Nonequilibrium Diversity Dynamics of the Lesser Antillean Avifauna Robert E. Ricklefs and Eldredge Bermingham

Note 14. For details of the molecular methods, see (1). Where samples were available, sequence was obtained from two individuals of each island and continental population. Genetic distances used in this analysis were based on the Tamura-Nei model of nucleotide substitution specified for each codon position (2). This model was selected by ModelTest (3) for all but one of the taxa included in this analysis and was therefore applied uniformly to all taxa. Relationships among populations within taxa were reconstructed by PAUP* using a neighbor-joining algorithm and midpoint rooting. Although we cannot compare rates between unrelated clades, nucleotide substitution rate did not violate the clock assumption within 38 of 39 clades. Because of the relatively small genetic distances among conspecific lineages and the concentration of substitutions at fourfold degenerate sites, we are comfortable that sequence divergence provides an index to relative time.

Note 15. Colonization events do not correspond exactly to species in the Lesser Antilles because four endemic thrashers (Margarops fuscatus, Allenia fusca, Cinclocerthia ruficauda complex, and Ramphocinclus brachyurus: Mimidae) (4) and two endemic hummingbirds (Eulampis jugularis and E. holosericeus: Trochilidae) (5) were produced by diversification within the archipelago from individual ancestral lineages. Four species include multiple colonization events: Columbina passerina (2), Elaenia flavogaster (2), Vireo altiloquus (2), and Troglodytes aedon (3).

Note 17. The greatest sequence divergence in the data set was for the Antillean crested hummingbird Orthorhynchus cristatus (d_A = 0.258), for which we could not identify a clear mainland or Greater Antillean sister taxon. Thus, d_A is a maximum value that might be reduced with increased sampling of potential outgroups. The next oldest sequence divergence (0.177) was for the Lesser Antillean tanager Tangara cucullata, whose sister taxon was also unclear. The genus is large, with 45 species in South America, of which 20 occur in Venezuela. Furthermore, the distance between Tangara and the next most-divergent lineage is significantly greater than expected. Excluding Orthorhynchus and Tangara, the average distance between successive genetic divergence values for d_A > 0.02 is 0.0072. Assuming an exponential distribution resulting from a homogeneous colonization process, the distance between Tangara (d_A = 0.177) and the next oldest lineage (0.143) is highly unlikely (P = 0.008). Thus, we dropped Orthorhynchus and Tangara from analyses to avoid questionable d_A values.

Note 30. Constant colonization is unrealistic because many species have poor powers of dispersal and the size of the mainland pool of species (P) places an upper limit to the cumulative number of extant colonists (L). Suppose that colonization rate decreases linearly with increasing diversity in the Lesser Antilles according to C = C₀(1 - L/P), which declines to 0 as L approaches the continental pool of species. C₀ is the initial colonization rate when there are no species in the archipelago (L = 0). In this case, the equilibrium number of colonists is C₀P/(C₀ + PM) and the cumulative number of colonists follows a function identical to that of Eq. 2, except that M is replaced by (M + C₀/P). When M = 0 and (M + C₀/P) is estimated to be close to 0, P must be large. For example, if C₀ = 1600, M = 0, and (M + C₀/P) = 4, as estimated in one of our analyses, then P would be 400, which greatly exceeds the contemporary number of land bird species the Lesser Antilles, not to mention the continental avifauna of Trinidad.

References

1. I. J. Lovette et al., Auk 115, 890 (1998).

2. K. Tamura, M. Nei, Mol. Biol. Evol. 10, 512 (1993).

3. D. Posada, K. A. Crandall, Bioinformatics 14, 817 (1998).

4. J. S. Hunt et al., Auk 118, 35 (2001).

5. E. Bermingham, R. E. Ricklefs, unpublished data.