The amount of carbon stored in deadwood is equivalent to about 8% of global forest carbon stocks 1 . Deadwood decomposition is largely governed by climate [2][3][4][5] with decomposer groups, such as microbes and insects, contributing to variations in decomposition rates 2,6,7 . At the global scale, the contribution of insects to deadwood decomposition and carbon release remains poorly understood 7 . Here we present a field experiment of wood decomposition across 55 forest sites on six continents. We find that deadwood decomposition rates increase with temperature, with the strongest temperature effect at high precipitation levels. Precipitation affects decomposition rates negatively at low temperature and positively at high temperatures. As net effect, including direct consumption and indirect effects via interactions with microbes, insects accelerate decomposition in tropical forests (3.9% median mass loss per year).In temperate and boreal forests we find weak positive and negative effects with a median mass loss of 0.9% and -0.1% per year, respectively. Furthermore, we apply the experimentally derived decomposition function to a global map of deadwood carbon synthesised from empirical and remote sensing data. This allows for a first estimate of 10.9 ± 3.2 Pg yr -1 of carbon released from deadwood globally, with 93% originating from tropical forests. Globally, the net effect of insects accounts for a carbon flux of 3.2 ± 0.9 Pg yr -1 or 29% of the total carbon released from deadwood, which highlights the functional importance of insects for deadwood decomposition and the global carbon cycle.

Aim The aim of this study was to assess the influence of island area, distance to source pool, latitude, habitat diversity and habitat type on species richness of forest birds in the Andaman islands. Location The Andaman islands (India) in the Bay of Bengal. Methods The distributions of 47 species of forest birds were surveyed on 45 islands in the Andaman islands across a latitudinal gradient. The size of the island and distance to the nearest large island were assessed on a satellite image of 1 : 250,000 scale. The number and types of habitats and the species richness of birds on each island were recorded during a field survey. The effects of the variables measured on the species richness of forest birds were assessed using regression analyses. The best fit models were selected for interpretation of the results. Separate analyses were conducted with selected islands to eliminate the effects of latitude and to control the effects of area and habitat diversity. Results The number of species of forest birds was strongly influenced by island area and habitat diversity. However the key determinant of species richness was habitat type, particularly the presence of wet forests. Wet forests, either semi‐evergreen or evergreen tropical forests are more common towards the southern islands and are usually restricted to larger islands. Main conclusions Area, habitat diversity and the presence of wet forests on islands significantly influenced species richness of forest birds. The wet forests maintain the biodiversity of the Andamans and should be regarded as a ‘keystone habitat’. This is probably because wet forests are species rich and also because the Andaman biota has affinities with that of the Malay peninsula where wet forests predominate. Therefore biogeographical history probably plays an important role in influencing biodiversity at a regional scale.

Fruit biomass and frugivore abundance were quantified over 3 y in a rain forest of the south Western Ghats, India. Fruit biomass was estimated by sampling fruit fall in the primary forest, and frugivore abundance by a 2.5-km transect. A total of 645 kg ha−1 of fruit was produced annually in the forest. Only 49% of this is edible to the frugivores and the remaining 51% is in the form of non-edible husks. Mammalian frugivores outnumbered avian frugivores and the majority of the mammals were seed predators. The total fruit biomass produced at Kakachi is lower than in the lowland forest and mountain forests in the neotropics but higher than in the wet sclerophyll forest of Australia. Lower diversity of trees and edaphic factors at Kakachi could be some of the reasons for these differences. On the other hand, paucity of fleshy fruits, low density of trees producing fleshy fruits and irregular fruiting of these species, account for the low number of obligate avian frugivores at Kakachi.

The distributional patterns of forest birds and butterflies in the Andaman islands, an oceanic chain located off SE Asia, were tested for nestedness. Both taxa were highly nested. Nestedness could be due to colonization or extinction processes, area or distance effects or nestedness of habitats. Nestedness in forest bird distributions were strongly influenced by area and habitat related factors. Habitats were significantly nested in all three island groups, however most strongly for the North Andamans. However forest bird distributions in the North Andamans, as indicated by row order in the packed matrix, was not correlated with habitat diversity, suggesting that habitat related factors alone cannot account for these patterns. Other causal influences could be passive sampling, where common and abundant species and habitats are more likely to have a widespread distribution than rare species and habitats. The nested subset pattern seen in two unrelated taxa suggests that the Andamans are extinction dominated and that the protection of forests on large islands is critical for the conservation of its biodiversity.

) and 32 epibenthos (through a naturalist dredge 40 × 40 cm) were collected that revealed 169 species representing 15 diverse groups. Gastropods, bivalves, polychaetes and decapods constituted the bulk of the population. Polychaetes were by far the most dominant group (64.98%) among the infauna followed by amphipods (25.23%), which were numerically more abundant (mean 1080 ± 292 ind. m ). There was a preponderance of decapods (26%), gastropods (20.7%), bivalves (8.3%) and several (28.3%) largesized polychaetes (e.g. Diopatra neapolitana, Eunice indica, Pista sp.) in the dredge hauls (mean abundance 42 ± 8 ind. haul -1 ). Infaunal diversity (Shannon-Wiener H ') was higher (H ': 2.455 ± 0.18) at 51 to 75 m relative to the sites at >150 m depth (H ': 0.981 ± 0.17). Multivariate analyses were used to define assemblages named after the most important (determining) taxon. Three epifaunal associations could be recognised off the northeast Indian shelf, namely the Charybdis Assemblage (30 to 50 m), the Liagore Assemblage (51 to 75 m), and the Amygdalum watsoni -Tibia delicatula Assemblage (>100 m). Infaunal associations consisted of Ampelisca (30 to 75 m), Nephtys (76 to 100 m) and Cossura coasta (>100 m). Canonical correspondence analysis (CCA) showed that depth, sand, sediment organic matter and sediment mean size influenced epifauna distribution, whereas for infauna, salinity, temperature, mean particle diameter, sand and depth proved important.