Range expansion by moose into coastal temperate rainforests of British Columbia, Canada

Ranges of species are dynamic and respond to long‐term climate change and contemporary effects such as habitat modification. We report here that moose (Alces alces) have recently colonized coastal temperate rainforests of British Columbia, Canada. Contrary to recent publications, field observations of moose and their sign, combined with their occurrence in wolf (Canis lupus) faeces, suggest that moose are now widespread on the coastal mainland and occur on least three islands. Traditional ecological knowledge (information accumulated by aboriginal peoples about their environment) suggests that colonization occurred during the mid 1900s, concomitant with logging of major watersheds that bisect the Coast Mountain Range. Range expansion by moose may have ecological consequences such as alteration of predator–prey dynamics and transmission of disease to native deer (Odocoileus hemionus).     

Diet of wolves <Emphasis Type="Italic">Canis lupus</Emphasis> returning to Hungary

At the end of the nineteenth century, the wolf Canis lupus was extinct in Hungary and in recent decades has returned to the northern highland area of the country. The diet of wolves living in groups in Aggteleki National Park was investigated using scat analysis (n = 81 scats) and prey remains (n = 31 carcasses). Throughout the year wolves (average, minimum two wolves per year) consumed mostly wild-living ungulates (mean percent of biomass consumed, B% 97.2%; relative frequency of occurrence, %O 74.0%). The wild boar Sus scrofa was the most common prey item found in wolf scat (%B 35.6%) and is also the most commonly occurring ungulate in the study areas. The second most commonly occurring prey item in wolf scat was red deer Cervus elaphus (B% 32.8%). Conversely, prey remain analyses revealed wild boar as the second most commonly utilised prey species (%O 16.1%) after red deer (%O 67.7%). The roe deer Capreolus capreolus that occurs at lower population densities was the third most commonly utilised prey species. The importance of low population density mouflon Ovis aries, livestock and other food types was low. The results are similar to those found in the northern part of the Carpathian Mountains.     

Assessment of anadromous salmon resources in the diet of the Alexander Archipelago wolf using stable isotope analysis

The Alexander Archipelago wolf (Canis lupus ligoni) is unique to southeast Alaska, occurring on islands south of Frederick Sound and along the mainland between Dixon Entrance and Yakutat Bay. Sitka black-tailed deer (Odocoileus hemionus sitkensis) are an important prey species for wolves across the southern part of the region. Spawning salmon (Onchorynchus sp.) are seasonally available but their presence in wolf diets has not previously been quantified. We examined the range of bone collagen δ¹³C and δ¹⁵N values for wolves throughout southeast (n = 163) and interior (n = 50) Alaska and used a dual-isotope mixing model to determine the relative contribution of salmon-derived marine protein in the diet. Southeast Alaska wolves consumed significantly more salmon (mean ± SE: 18.3 ± 1.2%) than did wolves from interior Alaska (9.1 ± 0.6%, P<0.001). Wolves on the southeast Alaska mainland appeared to have higher marine isotopic signatures than island wolves, although this difference was not significant. Variation among individual wolf diets was higher for southeast than for interior Alaska wolves, and variation was highest in coastal mainland wolf diets (P<0.001). Marine resources may augment the diet of southeast Alaska wolves during seasonal or annual fluctuations in the availability of deer, particularly in those areas on the mainland where densities of terrestrial ungulates are relatively low. Received: 1 July 1998 / Accepted: 23 February 1999     

Wolves,white‐tailed deer,and beaver: implications of seasonal prey switching for woodland caribou declines

Population increases of primary prey can negatively impact alternate prey populations via demographic and behavioural responses of a shared predator through apparent competition. Seasonal variation in prey selection patterns by predators also can affect secondary and incidental prey by reducing spatial separation. Global warming and landscape changes in Alberta's bitumen sands have resulted in prey enrichment, which is changing the large mammal predator–prey system and causing declines in woodland caribou Rangifer tarandus caribou populations. We assessed seasonal patterns of prey use and spatial selection by wolves Canis lupus in two woodland caribou ranges in northeastern Alberta, Canada, that have undergone prey enrichment following recent white‐tailed deer Odocoileus virginianus invasion. We determined whether risk of predation for caribou (incidental prey) and the proportion of wolf‐caused‐caribou mortalities varied with season. We found that wolves showed seasonal variation in primary prey use, with deer and beaver Castor canadensis being the most common prey items in wolf diet in winter and summer, respectively. These seasonal dietary patterns were reflected in seasonal wolf spatial resource selection and resulted in contrasting spatial relationships between wolves and caribou. During winter, wolf selection for areas used by deer maintained strong spatial separation between wolves and caribou, whereas wolf selection for areas used by beaver in summer increased the overlap with caribou. Changing patterns in wolf resource selection were reflected by caribou mortality patterns, with 76.2% of 42 adult female caribou mortalities occurring in summer. Understanding seasonal patterns of predation following prey enrichment in a multiprey system is essential when assessing the effect of predation on an incidental prey species. Our results support the conclusion that wolves are proximately responsible for woodland caribou population declines throughout much of their range.     

Population size and time since island isolation determine genetic diversity loss in insular frog populations

Understanding the factors that contribute to loss of genetic diversity in fragmented populations is crucial for conservation measurements. Land‐bridge archipelagoes offer ideal model systems for identifying the long‐term effects of these factors on genetic variations in wild populations. In this study, we used nine microsatellite markers to quantify genetic diversity and differentiation of 810 pond frogs (Pelophylax nigromaculatus) from 24 islands of the Zhoushan Archipelago and three sites on nearby mainland China and estimated the effects of the island area, population size, time since island isolation, distance to the mainland and distance to the nearest larger island on reduced genetic diversity of insular populations. The mainland populations displayed higher genetic diversity than insular populations. Genetic differentiations and no obvious gene flow were detected among the frog populations on the islands. Hierarchical partitioning analysis showed that only time since island isolation (square‐root‐transformed) and population size (log‐transformed) significantly contributed to insular genetic diversity. These results suggest that decreased genetic diversity and genetic differentiations among insular populations may have been caused by random genetic drift following isolation by rising sea levels during the Holocene. The results provide strong evidence for a relationship between retained genetic diversity and population size and time since island isolation for pond frogs on the islands, consistent with the prediction of the neutral theory for finite populations. Our study highlights the importance of the size and estimated isolation time of populations in understanding the mechanisms of genetic diversity loss and differentiation in fragmented wild populations.     

Utility of island populations in re‐introduction programmes – relationships between Arabian gazelles (Gazella arabica) from the Farasan Archipelago and endangered mainland populations

Understanding local adaptation and population differentiation is vital to the success of re‐introduction initiatives. As other mammals living on islands, Arabian gazelles (G. arabica) show reduced body size on the Farasan archipelago, which we corroborated in this study through morphometric analyses of skulls. In the light of the steep population decline on the Arabian Peninsula – but stable population development on the archipelago – we tested the potential suitability of Farasan gazelles as a source for re‐introductions on the mainland. We therefore investigated genetic differentiation between Farasan and mainland populations using eleven nuclear microsatellite loci and detected a distinct genetic cluster exclusively present on the archipelago, which we inferred to be separated from the mainland cluster for less than 2000 years. About 30% of sampled individuals from Farasan Islands showed assignment to a mainland cluster with signs of ongoing introgression. Analyses using the isolation‐with‐migration model confirmed recent (probably human‐induced) bidirectional exchange of gazelles between mainland and island populations. Hence, the surprisingly uniform island dwarfism most likely reflects phenotypic plasticity, that is, altered morphology as a direct consequence of harsh environmental conditions and resource limitation on the archipelago. Should a further decline of Arabian gazelles on the mainland necessitate restocking in the future, Farasan gazelles may thus become an additional source for captive breeding programmes.     

Nested distribution of amphibians in the Zhoushan archipelago, China: can selective extinction cause nested subsets of species?

Because of their poor dispersal ability, amphibians are well suited for testing the selective extinction theory on islands. Amphibian fauna in the Zhoushan archipelago, China, exhibit a high level of nestedness (C = 0.893), and the species number is lower on islands than on similar sized areas on the mainland. No correlation was found between island-specific species richness and the nearest distance from a larger island, distance from the mainland or density of human population. These results suggest that no amphibian colonisation has occurred in the archipelago since island isolation 7000–9000 years ago. Furthermore, the results imply that selective extinction contributes to the nestedness of amphibians in the Zhoushan archipelago. The incidence of a species on the islands is significantly correlated with log area of the smallest island occupied by the species and the number of provinces on the Chinese mainland in which the species occur. However, there is no correlation with average body length of adults and island occurrence. It is concluded that (1) the area of the smallest island occupied by a species is a good estimate of the minimum area for a viable population of the species and a good predictor of species incidence on islands, (2) species with a restricted distribution range are more vulnerable to extinction from islands than those with a wide distribution range and (3) the effect of body size on occurrence on the islands is uncertain, and may be specific to the archipelago and taxa studied. The observed nestedness of amphibian assemblages has two implications for conservation: (1) not only can all the species found in several small reserves be found on a large reserve of the same total size, but additional species can be found on the single large reserve; (2) for a reserve to maintain viable populations of all species in a region it should be at least as large as the smallest island occupied by the most vulnerable species. Received: 16 December 1996 / Accepted: 22 September 1997     

Recolonizing gray wolves increase parasite infection risk in their prey

The recent recolonization of Central Europe by the European gray wolf (Canis lupus) provides an opportunity to study the dynamics of parasite transmission for cases when a definitive host returns after a phase of local extinction. We investigated whether a newly established wolf population increased the prevalence of those parasites in ungulate intermediate hosts representing wolf prey, whether some parasite species are particularly well adapted to wolves, and the potential basis for such adaptations. We recorded Sarcocystis species richness in wolves and Sarcocystis prevalence in ungulates harvested in study sites with and without permanent wolf presence in Germany using microscopy and DNA metabarcoding. Sarcocystis prevalence in red deer (Cervus elaphus) was significantly higher in wolf areas (79.7%) than in control areas (26.3%) but not in roe deer (Capreolus capreolus) (97.2% vs. 90.4%) or wild boar (Sus scrofa) (82.8% vs. 64.9%). Of 11 Sarcocystis species, Sarcocystis taeniata and Sarcocystis grueneri occurred more often in wolves than expected from the Sarcocystis infection patterns of ungulate prey. Both Sarcocystis species showed a higher increase in prevalence in ungulates in wolf areas than other Sarcocystis species, suggesting that they are particularly well adapted to wolves, and are examples of “wolf specialists”. Sarcocystis species richness in wolves was significantly higher in pups than in adults. “Wolf specialists” persisted during wolf maturation. The results of this study demonstrate that (1) predator–prey interactions influence parasite prevalence, if both predator and prey are part of the parasite life cycle, (2) mesopredators do not necessarily replace the apex predator in parasite transmission dynamics for particular parasites of which the apex predator is the definitive host, even if meso‐ and apex predators were from the same taxonomic family (here: Canidae, e.g., red foxes Vulpes vulpes), and (3) age‐dependent immune maturation contributes to the control of protozoan infection in wolves.     

Diet of raccoon dogs in summer in the Finnish archipelago

The diet of raccoon dogs in summer was examined in the Finnish archipelago and compared with their diet on the mainland. Faeces were collected from latrines in the outer archipelago and the inhabited islands in May-July 1998 and 1999. Remains of waterfowl, especially those of female cider, occurred frequently in the faeces of raccoon dogs on the uninhabited islands, but less frequently on the inhabited islands and only occasionally on the mainland. Since many female eider died of disease during the summer in 1998 and 1999. and white-tailed eagles arc known to prey on brooding eider, it is probable that raccoon dogs did not themselves kill all the eider they consumed. Even if they did, they would only take a few percent of female eider population each year. It is thus improbable that raccoon dogs affect the size of the eider population in the Finnish archipelago. Frogs, reptiles, shrews and carrion occurred frequently in the diet on the mainland, but only occasionally in the outer archipelago. It is possible that raccoon dogs have decimated frog populations on small islands, as they are easy prey and occur frequently in the diet of raccoon dogs on the mainland. Small rodents are important prey for raccoon dogs on the mainland and on inhabited islands, but less important in the outer archipelago in early summer.     

Fragmented environment affects birch leaf endophytes

The effect of environmental fragmentation on the species distribution and frequency of horizontally transmitted endophytic fungi in Betula pubescens and Betula pendula leaves was studied in an archipelago in southwestern Finland. The study system consisted of 14 islands, ranging in size and distance to the mainland, and five mainland sites. Endophytic fungi were grown out from surface-sterilized leaves. The frequency of endophytic fungi mainly depended on the size of the island, explaining 32-35% of the variation, and the distance to the mainland explaining 29-35% of the variation. The birch trees on the largest islands near the mainland had the highest endophyte frequencies. Fusicladium betulae, Gnomonia setacea and Melanconium betulinum were the most commonly isolated fungi. Foliar endophytes of birch trees are able to disperse to fairly fragmented areas, but their frequencies seem to depend on environmental isolation and size of the island.     

Maximum body size among insular Komodo dragon populations covaries with large prey density

This study documents variation in maximum body size of Komodo dragons ( Varanus komodoensis ) among the four extant island populations in Komodo National Park and compares an indirect measure of deer density, the major prey item for large dragons, to differences in maximum body size among islands. The largest 15% of dragons from the large islands of Komodo and Rinca were significantly longer and heavier than the largest 15% of dragons on the small islands of Gili Motang and Nusa Kode. There was a 33% difference in snout vent length (SVL) between dragons found on Komodo and those found on Gili Motang, with mass varying by more than four-fold. Density of deer pellet groups between islands ranged from 5.86±0.75 groups per transect on Gili Motang to 20.73±1.02 groups per transect on Komodo Island. Maximal dragon SVL and mass was highly positively correlated with this index of deer density. Low prey density on the two small islands could constrain body size via energetic constraints. At present we can not deduce if insular body size variation has arisen through genotypic or phenotypic mechanisms.     

Optimal predator management for mountain sheep conservation depends on the strength of mesopredator release

Large predators often suppress ungulate population growth, but they may also suppress the abundance of smaller predators that prey on neonatal ungulates. Antagonistic interactions among predators may therefore need to be integrated into predator–prey models to effectively manage ungulate–predator systems. We present a modeling framework that examines the net impact of interacting predators on the population growth rate of shared prey, using interactions among wolves Canis lupus, coyotes Canis latrans and Dall sheep Ovis dalli dalli as a case study. Wolf control is currently employed on approximately 16 million ha in Alaska to increase the abundance of ungulates for human harvest. We hypothesized that the positive effects of wolf control on Dall sheep population growth could be counteracted by increased levels of predation by coyotes. Coyotes and Dall sheep adult females (ewes) and lambs were radiocollared in the Alaska Range from 1999–2005 to estimate fecundity, age‐specific survival rates, and causes of mortality in an area without wolf control. We used stage‐structured population models to simulate the net effect of wolf control on Dall sheep population growth (λ). Our models accounted for stage‐specific predation rates by wolves and coyotes, compensatory mortality, and the potential release of coyote populations due to wolf control. Wolves were the main predators of ewes, coyotes were the main predators of lambs, and wolves were the main source of mortality for coyotes. Population models predicted that wolf control could increase sheep λ by 4% per year in the absence of mesopredator release. However, if wolf control released coyote populations, our models predicted that sheep λ could decrease by up to 3% per year. These results highlight the importance of integrating antagonistic interactions among predators into predator–prey models, because the net effect of predator management on shared prey can depend critically on the strength of mesopredator release.     

Evolutionary history of Scinax treefrogs on land‐bridge islands in south‐eastern Brazil

Aim We investigated how Pleistocene refugia and recent (c. 12,000 years ago) sea level incursions shaped genetic differentiation in mainland and island populations of the Scinax perpusillus treefrog group. Location Brazilian Atlantic Forest, São Paulo state, south‐eastern Brazil. Methods Using mitochondrial and microsatellite loci, we examined population structure and genetic diversity in three species from the S. perpusillus group, sampled from three land‐bridge islands and five mainland populations, in order to understand the roles of Pleistocene forest fragmentation and sea level incursions on genetic differentiation. We calculated metrics of relatedness and genetic diversity to assess whether island populations exhibit signatures of genetic drift and isolation. Two of the three island populations in this study have previously been described as new species based on a combination of distinct morphological and behavioural characters, thus we used the molecular datasets to determine whether phenotypic change is consistent with genetic differentiation. Results Our analyses recovered three distinct lineages or demes composed of northern mainland São Paulo populations, southern mainland São Paulo populations, and one divergent island population. The two remaining island populations clustered with samples from adjacent mainland populations. Estimates of allelic richness were significantly lower, and estimates of relatedness were significantly higher, in island populations relative to their mainland counterparts. Main conclusions Fine‐scale genetic structure across mainland populations indicates the possible existence of local refugia within São Paulo state, underscoring the small geographic scale at which populations diverge in this species‐rich region of the Atlantic Coastal Forest. Variation in genetic signatures across the three islands indicates that the populations experienced different demographic processes after marine incursions fragmented the distribution of the S. perpusillus group. Genetic signatures of inbreeding and drift in some island populations indicate that small population sizes, coupled with strong ecological selection, may be important evolutionary forces driving speciation on land‐bridge islands.     

Prey Responses to Predator's Sounds: A Review and Empirical Study

Many animals assess their risk of predation by listening to and evaluating predators' vocalizations. We reviewed the literature to draw generalizations about predator discrimination abilities, the retention of these abilities over evolutionary time, and the potential underlying proximate mechanisms responsible for discrimination. Broadly, we found that some prey possess an ability to respond to a predator after having been evolutionarily isolated from a specific predator (i.e., predators are allopatric) and that some prey are predisposed to respond to certain types of predators that they coevolved with but without having ecological experience. However, these types of studies are lacking, and relatively, few studies have examined predator discrimination abilities in ungulates. To begin addressing these knowledge gaps, we performed field experiments on Mule deer (Odocoileus hemionus) in which we investigated the ability of deer to discriminate among familiar predators [coyotes (Canis latrans) and mountain lions (Puma concolor)] and an evolutionary relevant predator with which deer have had no recent exposure [locally extinct wolves (Canis lupus)]. We found that Mule deer respond to and discriminate among predators based on predator vocalizations and have retained an ability to respond to wolves that have been extinct from the study area since the early 20th century. Previous playback studies have shown that responses vary among human‐habituated and non‐habituated populations and differ according to human proximity. Deer greater than 0.5 km from human residences allocated more time to heightened responses both before and after stimulus playback. Our findings may help predict how prey–predator interactions may change as a result of the recovering wolf population with a basis in ecological and evolutionary experience in predator discrimination and desensitization.     

Island biogeography of bats in Baja California, Mexico: patterns of bat species richness in a near-shore archipelago

Aim We studied the relationship between the size and isolation of islands and bat species richness in a near‐shore archipelago to determine whether communities of vagile mammals conform to predictions of island biogeography theory. We compared patterns of species richness in two subarchipelagos to determine whether area per se or differences in habitat diversity explain variations in bat species richness. Location Islands in the Gulf of California and adjacent coastal habitats on the Baja California peninsula in northwest Mexico. Methods Presence–absence surveys for bats were conducted on 32 islands in the Gulf of California using acoustic and mist‐net surveys. We sampled for bats in coastal habitats of four regions of the Baja peninsula to characterize the source pool of potential colonizing species. We fitted a semi‐log model of species richness and multiple linear regression and used Akaike information criterion model selection to assess the possible influence of log₁₀ area, isolation, and island group (two subarchipelagos) on the species richness of bats. We compared the species richness of bats on islands with greater vegetation densities in the southern gulf (n = 20) with that on drier islands with less vegetation in the northern gulf (n = 12) to investigate the relationship between habitat diversity and the species richness of bats. Results Twelve species of bats were detected on islands in the Gulf of California, and 15 species were detected in coastal habitats on the Baja peninsula. Bat species richness was related to both area and isolation of islands, and was higher in the southern subarchipelago, which has denser vegetation. Log₁₀ area was positively related to bat species richness, which increased by one species for every 5.4‐fold increase in island area. On average, richness declined by one species per 6.25 km increase in isolation from the Baja peninsula. Main conclusions Our results demonstrate that patterns of bat species richness in a near‐shore archipelago are consistent with patterns predicted by the equilibrium theory of island biogeography. Despite their vagility, bats may be more sensitive to moderate levels of isolation than previously expected in near‐shore archipelagos. Differences in vegetation and habitat xericity appear to be associated with richness of bat communities in this desert ecosystem. Although observed patterns of species richness were consistent with those predicted by the equilibrium theory, similar relationships between species richness and size and isolation of islands may arise from patch‐use decision making by individuals (optimal foraging strategies).     

Colonization of Sphagnum on land uplift islands in the Baltic Sea: time, area, distance and life history

Aim To test whether species richness of Sphagnum mosses on islands in a land uplift archipelago is related to island age, area or connectivity, and whether the frequency of different species can be predicted by their life history and autecology. Location The northern Stockholm archipelago in the Baltic Sea, east‐central Sweden, with a current land uplift rate of 4.4 mm year^?1. Methods We sampled 17 islands differing in area (0.55–55 ha), height (3.6–18 m, representing c. 800–4000 years of age) and distance from mainland (1.6–41 km). For each Sphagnum patch we measured area, height above sea level, horizontal distance from the shore and shading from vascular plants. Factors affecting island species richness, species frequency and habitats on the islands were tested by stepwise regressions. Species frequency was tested on nine life history and autecological variables, including estimated abundance and spore output on the mainland, habitat preference and distribution. Results We recorded 500 patches of 19 Sphagnum species, distributed in 83 rock pools on 14 islands. Island species richness correlated positively with island area and with degree of shelter by surrounding islands, while distance from the mainland, connectivity, height or age did not add to the model. Species frequency (number of colonized islands and rock pools) was mainly predicted by spore output on the mainland and by habitat preference (swamp forest species were more frequent than others), while spore size, for example, did not add to the model. Species differed in mean height above and horizontal distance from the shore, area of occupied rock pools and in the degree of shading of patches. The mean horizontal distance from the shore and the area of occupied rock pools correlated positively with the normal growth position above the water table among species. Spore capsules were found in only 2% of patches, mostly in the bisexual Sphagnum fimbriatum. Main conclusions The presence of Sphagnum in the Stockholm archipelago seems to be governed by regional spore production and habitat demands. Sphagnum does not appear to be dispersal limited at distances up to 40 km and time spans of centuries. Species with a high regional spore output have had a higher colonization rate, which, together with the rarity of spore capsules on the islands, indicate the mainland as a source for colonization rather than dispersal among islands. Swamp forest species seem more tolerant to the island conditions (summer droughts and some salt spray) than open mire species. The different distances from the sea occupied by the species indicate a slow, continuous succession and species replacement towards the island interior as islands are being uplifted and thus expand in area. This partly explains why larger islands harbour more species. Our results thus support some of the island biogeographical theories related to the species–area relationship.     

Are species–area relationships from entire archipelagos congruent with those of their constituent islands?

Aim To establish the extent to which archipelagos follow the same species–area relationship as their constituent islands and to explore the factors that may explain departures from the relationship. Location Thirty‐eight archipelagos distributed worldwide. Methods We used ninety‐seven published datasets to create island species–area relationships (ISARs) using the Arrhenius logarithmic form of the power model. Observed and predicted species richness of an archipelago and of each of its islands were used to calculate two indices that determined whether the archipelago followed the ISAR. Archipelagic residuals (ArcRes) were calculated as the residual of the prediction provided by the ISAR using the total area of the archipelago, standardized by the total richness observed in the archipelago. We also tested whether any characteristic of the archipelago (geological origin and isolation) and/or taxon accounts for whether an archipelago fits into the ISAR or not. Finally, we explored the relationship between ArcRes and two metrics of nestedness. Results The archipelago was close to the ISAR of its constituent islands in most of the cases analysed. Exceptions arose for archipelagos where (i) the slopes of the ISAR are low, (ii) observed species richness is higher than expected by the ISAR and/or (iii) distance to the mainland is small. The archipelago's geological origin was also important; a higher percentage of oceanic archipelagos fit into their ISAR than continental ones. ArcRes indicated that the ISAR underpredicts archipelagic richness in the least isolated archipelagos. Different types of taxon showed no differences in ArcRes. Nestedness and ArcRes appear to be related, although the form of the relationship varies between metrics. Main conclusions Archipelagos, as a rule, follow the same ISAR as their constituent islands. Therefore, they can be used as distinct units themselves in large‐scale biogeographical and macroecological studies. Departure from the ISAR can be used as a crude indicator of richness‐ordered nestedness, responsive to factors such as isolation, environmental heterogeneity, number and age of islands.     

Biogeographical modules and island roles: a comparison of Wallacea and the West Indies

Aim In order to advance our understanding of the assembly of communities on islands and to elucidate the function of different islands in creating regional and subregional distribution patterns, we identify island biogeographical roles on the basis of the distribution of the islands’ biota within the archipelago. We explore which island characteristics determine island biogeographical roles. Furthermore, we identify biogeographical subregions, termed modules. Location Wallacea in Indonesia, and the West Indies in the Caribbean Sea. Methods We use a network approach to detect island biogeographical roles and avian biogeographical modules. To designate the biogeographical role of an island, each island is assigned two coordinates, l and r. The position of an island in l–r space characterizes its role, namely as peripheral, connector, module hub, or network hub. Island characteristics are tested as predictors of l and r. Results Both Wallacea and the West Indies were found to be significantly modular and divided into four biogeographical modules. The four modules identified within Wallacea each contain all existing island roles, whereas no module in the West Indies represents all possible roles. Island area and elevation appeared to be the most important determinants of an island’s l score, while measurements of isolation essentially determined the r score. Main conclusions In both Wallacea and the West Indies, the geographic structuring into biogeographical modules corresponds well with our knowledge of past connections and contemporary factors. In both archipelagos, large, mountainous islands are identified as hubs and are thus responsible for faunal coherence within modules (module hubs) and across the entire archipelago (network hubs). We thus interpret these as source islands for the surrounding islands in their module (module hubs) or for the entire archipelago (network hubs). Islands positioned marginally in their module and distant from the mainland are identified as connectors or network hubs, behaving as sinks and stepping stones for dispersing species. Modularity and predictors of biogeographical roles are similar for Wallacea and the West Indies, whereas the build‐up of biogeographical modules and the assortment of roles depend on the spatial constellation of islands in each archipelago.