Abstract
Aim The species–area relationship ( SAR ) is often modelled by the linearized power function log S = log c + z log A , where S is species richness, A is area, log c is the intercept and z is the slope. Although investigating how c and z values vary across taxa and archipelagos can provide insights into the biology of the SAR , this approach has many caveats. In this study, we aim to clarify how and why SAR s should be properly compared for the same taxon among different areas, or among different taxa in the same area. Location Mediterranean. We considered 18–46 Tyrrhenian islands (0.000024–223 km 2 ) and 32–65 Aegean islands (0.0058–8261 km 2 ). Methods We used OLS regressions to estimate c and z values for various taxonomic groups: land snails, isopods, centipedes, tenebrionids and reptiles. We used ANCOVA s to test (1) if different taxa have different z and c values within the same island group (possibly due to their dispersal ability and ecological characteristics), and (2) if the same taxon has different z and c values in different island groups (possibly due to differences in historical processes and isolation). Results z varied between 0.141 and 0.309, while c varied between 2.717 and 12.286 species per unit area (1 km 2 ). For tenebrionids, centipedes and land snails, we found higher c values in the Tyrrhenian islands than in the Aegean islands. Overall, c values were highest for land snails. Main conclusions Our results demonstrate the importance of comparing SAR s either of different groups within the same area, or of the same group in different areas. Furthermore, we identify the intercept, rather than the slope, as being dependent on the biogeographical dynamics (relict versus equilibrium faunas) and species ecology (dispersal capabilities and population abundance).