MEPS 319:85-91 (2006)  -  doi:10.3354/meps319085

ABSTRACT: We examine relationships between species richness (S), rarity (R) and average taxonomic distinctness (Δ⁺) from analysis of a comprehensive dataset for benthic marine algae (including Cyanophyta). This comprises 2894 species from 66 sites across the Indian Ocean. Ranked values for the sites, determined according to the 3 metrics, show significant positive correlation (p ≤ 0.01); Mauritius, India and Aldabra emerge as biodiversity ‘hotspots’, while Indonesia (Nias Island), Maldives (Male Atoll) and the Gulf of Aden are ‘coldspots’. Concordance between metrics was unexpected, given their disparity in robustness to sampling rigour and particularly since Δ⁺ is conceptually unrelated to S and R. Lack of significant latitudinal correlations was also evident except for Δ⁺, which increased towards temperate waters in the southern hemisphere. This contrasts with the variable patterns observed with longitude, for which significant correlations (negative, i.e. towards the west) were prevalent only for S (algae overall and separate categories except Phaeophyta), evident for R (Cyanophyta only) and absent for Δ⁺. Hence, use of one floral category as a surrogate for biodiversity in another is not guaranteed. Aquatic biodiversity patterns are complex, in accordance with recent findings derived mainly from faunal datasets. Relationships between different metrics can depend on both the group(s) selected and the environmental or geographical factor(s) examined. Our study is based on analysis of extensive but low resolution (presence/absence) data (Silva et al. 1996) collected from sites of variable size that were not sampled evenly. We address these constraints, but acknowledge the possibility that some patterns may prove to be artefacts, pending analysis of data from recent and ongoing studies. However, we do not expect this to significantly affect our overall conclusions.

KEY WORDS: Biodiversity · Species richness · Rarity · Average taxonomic distinctness · Latitudinal and longitudinal patterns · Hotspots