Influence of island geography,age and landscape on species composition in different animal groups

Aim To study the importance of ecological and geographical factors in explaining arthropod species composition on islands. Location The Aeolian Islands, a volcanic archipelago in the central Mediterranean, near Sicily. Methods The influence of island area, age, distance to the mainland, distance to the nearest island and land cover categories on species composition of arthropod groups was analysed using canonical correspondence analysis (CCA). The use of multiple animal groups in the same archipelago allowed the development of two complementary approaches based on CCA – a ‘taxon‐focused’ approach and an ‘island‐focused’ approach – to elucidate, respectively, how different taxa respond to the same environmental factors, and which factors are mainly responsible for the composition of the faunas in different locations. Results Island area was an important factor in explaining species composition in Chilopoda, Orthoptera and Tenebrionidae. Distance to the mainland was important mainly for Carabidae. Distance to the closest island was important for many groups. By contrast, island age exerted a significant influence only on the species composition of Orthoptera. Various groups were influenced by a combination of broad‐leaved forest and natural grassland. Main conclusions The example of the arthropods of the Aeolian Islands indicates that the influence of a given island characteristic on species composition varies among groups, although measures of inter‐island isolation were typically more important for taxa than isolation from the mainland source. This suggests that colonization of islands may occur mostly by stepping‐stone dispersal.     

A contribution to the knowledge of myxomycetes diversity in volcanic islands

A list of myxomycetes collected in Pantelleria, a volcanic island located 110 km southwest of the island of Sicily (Italy), is presented with data on distribution. Forty-nine taxa were identified, 38 of which are new for Sicily. The myxomycete diversity of Pantelleria was also compared with that of other volcanic islands located in the Atlantic and Pacific oceans.     

Orchids of the West Indies: predictability of diversity and endemism

Aim We examined phytogeographical patterns of West Indian orchids, and related island area and maximum elevation with orchid species richness and endemism. We expected strong species–area relationships, but that these would differ between low and montane island groups. In so far as maximum island elevation is a surrogate for habitat diversity, we anticipated a strong relationship with maximum elevation and both species richness and endemism for montane islands. Location The West Indies. Methods Our data included 49 islands and 728 species. Islands were classified as either montane (≥ 300 m elevation) or low (< 300 m). Linear and multivariate regression analyses were run to detect relationships between either area or maximum island elevation and species richness or the number of island endemic species. Results For all 49 islands, the species–area relationship was strong, producing a z‐value of 0.47 (slope of the regression line) and explaining 46% of the variation. For 18 relatively homogeneous, low islands we found a non‐significant slope of z = −0.01 that explained only 0.1% of the variation. The 31 montane islands had a highly significant species–area relationship, with z = 0.49 and accounting for 65% of the variation. Species numbers were also strongly related to maximum island elevation. For all islands < 750 km², we found a small‐island effect, which reduced the species–area relationship to a non‐significant z = 0.16, with only 5% of the variation explained by the model. Species–area relationships for montane islands of at least 750 km² were strong and significant, but maximum elevation was the best predictor of species richness and accounted for 79% of the variation. The frequency of single‐island endemics was high (42%) but nearly all occurred on just nine montane islands (300 species). The taxonomic distribution of endemics was also skewed, suggesting that seed dispersability, while remarkable in some taxa, is very limited in others. Montane island endemics showed strong species–area and species–elevation relationships. Main conclusions Area and elevation are good predictors of orchid species diversity and endemism in the West Indies, but these associations are driven by the extraordinarily strong relationships of large, montane islands. The species richness of low islands showed no significant relationship with either variable. A small‐island effect exists, but the montane islands had a significant relationship between species diversity and maximum elevation. Thus, patterns of Caribbean orchid diversity are dependent on an interplay between area and topographic diversity.     

island biogeography of Day Geckos (Phelsuma) in the Indian Ocean

Summary Step-wise multiple regression was employed to probe the determinants of species diversity of day geckos (Phelsuma) in the Indian Ocean. Independent variables were area, elevation, and two measures of isolation. Distance from Madagascar and island height (an indicator of habitat diversity) were the two most important predictors of species richness. Similar studies on other taxa rarely find isolation to be a major factor. The relatively poor dispersal abilities of reptiles may explain why isolation, rather than attributes of the islands, are more important in this case. The regressions also indicate that habitat diversity (assumed to correlate with maximum island elevation) is more important than area per se in determining species diversity. These results agree with predictions of the equilibrium theory of island biogeography, but historical processes have also greatly influenced species richness.     

Organization of Squamata (Reptilia) assemblages in Mediterranean archipelagos

Mediterranean islands have complex reptile assemblages, but little is known about the factors that determine their organization. In this study, the structure of assemblages of Squamata was evaluated based on their species richness and two measures of phylogenetic diversity (variability and clustering). I evaluated the composition of the assemblages comparing distinct biogeographic subregions within the Mediterranean: Adriatic, Aegean, Balearic, Corsica–Sardinia, Crete, Gulf of Gabés, Ionian Sea, Ligurian Sea, Malta, Sicily, and Tyrrhenian Sea. The effect of island environments and geographical isolation on the diversity metrics was assessed using generalized linear models. The analyses indicated that species richness was mostly influenced by island area and geographical isolation. Assemblages on smaller islands were poorer in species and phylogenetically dispersed, possibly as an effect of interspecific competition. The species composition of the assemblages was determined by similar environmental drivers within the biogeographic subregions, including island area, island elevation, geographical isolation, and aridity. In several subregions, significant patterns of phylogenetic attraction were found in species co‐occurrences, caused by the limits imposed by the island size on large predatory species.     

History and the species-area relationship in Lesser Antillean birds

We examined the species-area relationship for three historically distinct subsets of Lesser Antillean birds identified by molecular phylogenetic analysis of island and continental populations. The groups comprised recent colonists from continental or Greater Antillean source populations, old taxa having recently expanded distributions within the Lesser Antilles, and old endemic taxa lacking evidence of recent dispersal between islands. The number of young taxa was primarily related to distance from the source of colonists in South America. In a multiple regression, the logarithmic slope of the species-area relationship for this group was shallow (0.066+/-0.016). Old endemic taxa were restricted to islands with high elevation, and within this subset, species richness was related primarily to island area, with a steep slope (0.719+/-0.110). The number of recently spread endemic taxa was related primarily to island elevation, apparently reflecting the persistence of such populations on islands with large areas of forested and montane habitats. Historical analysis of the Lesser Antillean avifauna supports the dynamic concept of island biogeography of MacArthur and Wilson, rather than the more static view of David Lack, in that colonists exhibit dispersal limitation and extinction plays a role in shaping patterns of diversity. However, the avifauna of the Lesser Antilles is probably not in equilibrium at present, and the overall species-area relationship might reflect changing proportions of historically distinguishable subsets of species.     

Butterflies of European islands: the implications of the geography and ecology of rarity and endemicity for conservation

Depending on their faunal content islands can function as important ‘vehicles’ for conservation. In this study, we examine data on 440 butterfly species over 564 European islands in 10 island groups. To determine the status of the butterfly fauna, we have adopted two approaches, island-focused and species-focused, examined using principal components analysis and regression modelling. In the former, we relate species richness, rarity and endemicity to island geography (area, elevation, isolation and location in latitude and longitude); in the latter, species occurrence on islands is examined in relation to distribution, range, range boundaries, and altitudinal limits on the continent as well as species’ ecology (number of host plants) and morphology (wing expanse). Species on islands are also assessed for their status on the continental mainland, their distributional dynamics (extinctions, distribution changes) and conservation status (Red Data Book, European Habitat Directive, Species of European Conservation Concern and Bern Convention listing. Unexpectedly, we find that a large fraction of the European butterfly species is found on the islands (63.4%; 59% on small islands) comprising some 6.2% of the land area of Europe. Although species occurring on the islands tend, on the whole, to have lower conservation status and are not declining over Europe, 45 species are endemics restricted to the islands. Species richness shows only a weak locational pattern and is related as expected to isolation from the continental source and island area; but, both rarity and endemicity have distinctive geographical bias to southern Europe, on islands now under increasing pressure from climate change and increasingly intensive human exploitation. The vulnerability of species on islands is emphasised in the relationship of island occurrence (% occurrence and presence/absence of species on any island) with continental distributions. A large proportion of the variation (84%) is accounted by continental distribution, the southern range limit and lower altitudinal limit. Most species (69%) occur on very few islands (<5%). In view of ongoing species dynamics on islands, migrations and extinctions of species, island repositories of species depend in large part on conservation of butterflies at continental sources. The unique faunas and rare species on islands also depend on appropriate concern being given to the island faunas. Conservation of European islands is thus a two-way process, sustaining sources and conserving island refuges. Residuals from the regressions (islands with more or fewer species, rare and endemic species; species occurring more or less frequently than expected on islands) provide warning signals of regions and islands deserving immediate attention.     

Biogeographical determinants for total and endemic species richness in a continental archipelago

We examined the relationship between plant species richness and biogeographical variables (island area, island maximum elevation, distance from nearest inhabited island, distance from nearest mainland) using a data set comprising 201 islands of the Aegean archipelago. We found that endemic species richness was strongly correlated to total species richness. Single-island endemic species richness was most strongly correlated to island maximum elevation, and then to island area, with an apparent small island effect for islands smaller than 47 km². Total species richness was most strongly correlated to island area (with no apparent small island effect), and less strongly correlated to island maximum elevation. Distance from the mainland or other inhabited islands displayed limited predictive value in our data set. The slope of the relationship between species richness and geographical factors (island area, elevation, distance from island/mainland) was steeper for endemic species richness than for total richness. Finally, the different scales of endemicity (single-island endemics, island group endemics and Aegean regional endemics) displayed similar qualitative trends and only differed quantitatively. Thus, we conclude that different biogeographical factors act as drivers for total species richness than for endemic species richness.     

Mammals of Australian islands: factors influencing species richness

Distribution patterns of indigenous non-volant terrestrial mammals on 257 Australian islands were examined in relation to environmental parameters and the effects of human-induced disturbance during prehistoric and historic times on island species numbers. Species occurrence for individual species, for taxonomic and trophic groups, and for all species together was related to environmental parameters using regression analysis and the extreme-value function model. Patterns of occurrence were examined separately within three major biogeographic regions derived by pattern analysis. The number of species known to have occurred on these islands during historic times was adequately predicted from area alone. No statistically significant improvement in predicted species number was gained by including island elevation, mean annual rainfall, isolation from the mainland or the number of potentially competing species present on the island. Similarly, no single factor other than area was found to influence consistently the presence of individual species. We conclude that the occurrence of indigenous non-volant terrestrial mammal species on these islands indicates a relictual rather than equilibrial fauna. Visitation by Aboriginal people during prehistoric times did not significantly increase mammal extinctions on islands. Examination of patterns of species richness for a given area on a regional basis showed that islands in and around Bass Strait and Tasmania (Bass Region) were the most species-rich, islands off the northern coasts were slightly less rich, and islands off the south western coasts had fewest species. This is in contrast to the usual latitudinal gradient in species richness patterns. However, islands off the northern and eastern coasts had an overall greater number of different species. When considered in relation to the number of different species of mammals occurring within each region, islands of a given size in Bass Region typically bore a higher proportion of this species pool than other regions. The Bass Region was found to be particularly rich in macropoid herbivores and dasyurid carnivores and insectivores. Analyses indicated that there is a very strong relationship between the presence of exotics as a whole and the local extinction of native mammals. Many mammal species formerly widespread on the Australian mainland are now restricted totally to islands (nine species) or are threatened with extinction on the mainland and have island populations of conservation significance (ten species). In all, thirty-five islands protect eighteen taxa of Australian threatened mammals. The land-use and management of these islands is of considerable importance to nature conservation. The introduction of exotic mammals to these islands should be prevented; any introductions that occur should be eradicated immediately.     

Patterns of island treeline elevation – a global perspective

Treeline research has strongly focused on mountain systems on the mainland. However, island treelines offer the opportunity to contribute to the global framework on treeline elevation due to their island‐specific attributes such as isolation, small area, low species richness and relative youth. We hypothesize that, similar to the mainland, latitude‐driven temperature variation is the most important determinant of island treeline elevation on a global scale. To test this hypothesis, we compared mainland with island treeline elevations. Then we focused 1) on the global effects of latitude, 2) on the regional effects of island type (continental vs oceanic islands) and 3) the local effects of several specific island characteristics (age, area, maximum island elevation, isolation and plant species richness). We collected a global dataset of islands (n = 86) by applying a stratified design using GoogleEarth and the Global Island Database. For each island we extracted data on latitude and local characteristics. Treeline elevation decreased from the mainland through continental to oceanic islands. Island treeline elevation followed a hump‐shaped latitudinal distribution, which is fundamentally different from the mainland double‐hump. Higher maximum island elevation generated higher treeline elevation and was found the best single predictor of island treeline elevation, even better than latitude. Lower island treeline elevation may be the result of a low mass elevation effect (MEE) influencing island climates and an increasingly impoverished species pool but also trade wind inversion‐associated aridity. The maximum island elevation effect possibly results from an increasing mass elevation effect (MEE) with increasing island elevation but also range shifts during climatic fluctuations and the summit syndrome (i.e. high wind speeds and poor soils in peak regions). Investigating islands in treeline research has enabled disentangling the global effect of latitude from regional and local effects and, at least for islands, a comprehensive quantification of the MEE.     

Drivers of diversity in Macaronesian spiders and the role of species extinctions

Aim To identify the biogeographical factors underlying spider species richness in the Macaronesian region and assess the importance of species extinctions in shaping the current diversity. Location The European archipelagos of Macaronesia with an emphasis on the Azores and Canary Islands. Methods Seven variables were tested as predictors of single‐island endemics (SIE), archipelago endemics and indigenous spider species richness in the Azores, Canary Islands and Macaronesia as a whole: island area; geological age; maximum elevation; distance from mainland; distance from the closest island; distance from an older island; and natural forest area remaining per island – a measure of deforestation (the latter only in the Azores). Different mathematical formulations of the general dynamic model of oceanic island biogeography (GDM) were also tested. Results Island area and the proportion of remaining natural forest were the best predictors of species richness in the Azores. In the Canary Islands, area alone did not explain the richness of spiders. However, a hump‐shaped relationship between richness and time was apparent in these islands. The island richness in Macaronesia was correlated with island area, geological age, maximum elevation and distance to mainland. Main conclusions In Macaronesia as a whole, area, island age, the large distance that separates the Azores from the mainland, and the recent disappearance of native habitats with subsequent unrecorded extinctions seem to be the most probable explanations for the current observed richness. In the Canary Islands, the GDM model is strongly supported by many genera that radiated early, reached a peak at intermediate island ages, and have gone extinct on older, eroded islands. In the Azores, the unrecorded extinctions of many species in the oldest, most disturbed islands seem to be one of the main drivers of the current richness patterns. Spiders, the most important terrestrial predators on these islands, may be acting as early indicators for the future disappearance of other insular taxa.     

Species richness and structure of ant communities in a dynamic archipelago: effects of island area and age

Aim To assess how ant species richness and structure of ant communities are influenced by island age (disturbance history) in a dynamic archipelago. Location Cabra Corral dam, Salta Province, north‐west Argentina (25°08′ S, 65°20′ W). Methods Ant species richness on remaining fragments (islands) of a flooded forest was determined, as well as island area, isolation and age. Simple linear regressions were performed to assess relationships between ant species richness and those insular variables. Furthermore, a stepwise multiple linear regression analysis was conducted in order to determine the relative influence of each insular variable on ant species richness. Islands were categorized in two age classes (old and young) and co‐occurrence analyses were applied within each class to evaluate changes in community structure because of interspecific competition. Results Simple regression analyses indicated a moderate, positive effect of island area on ant species richness. Weak, marginally non‐significant relationships were found between ant species richness and both island isolation and island age, showing the tendency for there to be a decrease in ant species richness with island isolation and that ant species richness might be higher in old islands. The multiple regression analysis indicated that island isolation and age had no significant effects on the number of ant species, island area being the only independent variable retained in the analysis. On the contrary, whereas a random pattern of species co‐occurrence was found on young islands, ant communities in old islands showed a significantly negative pattern of species co‐occurrence, suggesting that the effect of competition on community structure was stronger on older islands than on younger islands. Main conclusions Island area was the most important variable explaining ant species richness on the islands of Cabra Corral dam. However, both island isolation and island age (or disturbance history) might also contribute to shape the observed community patterns. The present study also shows that island age significantly affects the strength with which interspecific interactions structure ant communities on islands.     

Plant species diversity and polyploidy in islands of natural vegetation isolated in extensive cultivated lands

Natural polyploidy is often related to a longer life span, vegetative reproduction and higher competitive ability. In this paper, we test the possibility that these characteristics may favour the survival of polyploid taxa under conditions of long-term habitat fragmentation. In islands of natural vegetation isolated in extensive vineyards located in the South of France and in a large neighbouring area of natural vegetation, plant species richness and the relative abundance of polyploid taxa were assessed according to island size, isolation and vegetation structure. High species richness was observed, with numerous species restricted to the islands, suggesting that these may constitute refugia. However, species richness was not related to island size or to degree of isolation except for the flora of the woody areas. A very positive effect of area fragmentation on plant richness was observed, which is probably attributable to relatively low species overlap among the islands. Particularly high species richness was observed in open areas, provided that these were not extensively colonized by shrubs which seem to be responsible for local extinction of many annual taxa. Polyploids, which comprised mostly perennial herbs and woody species, were predominant in all the islands and in the large reference area. In open habitats invaded by shrubby species, a higher relative frequency of polyploids was observed in islands than in the reference area. Moreover, polyploid taxa were present in a larger number of islands than the diploid taxa, which were often restricted to a single island, suggesting that, after a long period of isolation, the polyploids may still have a lower probability of extinction. Evidence was obtained from vegetation structure analysis that diploid and polyploid annual herbs were restricted to open habitats and were both eliminated by shrubby species. Conversely, the diploid perennial herbs were also significantly affected by shrub colonization whereas the polyploids were mostly present in shrubby areas. This suggests that the higher competitive ability of polyploid perennial herbs may constitute a critical factor responsible for their wider distribution over the islands. We report the implications of our findings on conservation strategies, more particularly for a Mediterranean flora.     

How island size and isolation affect bee and wasp ensembles on small tropical islands: a case study from Kepulauan Seribu,Indonesia

Aim To study the effects of isolation and size of small tropical islands on species assemblages of bees (superfamily Apoidea) and wasps (superfamily Vespoidea). Location Twenty islands in the Kepulauan Seribu Archipelago off the coast of west Java, Indonesia. The size of surveyed islands ranges between 0.75 and 41.32 ha; their distance from the coast of Java varies between 3 and 62 km. Methods Field work was conducted from February to May 2005. Bees and wasps were caught with a sweep net during sampling units of 15 min, continuing until four consecutive samples revealed no new species. Total species richness was quantified by the estimators Chao 2, first‐order jackknife and Michaelis–Menten. The software binmatnest was used to test for nestedness of species assemblages. Similarities of species composition between islands were quantified by Sørensen’s similarity index. Results Eighty‐two species were recorded on the 20 surveyed islands. Species richness declined with increasing isolation of islands from the source area, Java. Although the size of the largest island exceeded that of the smallest island by a factor of almost 60, island size only very weakly affected species richness of bees; no effect of island size was found for wasps. Mean body size of species decreased with increasing island isolation. Nestedness of island faunas was only weakly developed. Species composition of both superfamilies was affected by island isolation, but not by island size. Main conclusions While the species–isolation relationship on the very small islands of Kepulauan Seribu followed the prediction of MacArthur and Wilson’s equilibrium theory, the absence of a species–area relationship indicated a weak ‘small‐island effect’, at least in wasps. The combination of an only weakly developed pattern of nested species subsets, the shift in species compositions and the decline of mean body size with increasing island isolation from the source area indicates that biotic interactions and different species traits contribute to the shaping of communities of bees and wasps within the archipelago. The potential of biotic interactions for generating distribution patterns of species within the archipelago is also emphasized by the observed restriction of some species with apparently high dispersal abilities to outer islands.     

Human activities alter biogeographical patterns of reptiles on Mediterranean islands

Aim The theory of island biogeography predicts species richness based on geographical factors that influence the extinction–colonization balance, such as area and isolation. However, human influence is the major cause of present biotic changes, and may therefore modify biogeographical patterns by increasing extinctions and colonizations. Our aim was to evaluate the effect of human activities on the species richness of reptiles on islands. Location Islands in the Mediterranean Sea and Macaronesia. Methods Using a large data set (n = 212 islands) compiled from the literature, we built spatial regression models to compare the effect of geographical (area, isolation, topography) and human (population, airports) factors on native and alien species. We also used piecewise regression to evaluate whether human activities cause deviation of the species–area relationship from the linear (on log–log axes) pattern, and path analysis to reveal the relationships among multiple potential predictors. Results The richness of both native and alien species was best explained by models combining geographical and human factors. The richness of native species was negatively related to human influence, while that of alien species was positively related, with the overall balance being negative. In models that did not take into account human factors, the relationship between island area and species richness was not linear. Large islands hosted fewer native species than expected from a linear (on log–log axes) species–area relationship, because they were more strongly affected by human influence than were small islands. Path analysis showed that island size has a direct positive effect on reptile richness. However, area also had a positive relationship with human impact, which in turn mediated a negative effect on richness. Main conclusion Anthropogenic factors can strongly modify the biogeographical pattern of islands, probably because they are major drivers of present‐day extinctions and colonizations and can displace island biodiversity from the equilibrium points expected by theory on the basis of geographical features.     

Prehistoric species richness of birds on oceanic islands

Using data on prehistoric and modern birds from seven islands in the Kingdom of Tonga, we demonstrate that there is no positive relationship between species richness (S) and island area (A) over the observed range of A (1.8–259 km²). The uniform S‐values occur across more than three orders of magnitude of A when prehistoric data are included, and the strongest predictor of S on any island is the level of fossil sampling (number of identified bones). Below a minimum value for A (in Tonga < 1.8 km²), S declines to zero as A does the same. Within the ranges of island elevation (E) and inter‐island isolation (I) among the seven islands, neither E (11–312 m) nor I (0.6–38 km) has much if any effect on S. Under natural (pre‐human) conditions, a positive species‐area relationship may not be a valid generalization for birds on oceanic islands.     

Richness and composition of plants and birds on land‐bridge islands: effects of island attributes and differential responses of species groups

Aim To test relationships between the richness and composition of vascular plants and birds and attributes of habitat fragments using a model land‐bridge island system, and to investigate whether the effects of fragmentation differ depending on species natural history traits. Location Thousand Island Lake, China. Methods We compiled presence/absence data of vascular plant and bird species through exhaustive surveys of 41 islands. Plant species were assigned to two categories: shade‐intolerant and shade‐tolerant species; bird species were assigned to three categories: edge, interior, and generalist species. We analysed the relationships between island attributes (area, isolation, elevation, shape complexity, and perimeter to area ratio) and species richness using generalized linear models (GLMs). We also investigated patterns of composition in relation to island attributes using ordination (redundancy analysis). Results We found that island area explained a high degree of variation in the species richness of all species groups. The slope of the species–area relationship (z) was 0.16 for all plant species and 0.11 for all bird species. The lowest z‐value was for generalist birds (0.04). The species richness of the three plant species groups was associated with island area per se, while that of all, generalist, and interior birds was explained mainly by elevation, and that of edge bird species was associated primarily with island shape. Patterns of species composition were most strongly related to elevation, island shape complexity, and perimeter to area ratio rather than to island area per se. Species richness had no significant relationship with isolation, but species composition did. We also found differential responses among the species groups to changes in island attributes. Main conclusions Within the Thousand Island Lake system, the effects of fragmentation on both bird and plant species appear to be scale‐dependent and taxon‐specific. The number of plant species occurring on an island is strongly correlated with island area, and the richness of birds and the species composition of plants and birds are associated with variables related to habitat heterogeneity. We conclude that the effects of fragmentation on species diversity and composition depend not only on the degree of habitat loss but also on the specific patterns of habitat fragmentation.     

Increase of island endemism with altitude – speciation processes on oceanic islands

Understanding speciation on oceanic islands is a major topic in current research on island biogeography. Within this context, it is not an easy task to differentiate between the influence of elevation as an indicator for habitat diversity and island age as an indicator for the time available for diversification. One reason for this is that erosion processes reduce the elevation of islands over time. In addition, the geographic distance to source ecosystems might differ among habitats, which could lead to habitat‐specific reduction of species immigration, niche occupation and diversification. We used the percentage of single island endemic species (pSIE) in five different zonal ecosystems (distributed in altitude) on the Canary Islands as an indicator for diversification. We tested whether diversification increases with altitude due to a greater ecological isolation of high elevation ecosystems on oceanic islands under the assumption of a low elevation source region on the mainland. In addition we tested whether the ‘hump‐shaped’ (unimodal) relationship between pSIE and island age as well as the linear relationship between species richness and pSIE is consistent across spatial scales. We also analyse a potential influence of island area and habitat area. We found that pSIE increases with elevation. The relations between species richness as well as age with pSIE are consistent across scales. We conclude that high elevation ecosystems are ecologically isolated. Surprisingly, the altitudinal belt with the strongest human influences has the highest values of pSIE. We successfully transfer the ‘general dynamic theory of island biogeography’ to the ecosystem scale, which provides multiple opportunities for future studies. With this approach we find that the effects of elevation on diversification can be separated from those of island age.     

Time, area and isolation: factors driving the diversification of Azorean arthropods

Aim R. J. Whittaker et al. recently proposed a ‘general dynamic model of oceanic island biogeography’ (GDM), providing a general explanation of island biodiversity patterns by relating fundamental biogeographical processes – speciation, immigration, extinction – to area (A) and time (T; maximum island geological age). We adapt their model, which predicts a positive relationship with area combined with a humped relationship to time (designated the ATT² model), to study the factors promoting diversification on the Azores for several arthropod groups. Location The Azorean archipelago (North Atlantic; 37–40° N, 25–31° W). Methods We use the number of single‐island endemics (SIEs) as a measure of diversification, to evaluate four different predictions for the variation in SIEs between different islands, derived from the GDM theory and our knowledge of the fauna and history of the Azores. We calculated the number of SIEs for seven out of the nine Azorean islands and six groups of species (all arthropods, beetles, cavernicolous and non‐cavernicolous species, and taxa with high and low dispersal abilities). Several variables accounting for island characteristics (area, geological age, habitat diversity and isolation) and generalized linear models were used to evaluate the reliability of each prediction. Results A linear and positive relationship between SIEs and an AT (area + time) model was the most parsimonious explanation for overall arthropod diversification. However, cavernicolous species showed the opposite pattern (more SIEs inhabiting the youngest islands). Also, isolation was an important predictor of diversification for all groups except for the species with high dispersal ability; while the former were negatively related to the distance from the main source of colonizing lineages (Santa Maria island in most cases), the latter were related to area. Dispersal ability was also a key factor affecting the diversification of most groups of species. Main conclusions In general, the diversification of Azorean arthropods is affected by age, area and isolation. However, different groups are affected by these factors in different ways, showing radically different patterns. Although the ATT² model fails to predict the diversification pattern of several groups, it provides a framework for integrating these deviations into a general theory. Further improvements of the GDM theory need to take into account the particular traits of each group and the role of isolation in shaping island diversity.     

The influence of geographical and ecological factors on island beta diversity patterns

Aim To investigate the importance of various island characteristics in determining spatial patterns of variations in beta diversity for various animal groups. Location Analyses are presented for 10 animal groups living on the Aeolian Islands, a volcanic archipelago in the central Mediterranean, near Sicily. Methods Three hypotheses were formulated to explain patterns of beta diversity: the target‐area–distance effect, stepping stone dispersal and island age. Matrices of inter‐island dissimilarities were constructed under each hypothesis and correlated with matrices of faunal dissimilarities using Mantel tests. For the ‘target‐area–distance effect’ hypothesis, inter‐island dissimilarities were calculated using island sizes and distances to nearest mainland areas. For the ‘stepping stone dispersal’ hypothesis, inter‐island distances were measured. Finally, for the ‘island age’ hypothesis, inter‐island dissimilarities were calculated on the basis of the geological age of the islands. Cluster analysis was used to investigate inter‐island faunal relationships. Results Support for a target‐area–distance effect was found only for birds. For these highly mobile animals, inter‐island distances had no significant effects on beta diversity. Birds are known to colonize islands by crossing large sea barriers and thus they can easily reach the Aeolian Islands, which are close to source areas (notably Sicily). Inter‐island distances had a significant role in determining patterns of beta diversity in most invertebrates. For Mollusca, Opiliones, Chilopoda, Heteroptera, coprophagous Scarabaeoidea, and Tenebrionidae, even relatively short distances preclude invertebrates from colonizing an island regularly from the mainland, and most colonization probably results from inter‐island faunal exchanges. Island age was proved to be important only for orthopterans. Main conclusions The origin of most of the Aeolian invertebrate fauna is quite recent, and species appear to have established on the islands predominantly by stepping stone dispersal. Birds, which are highly mobile organisms, follow more direct mainland–island dynamics. As further studies on other islands become available, comparative analyses will confirm whether the factors influencing variations in beta diversity in this study and their relationships with species dispersal ability are consistent across scales and geographical context.