Aim Some species reach extraordinary densities on islands. Island assemblages have fewer species, however, and it is possible that island species differ from their mainland counterparts in average mass. Island assemblages could be partitioned differently (fewer species or smaller individuals) from mainland sites without differing in aggregate biomass (density compensation). Our objective was to determine the generality of excess density compensation in island herpetofaunal assemblages.

Location Our bounded removal plot data were obtained from Pacific Island sites (Guam, Saipan and Rota), the West Indies (British Virgin Islands), and the Indian Ocean (Ile aux Aigrettes off Mauritius). The literature values were taken from several locales. Other island locations included Barro Colorado Island, Bonaire, Borneo, Philippine Islands, Seychelle Islands, Barrow Island (Australia), North Brother Island (New Zealand), Dominica and Puerto Rico. Mainland sites included Costa Rica, Ivory Coast, Cameroon, Australia, Thailand, Peru, Brazil, Panama and the USA.

Method We added our thirty-nine bounded total removal plots from sixteen island habitats to fifteen literature records to obtain seventy-five venues with estimable density and biomass of arboreal or terrestrial herpetofaunal assemblages. These biomass estimates were evaluated geographically and in relation to sampling method, insularity, latitude, disturbance regime, seasonality, community richness, vegetative structure and climate. Direct data on trophic interactions (food availability, parasitism and predation pressure) were generally unavailable. Sampling problems were frequent for arboreal, cryptic and evasive species.

Results and main conclusions We found strong evidence that herpetofaunal assemblages on small islands (mostly lizards) exhibit a much greater aggregate density of biomass (kg ha⁻¹) than those of larger islands or mainland assemblages (small islands show excess density compensation). High aggregate biomass density was more strongly associated with the degree of species impoverishment on islands than it was on island area or insularity per se. High aggregate biomass density was not strongly associated with latitude, precipitation, canopy height or a variety of other physical characteristics of the study sites. The association between high aggregate biomass density and species-poor islands is consistent with the effects of a reduced suite of predators on depauperate islands, but other features may also contribute to excess density compensation.