Nested Subsets and the Distribution of Birds on Isolated Woodlots

Abstract. Distribution of bird species among isolated habitat patches (e.g., woodlots in an agricultural landscape) often appears to be nonrandom; species present in small, species-poor patches also are found in larger patches that support more species. Bird communities form 'nested subsets' (after Patterson & Atmar 1986) if all species found in small faunas also are found in more species-rich assemblages. Occurrence of a nested subset pattern implies an underlying, nonrandom pattern of species distributions. I used computer simulations to analyze the degree of nestedness exhibited by bird communities in east-central Illinois. Results demonstrated that the distribution of bird species breeding in isolated woodlots (1.8 to 600 ha) differed significantly from that predicted by a random distribution model; species assemblages were more "nested" than expected by chance. Most species present in small, species-poor woodlots also were found in larger, species-rich woodlots. As groups, species requiring forest interior habitat for breeding and species wintering in the tropics showed highly nested distributions. In contrast, short-distance migrants and species breeding in forest edge habitat showed more variable distribution patterns, species recorded on smaller woodlots were not always recorded on larger, more species-rich wood-lots. Apparent absences from larger woodlots may have reflected real distribution patterns or insufficient sampling of edge habitats. These results support previous conclusions that small habitat patches are insufficient for preservation of many species.     

Factors affecting the floristic diversity and nestedness in the islets of Lake Bardawil, North Sinai, Egypt: implications for conservation

Island biogeography theory and nestedness are two relevant and important concepts in biogeography and conservation. However, their integration and application has not been examined for maintaining or maximizing species diversity, especially in arid regions. The aim of this study is to investigate the ecological mechanisms related to the observed pattern of floristic diversity and nestedness in the Lake Bardawil islets on the Mediterranean coast of Sinai Peninsula before the execution of the North Sinai Agricultural Development Project (NSADP) that will threaten the ecosystem of the lake. Plant species on 15 islets were identified and categorized into ecological groups related to their life-form, salt tolerance and succulence. Richness of total plant species and their ecological groups were positively correlated with islet area, number of habitats and elevation, and negatively with disturbance index. The temperature calculator detected highly significant nestedness for the entire flora and all ecological groups. Distance from the mainland had no effect on either species richness or nested pattern of total plant species and their ecological groups. Both the analyses of species richness and nested distribution yielded evidence for a positive species-area relationship. Although nestedness was detected for the entire plant species and their ecological groups, many species and islets exhibited idiosyncratic distribution. Idiosyncrasies could be mainly attributed to demographic stochasticity which is an important character for species interaction and diversity maintenance in arid regions. Therefore, in developing conservation strategies for the Lake Bardawil islets, it is important to incorporate spatial and temporal stability of populations.     

Nested Faunas and Extinction in Fragmented Habitats

Abstract: The faunas of habitat islands, such as those produced by fragmentation of formerly continuous habitats, are commonly made up of nonrandom subsets of the total available species pool. Faunas within an archipelago may form a nested series, with depauperate faunas made up of subsets of more species-rich faunas. The pattern is seldom perfect: widespread species may be absent from otherwise rich faunas (holes), and uncommon species may occur in depauperate faunas (outliers). The "nestedness" of an assemblage of faunas (its fit to the "nested subset model"; Patterson & Atmar 1986) can be measured by summing the holes and outliers.
The distributions of boreal mammals and birds among mountain ranges in the Great Basin of western North America were analyzed by this method Despite differences in their derivation (mammalian faunas are thought to be relicts, isolated since the Pleistocene; bird faunas have probably experienced recurrent colonization throughout their history] both groups show an approximately equal (and highly significant) fit to the nested subset model. They differ, however, in the relative numbers of holes and outliers in their patterns. The mammalian pattern is hole-rich; the bird pattern is outlier-rich.
The present-day composition of the mammalian faunas is the result of selective extinction of species of originally richer faunas. The existence of nested subsets in these faunas suggests that extinction is a highly deterministic process: extinctions occurred in approximately the same sequence throughout the region, despite wide variation in extinction rates Extinction sequence has, in fact, been less variable than extinction rates.     

Floristic composition of arid off –shore islands in Abu Dhabi,United Arab Emirates

Data on flora and vegetation of 14 off-shore islands representing different habitat types were studied for investigating floristic composition, vegetation types and correlation with their habitats and ecosystem. The results demonstrate considerable plant distribution and diversity among the islands within the limited spectrum of species; encompassing a total of 47 species, belonging to 43 genera under 24 families. The most representative families were Amaranthaceae (9 species), Fabaceae (4 species), Poaceae (4 species) and Asteraceae and Aizoaceae (3 species each). Fourteen families are represented by only one species. The mangrove species, Avicennia marina, occurs throughout the coast of Abu Dhabi in discontinuous patches and in different water salinities. The studies also seek to underline that dominant species that constitutes the flora of Abu Dhabi are salt tolerant in nature. Compared to other Arabian Peninsula ecosystems, UAE perhaps has a higher coastal: mainland area ratio, which has contributed to a natural dominance of salt tolerant species among the community. The studies show 40% similarity among islands with respect to the species diversity. The floristic composition of the off-shore islands also indicates a need to consider these sites as protected sites.     

Food webs are built up with nested subwebs

Nested structure, in which specialists interact with subsets of species with which generalists interact, has been repeatedly found in networks of mutualistic interactions and thus is considered a general feature of mutualistic communities. However, it is uncertain how exclusive nested structure is for mutualistic communities since few studies have evaluated nestedness in other types of networks. Here, we show that 31 published food webs consist of bipartite subwebs that are as highly nested as mutualistic networks, contradicting the hypothesis that antagonistic interactions disfavor nested structure. Our findings suggest that nested networks may be a common pattern of communities that include resource-consumer interactions. In contrast to the hypothesis that nested structure enhances biodiversity in mutualistic communities, we also suggest that nested food webs increase niche overlap among consumers and thus prevent their coexistence. We discuss potential mechanisms for the emergence of nested structure in food webs and other types of ecological networks.     

Distinguishing stressors acting on land bird communities in an urbanizing environment

Urbanization has profound influences on ecological communities, but our understanding of causal mechanisms is limited by a lack of attention to its component stressors. Published research suggests that at landscape scales, habitat loss and fragmentation are the major drivers of community change, whereas at local scales, human activity and vegetation management are the primary stressors. Little research has focused on whether urbanization stressors may supplant natural factors as dominant forces structuring communities. We used model selection to determine the relative importance of urban development, human activity, local and landscape vegetation, topography, and geographical location in explaining land bird species richness, abundance, and dominance. We analyzed the entire community and groups of species based on ecological characteristics, using data collected in remnant forests along a gradient of urban development in the Lake Tahoe basin, California and Nevada, USA. Urbanization stressors were consistently among the principal forces structuring the land bird community. Strikingly, disturbance from human activity was the most important factor for richness in many cases, surpassing even habitat loss from development. Landscape-scale factors were consistently more important than local-scale factors for abundance. In demonstrating considerable changes in land bird community structure, our results suggest that ecosystem function in urban areas may be severely compromised. Such changes compel local- and landscape-scale management, focused research, and long-term monitoring to retain biodiversity in urban areas to the extent possible.     

Evaluating Effects of Habitat Loss and Land-Use Continuity on Ant Species Richness in Seminatural Grassland Remnants

Abstract:  Seminatural grasslands in Europe are susceptible to habitat destruction and fragmentation that result in negative effects on biodiversity because of increased isolation and area effects on extinction rate. However, even small habitat patches of seminatural grasslands might be of value for conservation and restoration of species richness in a landscape with a long history of management, which has been argued to lead to high species richness. We tested whether ant communities have been negatively affected by habitat loss and increased isolation of seminatural grasslands during the twentieth century. We examined species richness and community composition in seminatural grasslands of different size in a mosaic landscape in Central Sweden. Grasslands managed continuously over centuries harbored species-rich and ecologically diverse ant communities. Grassland remnant size had no effect on ant species richness. Small grassland remnants did not harbor a nested subset of the ant species of larger habitats. Community composition of ants was mainly affected by habitat conditions. Our results suggest that the abandonment of traditional land use and the encroachment of trees, rather than the effects of fragmentation, are important for species composition in seminatural grasslands. Our results highlight the importance of considering land-use continuity and dispersal ability of the focal organisms when examining the effects of habitat loss and fragmentation on biodiversity. Landscape history should be considered in conservation programs focusing on effects of land-use change.     

A Natural Experiment on the Impact of Overabundant Deer on Forest Invertebrates

Abstract:  In large parts of North America and Europe, deer overabundance threatens forest plant diversity. Few researchers have examined its effects on invertebrate assemblages. In a natural experiment on Haida Gwaii (British Columbia, Canada), where Sitka black-tailed deer ( Odocoileus hemionus sitkensis ) were introduced, we compared islands with no deer, with deer for fewer than 20 years, and with deer for more than 50 years. We sampled invertebrates in three habitat categories: forest edge vegetation below the browse line, forest interior vegetation below the browse line, and forest interior litter. In forest edge vegetation, invertebrate abundance and species density decreased with increasing length of browsing history. In forest interior vegetation, decrease was significant only on islands with more than 50 years of browsing. Insect abundance in the vegetation decreased eightfold and species density sixfold on islands browsed for more than 50 years compared with islands without deer. Primary consumers were most affected. Invertebrates from the litter showed little or no variation related to browsing history. We attributed the difference between vegetation-dwelling and litter-dwelling invertebrates to differences in the effect of browsing on their habitat. In the layer below the browse line deer progressively removed the habitat. The extent of litter habitat was not affected, but its quality changed. We recommend more attention be given to the effect of overabundant ungulates on forest invertebrate conservation with a focus on edge and understory vegetation in addition to litter habitat.     

Bat Diversity and Abundance as Indicators of Disturbance in Neotropical Rainforests

Abstract: Evaluating the degree of disturbance of any region to determine its relative importance for conservation purposes requires procedures that are relatively inexpensive and that yield accurate results fast. Because bats are abundant, diverse, and easy to sample, especially in the Neotropical rainforest, they fulfill several of the requirements of indicator species as identified in the literature. For 10 months we sampled bat communities in the Selva Lacandona in Chiapas, Mexico, at 15 sites representing five habitats. We also measured 10 variables representing vegetation structure and diversity at each site. With fuzzy-set techniques we produced a gradient classification of disturbance for the 15 sites based on the vegetation data. We explored the relationship between vegetation conditions, described as the membership degrees in the construct "fuzzy forest set" (the complementary fuzzy set of "disturbance"), and four bat community variables. Bat species richness, number of rare bat species, and the bat diversity index were positively correlated with the vegetation scores, and relative abundance of the most abundant bat species was negatively correlated with vegetation scores. A high number of phyllostomine species in a community is a good indicator of low levels of disturbance. Although a single indicator group will probably not be sufficient for decision-making processes in conservation, evaluating bat populations may be a good first step in assessing an area's conservation value, especially in rainforest regions.     

Differences in gypsum plant communities associated with habitat fragmentation and livestock grazing.

The negative consequences of habitat fragmentation for plant communities have been documented in many regions of the world. In some fragmented habitats, livestock grazing has been proposed to be a dispersal mechanism reducing isolation between fragments. In others, grazing acts together with fragmentation in a way that increases habitat degradation. Iberian gypsum plant communities have been grazed and fragmented by agricultural practices for centuries. Although their conservation is considered a priority by the European Community, the effects of fragmentation on gypsum plant communities and the possible role of livestock grazing remain unknown. In addition, a substantial proportion of plant species growing in gypsum environments are gypsum specialists. They could be particularly affected by fragmentation, as was found for other habitat specialists (i.e., serpentine and calcareous specialists). In this study (1) we investigated the effect of fragmentation and grazing on gypsum plant community composition (species and life-forms), and (2) we tested to see if gypsum specialists were differently affected by fragmentation and grazing than habitat generalists. A vegetation survey was conducted in the largest gypsum outcrop of Europe (Middle Ebro Valley, northeast Spain). Fragmented and continuous sites in grazed and ungrazed areas were compared. Measurements related to species and composition of life-forms were contrasted first for the whole gypsum plant community and then specifically for the gypsum specialists. In the whole community, our results showed lower plant species diversity in fragmented sites, mainly due to the larger dominance of species more tolerant to fragmented habitat conditions. With livestock grazing, the plant species richness and the similarity in plant species composition between remnants was larger, suggesting that animals were acting as dispersal agents between fragments. As expected, gypsum specialists were less abundant in fragmented areas, and grazing led to the disappearance of the rare gypsum specialist Campanula fastigiata. According to our results, conservation strategies for gypsum plant communities in human-dominated landscapes should consider that fragmentation and grazing modify plant community composition affecting gypsum specialists in particular.     

Planting gardens to support insect pollinators

Global insect pollinator declines have prompted habitat restoration efforts, including pollinator-friendly gardening. Gardens can provide nectar and pollen for adult insects and offer reproductive resources, such as nesting sites and caterpillar host plants. We conducted a review and meta-analysis to examine how decisions made by gardeners on plant selection and garden maintenance influence pollinator survival, abundance, and diversity. We also considered characteristics of surrounding landscapes and the impacts of pollinator natural enemies. Our results indicated that pollinators responded positively to high plant species diversity, woody vegetation, garden size, and sun exposure and negatively to the separation of garden habitats from natural sites. Within-garden features more strongly influenced pollinators than surrounding landscape factors. Growing interest in gardening for pollinators highlights the need to better understand how gardens contribute to pollinator conservation and how some garden characteristics can enhance the attractiveness and usefulness of gardens to pollinators. Further studies examining pollinator reproduction, resource acquisition, and natural enemies in gardens and comparing gardens with other restoration efforts and to natural habitats are needed to increase the value of human-made habitats for pollinators.     

Effects of Patch Size on Birds in Old-Growth Montane Forests

Following habitat alteration or fragmentation, competition, parasitism, and predation from species that live in the new habitats may reduce the survival and reproductive success of species living in the original habitats. Negative influences from species that live outside the remnant patches are expected to be greater in small rather than in large remnant patches because more "external" species are expected to move through the centers of small remnant patches. We surveyed birds within remnant patches of old-growth montane forests on Vancouver Island, Canada, (1) to evaluate whether the richness and abundance of non-old-growth bird species were larger at the center of small rather than large patches and (2) to evaluate whether the opposite was true of old-growth bird species. More non-old-growth bird species were present at the center of small remnant patches of old growth than in large old-growth patches. We found no relationship, however, between patch size and richness or abundance of old-growth bird species at the center of remnant patches of old growth. This was true for old-growth species with open, cup-shaped nests and cavity nests. Old-growth birds may have been affected less in our study area than in other areas because they evolved within heterogeneous montane forests and interacted with non-old-growth species throughout their evolutionary histories or because the contrast between old-growth forests and logged areas was less than that between the forests and agricultural/urban areas that were surveyed in other studies.     

Changes in hardwood forest understory plant communities in response to European earthworm invasions

European earthworms are colonizing earthworm-free northern hardwood forests across North America. Leading edges of earthworm invasion provide an opportunity to investigate the response of understory plant communities to earthworm invasion and whether the species composition of the earthworm community influences that response. Four sugar maple-dominated forest sites with active earthworm invasions were identified in the Chippewa National Forest in north central Minnesota, USA. In each site, we established a 30 x 150 m sample grid that spanned a visible leading edge of earthworm invasion and sampled earthworm populations and understory vegetation over four years. Across leading edges of earthworm invasion, increasing total earthworm biomass was associated with decreasing diversity and abundance of herbaceous plants in two of four study sites, and the abundance and density of tree seedlings decreased in three of four study sites. Sample points with the most diverse earthworm species assemblage, independent of biomass, had the lowest plant diversity. Changes in understory plant community composition were most affected by increasing biomass of the earthworm species Lumbricus rubellus. Where L. rubellus was absent there was a diverse community of native herbaceous plants, but where L. rubellus biomass reached its maximum, the herbaceous-plant community was dominated by Carex pensylvanica and Arisaema triphyllum and, in some cases, was completely absent. Evidence from these forest sites suggests that earthworm invasion can lead to dramatic changes in the understory community and that the nature of these changes is influenced by the species composition of the invading earthworm community.     

异质生境空心莲子草植物群落组成及物种多样性研究

研究草海湿地空心莲子草（Alternanthera philoxeroides）入侵对异质生境植物群落生活型、物种组成、重要值及物种多样性的影响,结果表明,①异质生境中群落生活型组成比例有差异,水生生境地面芽植物（H）∶隐芽植物（Cr）∶一年生植物（Th）为56∶38∶6;湿生生境为43∶36∶21。②异质生境中物种组成差异明显,水生生境试验样方包含9科13属16种,对照样方（无空心莲子草入侵）包含10科19属23种;湿生生境试验样方包含8科14属14种,对照样方有10科19属22种。空心莲子草入侵不同生境后群落内物种重要值变化及差异均较大,部分乡土物种在试验样方中减少甚至消失,水生生境和湿生生境群落中分别有30.44%、36.36%物种消失。③异质生境中空心莲子草重要值与物种多样性指数、生态优势度指数、均匀度指数均呈二项式函数关系;两种生境的多样性指数和均匀度指数均表现为随空心莲子草重要值先增加后减少,生态优势度指数均为先减少后增加,空心莲子草入侵对乡土植物群落物种多样性产生正向影响和负向影响,在群落中其重要值达25.8（水生生境）、18.9（湿生生境）时,是乡土植物多样性下降的临界点。     

Null model analysis of species nestedness patterns

Nestedness is a common biogeographic pattern in which small communities form proper subsets of large communities. However, the detection of nestedness in binary presence-absence matrices will be affected by both the metric used to quantify nestedness and the reference null distribution. In this study, we assessed the statistical performance of eight nestedness metrics and six null model algorithms. The metrics and algorithms were tested against a benchmark set of 200 random matrices and 200 nested matrices that were created by passive sampling. Many algorithms that have been used in nestedness studies are vulnerable to type I errors (falsely rejecting a true null hypothesis). The best-performing algorithm maintains fixed row and fixed column totals, but it is conservative and may not always detect nestedness when it is present. Among the eight indices, the popular matrix temperature metric did not have good statistical properties. Instead, the Brualdi and Sanderson discrepancy index and Cutler's index of unexpected presences performed best. When used with the fixed-fixed algorithm, these indices provide a conservative test for nestedness. Although previous studies have revealed a high frequency of nestedness, a reanalysis of 288 empirical matrices suggests that the true frequency of nested matrices is between 10% and 40%.     

Declines in Woodland Salamander Abundance Associated with Non-Native Earthworm and Plant Invasions

Abstract:  Factors that negatively affect the quality of wildlife habitat are a major concern for conservation. Non-native species invasions, in particular, are perceived as a global threat to the quality of wildlife habitat. Recent evidence indicates that some changes to understory plant communities in northern temperate forests of North America, including invasions by 3 non-native plant species, are facilitated by non-native earthworm invasion. Furthermore, non-native earthworm invasions cause a reduction in leaf litter on the forest floor, and the loss of forest leaf litter is commonly associated with declines in forest fauna, including amphibians. We conducted a mark-recapture study of woodland salamander abundance across plant invasion fronts at 10 sites to determine whether earthworm or plant invasions were associated with reduced salamander abundance. Salamander abundance declined exponentially with decreasing leaf litter volume. There was no significant relationship between invasive plant cover and salamander abundance, independent of the effects of leaf litter loss due to earthworm invasion. An analysis of selected salamander prey abundance (excluding earthworms) at 4 sites showed that prey abundance declined with declining leaf litter. The loss of leaf litter layers due to non-native earthworm invasions appears to be negatively affecting woodland salamander abundance, in part, because of declines in the abundance of small arthropods that are a stable resource for salamanders. Our results demonstrate that earthworm invasions pose a significant threat to woodland amphibian fauna in the northeastern United States, and that plant invasions are symptomatic of degraded amphibian habitat but are not necessarily drivers of habitat degradation.     

Complementary habitat use by wild bees in agro-natural landscapes

Human activity causes abrupt changes in resource availability across the landscape. In order to persist in human-altered landscapes organisms need to shift their habitat use accordingly. Little is known about the mechanisms by which whole communities persist in human-altered landscapes, including the role of complementary habitat use. We define complementary habitat use as the use of different habitats at different times by the same group of species during the course of their activity period. We hypothesize that complementary habitat use is a mechanism through which native bee species persist in human-altered landscapes. To test this idea, we studied wild bee communities in agro-natural landscapes and explored their community-level patterns of habitat and resource use over space and time. The study was conducted in six agro-natural landscapes in the eastern United States, each containing three main bee habitat types (natural habitat, agricultural fields, and old fields). Each of the three habitats exhibited a unique seasonal pattern in amount, diversity, and composition of floral resources, and together they created phenological complementarity in foraging resources for bees. Individual bee species as well as the bee community responded to these spatiotemporal patterns in floral availability and exhibited a parallel pattern of complementary habitat use. The majority of wild bee species, including all the main crop visitors, used fallow areas within crops early in the season, shifted to crops in mid-season, and used old-field habitats later in the season. The natural-forest habitat supported very limited number of bees, mostly visitors of non-crop plants. Old fields are thus an important feature in these arable landscapes for maintaining crop pollination services. Our study provides a detailed examination of how shifts in habitat and resource use may enable bees to persist in highly dynamic agro-natural landscapes, and points to the need for a broad cross-habitat perspective in managing these landscapes.     

Habitat biodiversity as a determinant of fish community structure on coral reefs

Increased habitat diversity is often predicted to promote the diversity of animal communities because a greater variety of habitats increases the opportunities for species to specialize on different resources and coexist. Although positive correlations between the diversities of habitat and associated animals are often observed, the underlying mechanisms are only now starting to emerge, and none have been tested specifically in the marine environment. Scleractinian corals constitute the primary habitat-forming organisms on coral reefs and, as such, play an important role in structuring associated reef fish communities. Using the same field experimental design in two geographic localities differing in regional fish species composition, we tested the effects of coral species richness and composition on the diversity, abundance, and structure of the local fish community. Richness of coral species overall had a positive effect on fish species richness but had no effect on total fish abundance or evenness. At both localities, certain individual coral species supported similar levels of fish diversity and abundance as the high coral richness treatments, suggesting that particular coral species are disproportionately important in promoting high local fish diversity. Furthermore, in both localities, different microhabitats (coral species) supported very different fish communities, indicating that most reef fish species distinguish habitat at the level of coral species. Fish communities colonizing treatments of higher coral species richness represented a combination of those inhabiting the constituent coral species. These findings suggest that mechanisms underlying habitat-animal interaction in the terrestrial environment also apply to marine systems and highlight the importance of coral diversity to local fish diversity. The loss of particular key coral species is likely to have a disproportionate impact on the biodiversity of associated fish communities.     

Tropical coastal habitats as surrogates of fish community structure, grazing, and fisheries value

Habitat maps are frequently invoked as surrogates of biodiversity to aid the design of networks of marine reserves. Maps are used to maximize habitat heterogeneity in reserves because this is likely to maximize the number of species protected. However, the technique's efficacy is limited by intra-habitat variability in the species present and their abundances. Although communities are expected to vary among patches of the same habitat, this variability is poorly documented and rarely incorporated into reserve planning. To examine intra-habitat variability in coral-reef fishes, we generated a data set from eight tropical coastal habitats and six islands in the Bahamian archipelago using underwater visual censuses. Firstly, we provide further support for habitat heterogeneity as a surrogate of biodiversity as each predefined habitat type supported a distinct assemblage of fishes. Intra-habitat variability in fish community structure at scales of hundreds of kilometers (among islands) was significant in at least 75% of the habitats studied, depending on whether presence/absence, density, or biomass data were used. Intra-habitat variability was positively correlated with the mean number of species in that habitat when density and biomass data were used. Such relationships provide a proxy for the assessment of intra-habitat variability when detailed quantitative data are scarce. Intra-habitat variability was examined in more detail for one habitat (forereefs visually dominated by Montastraea corals). Variability in community structure among islands was driven by small, demersal families (e.g., territorial pomacentrid and labrid fishes). Finally, we examined the ecological and economic significance of intra-habitat variability in fish assemblages on Montastraea reefs by identifying how this variability affects the composition and abundances of fishes in different functional groups, the key ecosystem process of parrotfish grazing, and the ecosystem service of value of commercially important finfish. There were significant differences in a range of functional groups and grazing, but not fisheries value. Variability at the scale of tens of kilometers (among reefs around an island) was less than that among islands. Caribbean marine reserves should be replicated at scales of hundreds of kilometers, particularly for species-rich habitats, to capture important intra-habitat variability in community structure, function, and an ecosystem process.     

Selection of Islands for Conservation in the Urban Archipelago of Helsinki, Finland

Abstract: The occurrence of vascular plants was surveyed on 207 islands (size range 0.01–390.2 ha, number of plant species 1–449) offshore from the city of Helsinki in the Baltic Sea to examine the conservation value of these islands. We calculated a rarity score for each species (1/number of islands occupied by the species) and a biodiversity score for each island (sum of the rarity scores of each species present on the island). Positive correlations between species number and biodiversity score (r _s = 0.97, p < 0.001) and between biodiversity score and island area (r _s = 0.87, p < 0.001) indicated that these parameters are heavily dependent on island size. With the goal of including at least one occurrence (island) of all plant species, an iterative selection algorithm chose a set of 41 islands whose average size (29.3 ha) was four times the average size of all existing islands (7.0 ha). Strong nestedness ( N < 54) explains the concentration of plant species diversity on large islands. An operational strategy for selection of sites for protection is to complement the set produced by a selection algorithm with target species not yet included (e.g., endangered species with several occurrences). Comprehensive mapping and analysis of a taxonomic group will help integrate conservation biology into land-use planning and increase the quality of the networks of protected areas.