community structure.

Thorough censuses of both native and exotic (introduced) plants occurring on 23 small islands in Lake Manapouri in the South Island of New Zealand (Wilson, 1988b), were the basis for computing a standard index of association for every pair of species:

evidence from the divergent evolution of lizards on islands evidence from 'checkerboard' distributions in ... ... island birds...

... and native and exotic plants on islands in a lake

Two-morph stage

Twig

Generalist

Twig

Generalist

Three-morph stage

Twig

Crown-giant

Crown-giant

Trunk-ground

Twig

Trunk-ground

Crown

Trunk-ground

Four-morph stage

Twig

Crown-giant

Crown-giant

Trunk- Trunk-crown ground

Twig

Trunk-ground

Trunk-ground

Crown- Trunk-giant crown

Crown

Five-morph stage

Twig Crown- Trunk- Grass- Trunk-giant crown bush ground

Twig Crown- Trunk- Grass- Trunk-giant crown bush ground

Figure 19.10 The evolution of Anolis communities (a) on Puerto Rico and (b) on Jamaica for two-, three-, four- and, in the case of Puerto Rico, five-ecomorph communities. Labels at nodes in the trees are the estimated ecological characteristics of the ancestors. (After Losos, 1992.)

Figure 19.10 The evolution of Anolis communities (a) on Puerto Rico and (b) on Jamaica for two-, three-, four- and, in the case of Puerto Rico, five-ecomorph communities. Labels at nodes in the trees are the estimated ecological characteristics of the ancestors. (After Losos, 1992.)

Checkerboard Distribution Ecology

Figure 19.11 Checkerboard distribution of two small Macropygia cuckoo-dove species in the Bismarck region. Islands whose pigeon faunas are known are designated as M (M. mackinlayi resident), N (M. nigrirostris resident) or O (neither species resident). Note that most islands have one of these species, no island has both and some islands have neither. (After Diamond, 1975.)

Figure 19.12 A comparison between the observed values of association between pairs of (a) native plant species and (b) exotic plant species on islands in Lake Manapouri (histograms), and the distributions expected on the basis of a neutral model (o). (After Wilson, 1988b.)

Figure 19.12 A comparison between the observed values of association between pairs of (a) native plant species and (b) exotic plant species on islands in Lake Manapouri (histograms), and the distributions expected on the basis of a neutral model (o). (After Wilson, 1988b.)

Difference Richness

where dik is the difference between the observed (Oik) and the expected (Eik) number of islands shared by species i and k, expressed in terms of the standard deviation of the expected number (SDik).

The resulting sets of association values for the real communities of native and exotic species are presented as histograms in Figure 19.12. These can be compared with null model communities in which island species richnesses and species frequencies of occurrence were fixed at those observed, but species occurrences on islands were randomized (Wilson, 1987). One thousand randomizations were performed, yielding a mean frequency in each dik category (the circles in Figure 19.12). The analysis of native plants showed an excess of negative associations (highly statistically significant for the bottom four categories) and of positive associations (highly significant for the top five categories), with a corresponding deficit of associations near zero. In contrast, the analysis of exotic plants showed no significant departure from the null model.

In the case of native species, the excess of negative associations is consistent with the action of competitive exclusion, and this is particularly likely for the woody species. However, we cannot rule out an explanation based on a tendency of particular pairs of species to occur in different habitats, which themselves are not represented on every island (Wilson, 1988b). The most likely explanation for the excess of positive associations amongst native plants is a tendency for certain species to occur in the same habitats. The agreement of exotic species with the null model may reflect their generally weedy status and effective colonization abilities, or it may indicate that the exotics have not yet reached an equilibrium distribution (Wilson, 1988b).

The number of checkerboard pairs in a community can be readily calculated by counting the number of unique pairs of species that never co-occur. A less strict version of Diamond's assembly rule that 'some pairs of species never coexist' can be assessed using the C score of Stone and Roberts (1990). This index also measures the degree to which species co-occur but does not require perfect segregation between species. The C score is calculated for each pair of species as (Ri — S)(Rj — S) where Ri and Rj are the number of sites where species i and j occur, and S is the number of sites in which both species co-occur. This score is then averaged over all possible pairs of species in the matrix. For a community structured by competitive interactions, the number of checkerboard pairs should be greater and the C score should be larger than expected by chance.

Gotelli and McCabe (2002) checked the generality of negatively associated distributions (in support of a structuring role for competition) in a meta-analysis of various taxonomic groups in 96 data sets that reported the distribution of species assemblages across sets of replicated sites. For every real data set, 1000 randomized versions were prepared and an index of association dik was computed (as in Wilson, 1988b) but Gotelli and McCabe called this index the standardized effect size (SES). The results of this analysis for all 96 data sets together support the predictions that the C score and the number of checkerboard pairs should be larger than expected by chance (Figure 19.13a, b). The null hypothesis in each case is that the mean SES should be zero (real communities not different from simulated communities) and that 95% of the values should lie between —2.0 and +2.0. The null hypothesis can be rejected in both cases. Figure 19.13c shows that comparison of taxonomic groups

Figure 19.13 Frequency histograms for standardized effect sizes measured for 96 presence-absence matrices taken from the literature in the case of (a) the C score and (b) the number of species pairs forming perfect checkerboard distributions. (c) Standardized effect sizes for the C score for different taxonomic groups. The dashed line indicates an effects size of 2.0, which is the approximate 5% significance level. (After Gotelli & McCabe, 2002.)

Figure 19.13 Frequency histograms for standardized effect sizes measured for 96 presence-absence matrices taken from the literature in the case of (a) the C score and (b) the number of species pairs forming perfect checkerboard distributions. (c) Standardized effect sizes for the C score for different taxonomic groups. The dashed line indicates an effects size of 2.0, which is the approximate 5% significance level. (After Gotelli & McCabe, 2002.)

plants and homeothermic vertebrates tend to have higher SESs for the C score, indicating stronger tendencies towards negative species associations than the poikilotherms have (invertebrates, fish and reptiles), with the exception of ants.

Gotelli and McCabe (2002) do not go so far as to claim they have performed a definitive test of the role of competition. They note that some species may exhibit 'habitat checkerboards'

because they have affinities to nonoverlapping habitats. Others may reveal 'historical checkerboards', co-occurring infrequently because of restricted dispersal since allopatric speciation (i.e. having speciated in different places). However, these results add further weight to a widespread role for competition in structuring communities.

Was this article helpful?

0 0
Lawn Care

Lawn Care

The Secret of A Great Lawn Without Needing a Professional You Can Do It And I Can Show You How! A Great Looking Lawn Doesnt Have To Cost Hundreds Of Dollars Or Require The Use Of A Professional Lawn Care Service. All You Need Is This Incredible Book!

Get My Free Ebook


Responses

  • Gaudenzia
    Do competitors show negatively associated distributions?
    4 years ago

Post a comment