Determinants of bird community composition on patches in the suburbs of Paris, France

The ecological processes responsible for the spatial assemblages of breeding bird communities in urban landscapes are more and more investigated. Indeed, understanding these processes is imperative to plan relevant management policies. We investigated breeding bird communities on 67 patches in the suburbs of Paris, France. We examined the role of patch characteristics and geographic distance between patches in determining similarity between bird assemblages. To do this, we proposed a new SØrensen similarity index based on estimators of change in community composition taking into account the detection probability of species. The patch occupancy by sedentary and migratory species was also estimated to compare their sensitivity to urbanization. Patches close to each other supported more similar bird assemblages, suggesting an effect of the spatial distribution of patches on bird dispersal and a posteriori on local community composition. Accounting for spatial location of patches, bird assemblage similarity was related to the similarity of the surrounding level of urbanization but not to the similarity of patch size or to the similarity of patch vegetation. The mean estimated occupancy rate of sedentary species was higher than that of migratory species in the whole study area. While sedentary species occupied patches surrounded by both moderate and high levels of urbanization, migratory ones primarily occupied patches surrounded by moderate levels of urbanization. Human choices in degrees, styles, and extent of urbanization, including designation and design of patches within an urban matrix, affect the composition of local bird communities.     

Long-term effects of fragmentation and fragment properties on bird species richness in Hawaiian forests

Forest fragmentation is a common disturbance affecting biological diversity, yet the impacts of fragmentation on many forest processes remain poorly understood. Forest restoration is likely to be more successful when it proceeds with an understanding of how native and exotic vertebrates utilize forest patches of different size. We used a system of forest fragments isolated by volcanic activity 153 years ago in Hawaii to examine how long-term fragmentation, as well as fragment size and structural features affect the richness of native and exotic bird species. The total number of bird species increased rapidly with forest fragment size, with most of the native species pool found in patches <3 ha. Smaller fragments were dominated by native bird species with several exotic bird species found only in the largest fragments, suggesting that exotic bird species in this landscape show greater area-sensitivity than native species. We used airborne scanning light detection and ranging (LiDAR) to assess whether fragment area was correlated with estimates of fragment vegetation volume as well as measures of tree height. Fragment area was highly correlated with vegetation volume, maximum tree height, and canopy height heterogeneity, and these variables were strong predictors of bird richness, demonstrating that remote sensing can provide key insights into the relationship between fragment structural attributes and biodiversity indicators. Overall, this work demonstrates the value of conserving small remnant mid-elevation forest patches for native birds in Hawaii. This work also provides insight into how newly created forest patches might be used by native and exotic bird species in Hawaii.     

The influence of patch area and connectivity on avian communities in urban revegetation

Landscape restoration through revegetation is being increasingly used in the conservation management of degraded landscapes. To effectively plan restoration programs information is required on how the landscape context of revegetation influences biodiversity gains. Here, we investigate the relative influence of patch area and connectivity on bird species richness and abundance within urban revegetation patches in Brisbane, Australia. We carried out bird surveys at 20 revegetation sites, and used hierarchical partitioning and model selection to test the relative importance of patch area (the area of revegetation including all directly connected remnant vegetation) and landscape connectivity (the vegetated area connected by less than 10 m, 20 m, 30 m, 40 m and 50 m cleared gaps). We controlled for a number of possible confounding variables within the hierarchical partitioning procedure. Both the hierarchical partitioning and model selection procedures indicated that connectivity had an important influence on bird species richness. Patch area in combination with connectivity were important influencing factors on overall bird abundance. We also carried out the hierarchical partitioning procedure for bird abundance data within a range of feeding guilds, yielding results specific to species groups. Overall our data suggest that greater connectivity enhances the habitat area that colonists can arrive from (resulting in greater species richness), whereas increased patch area allows for increased abundance by expanding the habitat available to species already present in a patch. A combined approach where connectivity and overall habitat area is enhanced across the landscape is likely to be necessary to meet long-term conservation objectives.     

Matrix effects on bird assemblages in fragmented Afromontane forests in South Africa

We examine the effects of matrix type on forest bird species richness in historically fragmented forests of the KwaZulu-Natal midlands, South Africa. Bird species-area relationships for forest fragments within natural grassland were compared with those surrounded by plantation forestry (Pinus spp.). While fragments in grasslands displayed a species-area effect, no such effect (slope ∼0) was detected for fragments in plantations. The critical fragment size to avoid an island effect on species richness was 302 ha for fragments in grasslands, and contained 51 of the 61 forest bird species encountered. Small forest fragments (<50 ha) within grasslands were less species-rich than those surrounded by plantation, while the reverse was true for forests larger than 50 ha. Bird density was significantly lower in larger forest patches within plantation when compared to those in grassland. However, a slight decline in species density with increasing species richness in a patch (i.e. weak density compensation) characterised only those bird assemblages surrounded by a grassland matrix, suggesting these are approaching species saturation. Commercial plantations may increase the likelihood of colonisation of, and immigration from, small forest fragments by birds, reducing the incidence of area-dependent extinction in small fragments, but may also result in lower species richness in larger fragments. The latter is likely a consequence of the promotion by plantations of a wider distribution of generalist species in, and the loss of some specialised or rare species from, large patches. Management options are to avoid planting plantations near large forest patches and to increase the size of small patches where possible. Where afforestation is unavoidable, placing plantations in the vicinity of small forest patches rather than large forest patches may be preferable.     

Rapid avifaunal collapse along the Amazonian deforestation frontier

The combined effects of rapid habitat loss, fragmentation and disturbance on tropical forest avifaunas have not been examined to date. The southern Amazonian ‘arc of deforestation’ marks the boundary of the most aggressive agricultural frontier in tropical forests worldwide. We sampled 21 disturbed and undisturbed primary forest patches, ranging in size from 1.2 to 14,476 ha, to elucidate the synergistic effects of both forest disturbance and fragmentation on bird community structure, and pinpoint which species were the “winners” and “losers” from this process. A number of forest patch metrics, derived from an independent remote sensing approach, explained much of the resulting presence/absence matrix. The bird community exhibited a highly nested structure, with small patches being most dissimilar from one another. Bird species differed in their response to both forest patch size and forest canopy perforation according to their dependence on closed-canopy primary forest. Forest patch geometry, which clearly modulated the shape of species-area relationships accounted for 83-96% of the variation in species richness, but forest habitat quality resulting from logging and surface-fire disturbance was also a significant predictor of species richness for the most forest-dependent taxa.     

Nestedness of bird assemblages in fragmented Afromontane forest: the effect of plantation forestry in the matrix

Bird species composition (61 species) of Afromontane forest fragments (n = 19) embedded in a natural grassland matrix or a plantation forestry matrix were compared. Forest bird assemblages in a natural grassland matrix displayed a nested non-random species distribution. Species loss was significantly area-dependent and predictable. Assemblages in the plantation-dominated matrix were also nested but were neither area- nor isolation-dependent and these potential processes had equal influence over bird species composition. Forests patches <50 ha in plantation were more species rich than those in a grassland matrix and visa versa for patches >50 ha. The effect of the matrix is revealed here with arboreal plantation cover promoting dispersal across the landscape, favouring abundant generalist bird species in smaller patches than expected in the plantation matrix, and the loss of some specialists and species living at low densities from large (>50 ha) patches. Because of a significant species-area relationship and possible area-dependent extinction, the largest forest patch in the nested series in a grassland matrix protects most bird species (89%). However, in a plantation matrix the absence of a species-area relationship requires the combination of many larger forest patches to protect most bird species. Plantation forests in the matrix significantly alter forest bird assemblage structure and composition. To avoid the effects of insularisation and to maintain the viability of bird communities in Afromontane forests requires conservation of the larger forests in a natural matrix. However, for forests in a plantation matrix, the management of dispersal processes by manipulation of the matrix may be as important to conserving species and communities as is minimising extinctions. Avoiding placing plantations near large forests, increasing the size of small patches, and where afforestation is unavoidable, placing plantations in the vicinity of small rather than large forest patches, are preferable management practices.     

Values of village fengshui forest patches in biodiversity conservation in the Pearl River Delta, China

Chinese village fengshui forests (VFF) are small remnant forest patches that coexist with natural villages. The indigenous residents protect the forest patches under traditional Chinese geomancy beliefs (namely fengshui). However, the VFF community features and conservation values and relationships with the indigenous people remain poorly understood. In this study, we evaluated tree species diversity conservation of regional VFF patches by sampling a 1200 m² transect within each patch. We also tested our hypothesis that patch size did not significantly impact interior forest community features of well-protected VFF patches. Thirty-two well-protected VFF patches in the Pearl River Delta, China were investigated. The average coefficient of similarity between transects (CS) was employed to evaluate community heterogeneity. Five forest community parameters (tree species richness per 1200 m², tree stem density, tree basal area density, Shannon–Wiener diversity index (SWI), and Simpson diversity index (SI)) were measured and compared with regional well-developed evergreen broadleaved, coniferous and coniferous-broadleaved mixed forests. The relationships between the five parameters versus patch size and elevation were analyzed. A total of 266 tree species comprised of 57 families were recorded in 32 transects, of which 258 (97%) species were indigenous and eight (3%) were exotic. Ten tree species were endangered, rare or nationally protected by the Chinese government, and 57 species were endemic to China. The average CS was 0.38; and the average five forest community parameter values were as follows: 46.8 for tree species richness; 3403 plants/ha for stem density, 49.1 m²/ha for basal area, 4.04 for SWI and 0.90 for SI. These values were consistent with well-developed evergreen broadleaved forests and greater than coniferous and coniferous-broadleaved forests. No significant correlation was detected between the five community parameters and patch area or elevation. We conclude that VFF patches preserve abundant tree species and heterogeneous habitats, which are important for maintaining regional biodiversity. The interior community features of VFF patches were not significantly affected by patch size. We recommend protection of both large and small VFF patches, which can be substantially enhanced by the preservation of associated traditional relic village cultures.     

Composition and abundance of birds in Andean alder (Alnus acuminata) patches with past and present harvest in Bolivia

We compared species richness and abundance of birds between five patches under selective Alnus exploitation and five patches that have not been harvested for at least 10 years prior to our study, during the early dry season (April-July 2001), in Cotapata National Park. Using “point counts” we recorded birds and their distribution in two (<1.5 m and >1.5 m) forest layers. Simultaneously we evaluated the floristic structure (size [dbh] distribution, basal area, tree density, tree height, and vegetation cover) and composition (diversity) on three transects placed within each Alnus patch. Both bird diversity and vegetation cover were significantly higher in not presently used patches but only for the higher layer of the forest, whereas plant diversity was higher in presently used patches. Lack of differences between the two types of Alnus patches in any of the vegetation parameters measured in the lower layer was coupled with an indistinct avifauna. Small changes in habitat characteristics following a perturbation like selective logging have the potential to affect richness and abundance of birds, at least within the habitats directly affected by the perturbation.     

Small isolated aspen stands enrich bird communities in southwestern ponderosa pine forests

Small aspen stands are disappearing from the landscape in the Southwest, so it is important to understand their contribution to the avian community. We sampled birds in 53 small, isolated aspen stands and 53 paired plots within the ponderosa pine forest in northern Arizona, during the 1996 and 1997 breeding seasons. Bird species richness and abundance were higher in aspen than in pine. However, bird species richness and abundance did not vary with size of the aspen patch or isolation index. In addition, direct ordination of species distributions with habitat factors suggested no distinct avian communities. This suggests that aspen stands do not harbor separate populations, but rather are locations where the regional avifauna reaches high local density and richness and may be crucial to birds in years of resource scarcity. Thus it is important for avian conservation to maintain many aspen stands across the landscape, encompassing a diversity of vegetation structure and composition.     

Birds as agents of seed dispersal in a human-dominated landscape in southern Costa Rica

Birds play vital roles as seed dispersers helping to maintain and restore plant communities. With restoration increasingly key to global conservation, it is important to understand the landscape attributes and bird community characteristics that most influence avian seed dispersal in human-altered landscapes. We examined bird community structure and seed-dispersal patterns in agricultural countryside in Costa Rica that is typical of much of the Neotropics. Contrary to expectations, bird abundance, not richness, best predicted the richness of bird-dispersed seeds. Neither forest patch size or proximity, nor total tree cover, influenced seed dispersal. The richness and abundance of dispersed seeds, however, was strongly correlated with “wetness,” a remotely-sensed metric of vegetation, at several scales. These results suggest that in this human-dominated tropical region: (1) bird abundance, not species richness or size, may drive seed dispersal, and (2) remote-sensing combined with field verification can detect landscape elements that are helpful for maintaining the option of bird-mediated reforestation.     

Colonisation of newly established habitats by soil decomposer organisms: the effect of habitat corridors in relation to colonisation distance and habitat size

The aim of the present 2.5-year-long field experiment was to explore the ability of various members of the detrital food web to colonise newly established habitat patches in field conditions, either in the presence or absence of habitat corridors. Patch size and distance to the “mainland” (colonisation source) were manipulated to explore the scale dependency of the corridor effects. Sterilised humus patches, embedded in mineral soil regarded as uninhabitable (or non-preferred) matrix for the soil organisms, functioned as newly established habitats. Intact forest soil served as the source of colonisers. Three kinds of patches were established: large ones situated at relatively long distance from the colonisation source, and small ones situated either at long or short distance from the source. Corridors, when present, connected the patches to the intact forest soil. The presence of corridors consistently increased the species richness of soil fungi in the patches. Also bacterial species richness, number of microarthropod taxa and abundance of enchytraeid worms were sometimes increased by the presence of corridors. The response of bacterial species was assumed to be indirect, caused by the response of enchytraeid worms functioning as dispersal agents for the bacteria. Colonisation of the patches by soil organisms was virtually independent of patch size and colonisation distance, both in the presence and absence of corridors. This can be related to the resource-poor or otherwise sub-optimal conditions in the initially sterile humus soil, restricting the number of species able to establish themselves in the patches and/or utilise the corridors. Nevertheless, the results indicate that habitat corridors can facilitate the colonisation of unpopulated habitats by organisms with greatly different life history traits.     

Bird responses to broad-leaved forest patch area in a plantation landscape across seasons

Although plantation forests have widely replaced native forests worldwide, few studies have examined the effects of plantation forestry on organisms at the landscape level. In this study, we examined the effects of broad-leaved forest patch area (1.4–312 ha) on bird assemblages in a conifer plantation-dominated landscape during wintering and breeding seasons. We also surveyed birds in the plantation matrix and treated the patch area of the matrix as 0. We examined whether the detection rates of species changed suddenly across patch area (threshold) using segmented regression. We found that species richness increased with patch area in both seasons. Responses of detection rates to patch area varied and were categorized as follows: insensitive, linear increase and decrease, quadratic increase and decrease, concave and convex. Thresholds indicating that the detection rate suddenly increased above a certain patch area were found for two species only in the winter. Species responses varied more in the winter than in the breeding season. Certain ecological traits of species were consistently associated with the responses to patch area across seasons (migratory strategy, nesting substrate and foraging trait), while body weight was only associated in the winter. Because species richness and detection rates of many species were high in large patches, large broad-leaved forests (especially >40 ha) should be retained and restored with high priority. Additionally, bird responses to patch area in the winter should not be ignored because many species required large patches during the winter season.     

The effects of forest patch size and matrix type on changes in carabid beetle assemblages in an urbanized landscape

In this study we compared ground beetles (Carabidae) from a range of different forest fragments along an urbanization gradient in Brussels, Belgium. We address the following questions: (i) How does the degree of urbanization in the surrounding habitat affect forest beetles, and does it interact with the effects of patch size and distance to forest edge? (ii) Do these factors have a different effect at the level of individual species, habitat affinity groups or total community? During 2002 we sampled 13 forest plots in 10 forest patches, ranging in size from 5.27 to 4383 ha. The beetles were captured using transects of pitfall traps from the edge to a distance of 100 m into each woodland and identified to species level. Effects of urbanization, forest size and forest edge were evaluated on total species number, abundance and habitat affinity groups and ten abundant, widespread model carabid species. Overall, the effects of urbanization, forest size and edge effects slightly influenced total species richness and abundance but appeared to have a major effect on ground beetle assemblages through species specific responses. More urbanized sites had significantly fewer forest specialists and more generalist species. Large forest fragments were favoured by forest specialist species while generalist species and species frequently associated with forest (forest generalists) dominated the smaller forests. Forest edges mainly harboured generalist species while forest specialist species were more frequent into the forests if the forest patches were large enough, otherwise they disappeared due to the destruction or impoverishment of their habitat. Our results show the importance of differentiating between habitat affinity, especially habitat generalists versus specialists, the latter having a higher value in nature conservation, and merely the quantity of species represented in human-dominated areas.     

Effects of forestry practices on vegetation structure and bird community of Kibale National Park, Uganda

I examined the lingering effects of past timber management practices on the vegetation structure and bird community of Kibale National Park, Uganda. I compared four forest treatments: unlogged native forest (UL), two that were selectively logged at low (LL) and high (HL) intensities in the 1960s, and a conifer plantation (PL). Forest-dependent birds were best represented at UL. LL was similar to UL in both vegetation structure and bird community composition, although some forest-dependent bird species were missing from the former. HL had significantly less canopy closure and lower tree density than other plots as a result of the combination of extensive secondary damage and natural disturbance patterns that prevented the reclosure of the forest canopy. Thirty-one percent of the forest-dependent bird species observed during the study were not detected at HL. At PL, bird species richness and bird abundance were about a third of those observed in other plots. There were significant correlations between heterogeneity of tree distribution (horizontal heterogeneity) and abundance and species richness of birds across plots. Abundance and species richness of all, forest-dependent, and forest generalist birds were highest in plots with intermediate measures of horizontal heterogeneity, which were mostly unlogged or lightly logged. If reduced-impact logging practices are not implemented during selective logging operations in tropical forests, consequent long-term changes in vegetation structure may result in significant declines in the populations of some forest-dependent species, as was observed in Kibale National Park.     

Anthropogenic determinants of primate and carnivore local extinctions in a fragmented forest landscape of southern Amazonia

Habitat fragmentation has been shown to influence the abundance, movements and persistence of many species. Here, we examine the effects of forest patch and landscape metrics, and levels of forest disturbance on the patterns of local extinction of five primate and 14 carnivore species within 129 forest patches in a highly fragmented forest landscape of southern Brazilian Amazonia. Classic habitat area effects were the strongest predictors of species persistence, explaining between 42% and 55% of the overall variation in primate and carnivore species richness. Logistic regression models showed that anthropogenic disturbance, including surface wildfires, timber extraction and hunting pressure, had detrimental effects on the persistence of some species over and above those of fragment size. Different species ranged in their responses from highly sensitive to highly tolerant to forest fragmentation. Patterns of local extinction documented here were by no means chance events, and the nestedness of the overall species-by-site matrix was highly nonrandom in terms of the sets of species extirpated from the most to the least species-rich forest patches.     

Recent declines in northern tall-grass prairies and effects of patch structure on community persistence

Tall-grass prairies are a critically endangered ecosystem in North America. Our objectives were to evaluate potential roles of prairie patch structure (defined in terms of prairie patch area, matrix type, and edge effects) in explaining changes in number, size, and quality of northern tall-grass prairies over time. In 2006, we evaluated changes in remnant tall-grass prairies at the most northern extent of the tall-grass prairie range, by resurveying plant communities in 65 remnant patches in Manitoba, Canada, that were previously surveyed in 1987 or 1988. In 2007 and 2008 we conducted more detailed surveys of vegetation structure and composition at 580, 0.2 × 0.5 m quadrats distributed within 24 remnant patches of northern tall-grass prairie. Our findings suggest remnant northern tall-grass prairies continue to suffer from serious threats: 37% of the patches surveyed in 1987 or 1988 had changed to other habitat types by 2006; patches smaller than 21 ha tended to decrease in size, while larger patches increased in size; and most patches, particularly smaller ones, declined in quality. Both native and alien species responded more strongly to distance to edge than to patch size or matrix type. Edge effects may explain why prairie quality is lower and more likely to decline in smaller remnants. Richness of native plants was negatively correlated with cover and richness of alien species, suggesting that alien species may displace native species. Few existing northern tall-grass prairies are likely to be self-sustaining, and immediate active management is required to prevent further loss of remnant northern tall-grass prairies.     

Forest loss with urbanization predicts bird extirpations in Vancouver

Forest loss with accompanying urbanization is often permanent as forests are replaced by impervious concrete structures or surfaces. The remaining forest patches are usually isolated and too small to maintain viable populations of forest-dependent wildlife species. Consistent with the species-area function (S = cA^z), extirpation of species should follow forest loss. However, this has rarely ever been tested in the urban landscape. We determined whether the observed number of forest-dependent bird extirpations occurring after forest loss matches that predicted by the species-area function in Vancouver (study area ∼126.7 km²), British Columbia, Canada. Prior to European settlement in 1859, Vancouver was covered completely by coastal western hemlock forest. Based on published historical maps of forest cover and present Landsat imagery of Vancouver, we estimated that approximately 87% of the forest area in Vancouver has been converted to urban development. Using the species-area function, we then predicted that approximately 14 bird species closely associated with lowland forest should have been extirpated. This figure is significantly greater than the three bird species known to have been extirpated in Vancouver. When we constrained the comparison to bird species closely associated with lowland forest and restricted in their distribution in British Columbia, we found a close match between the observed and predicted number of extirpations, suggesting that bird species closely associated with lowland forest and restricted in their geographic distribution are most prone to extirpation as a result of forest loss. Given that urban areas are situated in productive habitats with inherently high species richness, it is important that local land-use plans incorporate the conservation of habitat fragments from the onset and place particular emphasis on the regionally restricted forest-dependent species.     

The relative importance of patch habitat quality and landscape attributes on a declining steppe-bird metapopulation

Metapopulation theory is one of the most popular approaches to identify the factors affecting the spatial and temporal dynamics of populations in fragmented habitat networks. Habitat quality, patch area and isolation are mainly focused on when analyzing distribution patterns in fragmented landscapes. The effects of landscape heterogeneity in the non-occupied matrix, however, have been largely neglected. Here, we determined the relative importance of patch quality and landscape attributes on the occurrence, density and extinction of the Dupont’s lark (Chersophilus duponti), an endangered steppe passerine whose habitat has been extremely reduced to highly isolated and fragmented patches embedded in a mainly unsuitable landscape matrix. Habitat patch quality, measured in terms of vegetation structure, grazing pressure, arthropod availability, predator abundance, and inter-specific competition, did not affect occurrence, density or extinction. At the landscape scale, however, the species’ occurrence was principally determined by the interactions among patch size, geographic isolation and landscape matrix. Isolation had the main independent contribution to explaining the probability of occurrence, followed by landscape matrix composition and patch size. The species’ density was negatively correlated to patch size, suggesting crowding effects in small fragments, while extinction events were exclusively related to isolation. Our findings suggest that landscape rather than local population characteristics are crucial in determining the patterns of distribution and abundance of non-equilibrium populations in highly fragmented habitat networks. Consequently, conservation measures for these species should simultaneously involve patch size, isolation and landscape matrix and apply to the entire metapopulation rather than to particular patches.     

The effect of the surrounding landscape matrix on mangrove bird community assembly in north Australia

Mangroves are highly threatened ecosystems yet their community ecology is poorly understood. We examined the ecological determinants of bird community assemblage in floristically depauperate mangroves. Birds were surveyed using line transect methods. Large mangrove patches supported fewer species than smaller patches. Patches did not comprise nested species subsets and the bird species richness of several small patches combined was greater than a single large area. The number of mangrove dependent species in a patch was area-dependent suggesting these species may be resource limited, although there was no species density compensation. There was a clear effect of the surrounding habitat, with matrix species accounting for ∼45% of bird species in a patch. Patches surrounded by tropical savanna were relatively species-poor, while regardless of size, patches including monsoon rainforest were relatively species rich. Null model analysis of non-random assemblage structure (nestedness and species co-occurrence) revealed no deterministic structure to the overall mangrove species assemblage. These analyses described a random pattern of bird distribution and with no evidence of density compensation this suggests that competition is a weak structuring force of mangrove bird assemblages. The lack of nestedness and the random co-occurrence of species are consistent with the matrix-dependence of bird community composition. Conservation plans should treat mangrove patches as part of a habitat mosaic and incorporate many smaller mangrove patches rather than just big ones. Consideration of the nature, extent and diversity of the surrounding matrices is vital in managing and conserving mangrove bird communities.     

The effects of extreme forest fragmentation on the bird community of Singapore Island

Singapore Island suffered one of the highest known deforestation rates in the tropics from the mid-to-late 19th century when over 95% of its native lowland forest was cleared. We compared the current bird community structure and composition among three habitat types, i.e., old (>50 years, 7-935 ha) and young (?50 years, 29-49 ha) naturally regenerating secondary forests and abandoned wooded plantations (27-102 ha) dominated by exotic species. Forest patch area had the strongest influence on the current species richness. The overall bird richness was not greater in most mature forest patches, but 20 species were only found in the old secondary forests and five of these were found in <50% of these patches. The rapid decrease in the number of forest species in plantations was offset by an increase in the number of open habitat species. Comparisons with current bird communities in nearby mainland forest sites (Peninsular Malaysia) suggest that the forest avifauna of Singapore is depauperate. The preservation of larger mature and maturing forests is therefore required for conserving the extant forest avifauna in Singapore. Connecting isolated patches can also be envisioned to facilitate movements of forest birds that have low densities and restricted distribution.