Seed Banks in Savanna,Forest Fragments,and Continuous Forest Edges Differ in a Tropical Dry Forest in Madagascar

Rapid deforestation has fragmented habitat across the landscape of Madagascar. To determine the effect of fragmentation on seed banks and the potential for forest regeneration, we sampled seed viability, density and diversity in 40 plots of 1 m² in three habitat types: forest fragments, the near edge of continuous forest, and deforested savanna in a highly fragmented dry deciduous forest landscape in northwestern Madagascar. While seed species diversity was not different between forest fragments and continuous forest edge, the number of animal‐dispersed seeds was significantly higher in forest fragments than in continuous forest edge, and this pattern was driven by a single, small‐seeded species. In the savanna, seeds were absent from all but three of the 40 plots, indicating that regeneration potential is low in these areas. Several pre‐ and post‐dispersal biotic and abiotic factors, including variation in the seed predator communities and edge effects could explain these findings. Understanding the extent to which seed dispersal and seed banks influence the regeneration potential of fragmented landscapes is critical as these fragments are the potential sources of forest expansion and re‐connectivity.     

Habitat loss and fragmentation reduce effective gene flow by disrupting seed dispersal in a neotropical palm

Habitat loss and fragmentation often reduce gene flow and genetic diversity in plants by disrupting the movement of pollen and seed. However, direct comparisons of the contributions of pollen vs. seed dispersal to genetic variation in fragmented landscapes are lacking. To address this knowledge gap, we partitioned the genetic diversity contributed by male gametes from pollen sources and female gametes from seed sources within established seedlings of the palm Oenocarpus bataua in forest fragments and continuous forest in northwest Ecuador. This approach allowed us to quantify the separate contributions of each of these two dispersal processes to genetic variation. Compared to continuous forest, fragments had stronger spatial genetic structure, especially among female gametes, and reduced effective population sizes. We found that within and among fragments, allelic diversity was lower and genetic structure higher for female gametes than for male gametes. Moreover, female gametic allelic diversity in fragments decreased with decreasing surrounding forest cover, while male gametic allelic diversity did not. These results indicate that limited seed dispersal within and among fragments restricts genetic diversity and strengthens genetic structure in this system. Although pollen movement may also be impacted by habitat loss and fragmentation, it nonetheless serves to promote gene flow and diversity within and among fragments. Pollen and seed dispersal play distinctive roles in determining patterns of genetic variation in fragmented landscapes, and maintaining the integrity of both dispersal processes will be critical to managing and conserving genetic variation in the face of continuing habitat loss and fragmentation in tropical landscapes.     

The effect of habitat fragmentation on the bee visitor assemblages of three Australian tropical rainforest tree species

Tropical forest loss and fragmentation can change bee community dynamics and potentially interrupt plant–pollinator relationships. While bee community responses to forest fragmentation have been investigated in a number of tropical regions, no studies have focused on this topic in Australia. In this study, we examine taxonomic and functional diversity of bees visiting flowers of three tree species across small and large rainforest fragments in Australian tropical landscapes. We found lower taxonomic diversity of bees visiting flowers of trees in small rainforest fragments compared with large forest fragments and show that bee species in small fragments were subsets of species in larger fragments. Bees visiting trees in small fragments also had higher mean body sizes than those in larger fragments, suggesting that small‐sized bees may be less likely to persist in small fragments. Lastly, we found reductions in the abundance of eusocial stingless bees visiting flowers in small fragments compared to large fragments. These results suggest that pollinator visits to native trees living in small tropical forest remnants may be reduced, which may in turn impact on a range of processes, potentially including forest regeneration and diversity maintenance in small forest remnants in Australian tropical countryside landscapes.     

How landscapes affect the evolution of dispersal behaviour in reef fishes: results from an individual‐based model

The vast majority of tropical reef fishes have a sedentary adult phase and pelagic larval phase that is potentially highly dispersive. Dispersal may be favoured by a wide range of factors including the arrangement of suitable habitat in space. In this paper the dispersal strategy of individuals is followed and allowed to evolve in a simplified model of three different landscapes: an enclosed sea, an open archipelago and a barrier reef. The three landscapes have very different characteristics, but all have similar spatial clumping of reef habitat. In all landscapes, as minimum time to settlement increases, evolved movement strategy also increases and longer settlement windows favour dispersal. In the archipelago movement is not maximized until the minimum pelagic duration is longer than in the other landscapes. The model predicts that, given the same pelagic duration, species from enclosed seas should have more dispersive behaviours than those from open archipelagos, because of the density of habitat and the aggregation of habitat in space affect the likelihood of larvae finding suitable habitat for settlement.     

Movement patterns of Rhyssomatus lineaticollis Say (Coleoptera: Curculionidae) within and among Asclepias syriaca (Asclepiadaceae) patches in a fragmented landscape

Abstract.  1. Dispersal capabilities of organisms are critical in determining the landscape population structure of species as well as their likelihood of survival in fragmented landscapes. Using mark–recapture techniques on the monophagous weevil Rhyssomatus lineaticollis Say (Curculionidae), within- and between-patch dispersal capabilities, landscape level population structure, and the role of beetle density and host patch characteristics in setting distances, amounts, and timing of dispersal were studied.
2. The data indicate that R. lineaticollis is sedentary, with 50% of recaptured beetles moving < 1 m and the maximum distance moved < 1 km. Within- and between-patch movement of beetles was unrelated to host plant patch characteristics and beetle densities.
3. Despite limited dispersal, R. lineaticollis probably functions as a patchy population in east-central Iowa, U.S.A. because dispersals between patches are common and because all host patches surveyed contained this herbivore, indicating a lack of suitable vacant patches, a prerequisite for metapopulation structure.
4. Between-patch distances are well within the dispersal capabilities of R. lineaticollis , although this may be the result of an increase in the density of patches of its host, Asclepias syriaca , in the landscape over the last 150 years as a result of human disturbance and this species' weedy habit.
5. Metapopulation structure in monophagous prairie herbivores may be most likely in species whose non-weedy host plants form highly predictable resources in space and time, but which are now widely scattered in habitat fragments.     

Functional connectivity and genetic diversity of Eulaema atleticana (Apidae,Euglossina) in the Brazilian Atlantic Forest Corridor: assessment of gene flow

The mobility and dispersal of organisms affect population genetics and dynamics, and consequently affect persistence and the risk of extinction. Thus, it is important to understand how organisms move in the fragmented landscapes in order to manage populations and predict the effects of habitat changes on species persistence. This study evaluated the functional connectivity of an orchid bee (Eulaema atleticana Nemésio, 2009) with a high fidelity to forest habitats in the Brazilian Atlantic Forest Corridor by analyzing genetic diversity, spatial genetic structure, and gene flow estimated from microsatellite and mitochondrial markers. Genetic diversity was not correlated with area of the forest fragments, or with forest isolation. At the mosaic scale, Eulaema atleticana showed no significant or low genetic differentiation, indicating genetic homogeneity among forest fragments. A previous field study indicated that Eulaema atleticana was one of the most sensitive Euglossina bees to forest fragmentation but the present molecular analyses demonstrates that current gene flow is sufficient to maintain genetic variability at the mosaic scale.     

Fragmentation of Atlantic Forest has not affected gene flow of a widespread seed‐dispersing bat

Habitat loss and resultant fragmentation are major threats to biodiversity, particularly in tropical and subtropical ecosystems. It is increasingly urgent to understand fragmentation effects, which are often complex and vary across taxa, time and space. We determined whether recent fragmentation of Atlantic forest is causing population subdivision in a widespread and important Neotropical seed disperser: Artibeus lituratus (Chiroptera: Phyllostomidae). Genetic structure within highly fragmented forest in Paraguay was compared to that in mostly contiguous forest in neighbouring Misiones, Argentina. Further, observed genetic structure across the fragmented landscape was compared with expected levels of structure for similar time spans in realistic simulated landscapes under different degrees of reduction in gene flow. If fragmentation significantly reduced successful dispersal, greater population differentiation and stronger isolation by distance would be expected in the fragmented than in the continuous landscape, and genetic structure in the fragmented landscape should be similar to structure for simulated landscapes where dispersal had been substantially reduced. Instead, little genetic differentiation was observed, and no significant correlation was found between genetic and geographic distance in fragmented or continuous landscapes. Furthermore, comparison of empirical and simulated landscapes indicated empirical results were consistent with regular long‐distance dispersal and high migration rates. Our results suggest maintenance of high gene flow for this relatively mobile and generalist species, which could be preventing or significantly delaying reduction in population connectivity in fragmented habitat. Our conclusions apply to A. lituratus in Interior Atlantic Forest, and do not contradict broad evidence that habitat fragmentation is contributing to extinction of populations and species, and poses a threat to biodiversity worldwide.     

Genetic signature of population fragmentation varies with mobility in seven bird species of a fragmented Kenyan cloud forest

Habitat fragmentation can restrict geneflow, reduce neighbourhood effective population size, and increase genetic drift and inbreeding in small, isolated habitat remnants. The extent to which habitat fragmentation leads to population fragmentation, however, differs among landscapes and taxa. Commonly, researchers use information on the current status of a species to predict population effects of habitat fragmentation. Such methods, however, do not convey information on species-specific responses to fragmentation. Here, we compare levels of past population differentiation, estimated from microsatellite genotypes, with contemporary dispersal rates, estimated from multi-strata capture-recapture models, to infer changes in mobility over time in seven sympatric, forest-dependent bird species of a Kenyan cloud forest archipelago. Overall, populations of sedentary species were more strongly differentiated and clustered compared to those of vagile ones, while geographical patterning suggested an important role of landscape structure in shaping genetic variation. However, five of seven species with broadly similar levels of genetic differentiation nevertheless differed substantially in their current dispersal rates. We conclude that post-fragmentation levels of vagility, without reference to past population connectivity, may not be the best predictor of how forest fragmentation affects the life history of forest-dependent species. As effective conservation strategies often hinge on accurate prediction of shifts in ecological and genetic relationships among populations, conservation practices based solely upon current population abundances or movements may, in the long term, prove to be inadequate.     

Use of Fruiting Trees by Birds in Continuous Forest and Riparian Forest Remnants in Los Tuxtlas,Veracruz, Mexico1

The composition of plant species in fragmented landscapes may be influenced by the pattern of visitation by birds to fruiting trees and by the movement of seeds among and within fragments. We compared bird visitation patterns to two tree species (Dendropanax arboreus, Araliaceae; Bursera simaruba, Burseraceae) in continuous forest and remnants of riparian vegetation in a region dominated by pasture in Los Tuxtlas, Veracruz, Mexico. We quantified frequency of visitation, fruit consumption, consistency of visitation (percentage of total tree observation periods during which a given bird species was recorded), and species composition of birds at individuals of both tree species in continuous forest and riparian remnants. Bird visitation rate, species richness, and fruit consumption rates were similar within both tree species in the two habitats. Species assemblages at D. arboreus were different between continuous forest and remnants. Species assemblages at B. simaruba did not differ by habitat. Our results demonstrate that habitat disturbance may influence avian visitation patterns, which may in turn affect subsequent recruitment patterns in some tree species. Our results, however, were not consistent between the tree species, suggesting that it is difficult to generalize concerning the effects of forest disturbance on avian species assemblages in fruiting trees.     

Edge Effects Act Differentially on Multiple Early Regeneration Stages of a Shade‐tolerant Tree Tapirira mexicana

Forest fragmentation is pervasive in tropical landscapes, and one pathway by which fragmentation may negatively impact populations is via edge effects. Early life‐stages are particularly important for species regeneration as they act as bottlenecks, but how edge effects may act differentially on different life‐stages is unknown. This study evaluated edge effects on multiple early life‐stages of a currently common animal‐dispersed, shade‐tolerant tree Tapirira mexicana (Anacardiaceae). The study was conducted in tropical premontane wet forest fragments in a highly deforested region of Costa Rica. The stages assessed were pre‐dispersal predation, primary dispersal, post‐dispersal predation, secondary dispersal, ex situ germination, in situ seed longevity, first and second year seedling abundance, second year seedling survivorship, and basal diameter growth. Results showed that impacts of edge effects were not equal across stages, but were limited to specific stages and times. One stage which may act as a bottleneck for species regeneration was pre‐dispersal predation. Over 60 percent of the seeds were predated by larvae, and predation was higher near the edge than interior habitat. Seeds lost viability within 10 d in the forest. Germination to first year seedling stage was also lower near edges, but such effect was eliminated within a year after that. Primary dispersal, seedling survivorship, and growth were not affected by proximity to edges, and both secondary dispersal and post‐dispersal predation were rare. This study demonstrates that current population abundance may not guarantee future species persistence and the importance of considering multiple life‐stages for a comprehensive assessment of forest fragmentation effects on species regeneration.     

The frailty of tropical restoration plantings

Conserving and restoring biodiversity are compelling challenges in the face of deforestation and fragmentation of tropical forests. Establishing restoration plantings that act as stepping‐stone corridors for animals and develop into forest islets is one way we can reconnect forest fragments split by active agricultural landscapes. However, this strategy's success is contingent on dispersal agents attracted from the forests that vary greatly in their dispersal services, diet, and mobility. Dispersal agents capable of traversing the agricultural matrix that also provide high‐quality and high‐quantity seed dispersal are often a small subset of the present fauna. They also tend to be large‐bodied birds with broad diets (e.g. toucans). This subset of dispersal agents (here termed “effective restoration agents”) plays a key role in driving succession in restoration plantings. Their absence or low numbers can compromise the strategy of using plantings to enhance connectivity in landscapes fragmented by crops, orchards, or extensive pastures. In this event, additional intervention may be required to attract other dispersal agents that would otherwise not disperse seeds at or play a significant role in restoration plantings.     

Gap-crossing movements predict species occupancy in Amazonian forest fragments

In fragmented landscapes, species persistence within isolated habitat patches is governed by a myriad of species life‐history, habitat patch and landscape characteristics. We investigated the inter‐specific variation in non‐forest gap‐crossing abilities of an entire tropical forest‐dependent avifauna. We then related this measure of dispersal ability to species life‐history characteristics and occupancy data from 31 variable‐sized forest patches sampled within the same fragmented forest landscape. A total of 5436 gap‐crossing movements of 231 forest‐dependent bird species were observed across ten linear forest gaps of varying widths, adjacent to large areas of undisturbed forest. Species persistence in isolated fragments was strongly linked to gap‐crossing ability. The most capable gap‐crossers were medium to large‐bodied species in the large insectivore, frugivore and granivore guilds, matching the most prevalent subset of species in small forest patches. However, some competent gap‐crossing species failed to occur in small patches, and minimum forest‐patch area requirements were more important in determining patch occupancy for these species. Narrow forest gaps (4–70 m) created by roads and power‐lines may become territory boundaries, thereby eliminating home‐range gap‐crossing movements for many forest species, but permit rarer dispersal events. Wider gaps (>70 m) may inhibit gap‐crossing behaviour for all but the most vagile species. Although patch size and quality may be the most important factors in structuring species assemblages in forest fragments, our results show that the degree of patch isolation and permeability of the surrounding matrix also explain which species can persist in forest isolates. Reducing the number and width of forest‐dividing gaps; maintaining and/or creating forest corridors and increasing matrix permeability through the creation and maintenance of ‘stepping‐stone’ structures will maximise the species retention in fragmented tropical forest landscapes.     

Possible effects of habitat fragmentation and climate change on the range of forest plant species

Global circulation models predict an increase in mean annual temperature between 2.1 and 4.6 °C by 2080 in the northern temperate zone. The associated changes in the ratio of extinctions and colonizations at the boundaries of species ranges are expected to result in northward range shifts for a lot of species. However, net species colonization at northern boundary ranges, necessary for a northward shift and for range conservation, may be hampered because of habitat fragmentation. We report the results of two forest plant colonization studies in two fragmented landscapes in central Belgium. Almost all forest plant species (85%) had an extremely low success of colonizing spatially segregated new suitable forest habitats after c . 40 years. In a landscape with higher forest connectivity, colonization success was higher but still insufficient to ensure large-scale colonization. Under the hypothesis of net extinction at southern range boundaries, forest plant species dispersal limitation will prevent net colonization at northern range boundaries required for range conservation.     

Increased Productivity and Reduced Seed Predation Favor a Large‐seeded Palm in Small Atlantic Forest Fragments

Large‐seeded plants are especially vulnerable to the loss of seed dispersers in small forest fragments. The palm Attalea humilis goes against this trend by reaching high abundances in small remnants. Productivity, seed dispersal and seed predation of A. humilis were investigated in two large (2400 and 3500 ha) and three small (19, 26 and 57 ha) Atlantic Forest fragments in southeastern Brazil. Palms in the small fragments produced more female inflorescences, resulting in a higher fruit production in these places. Seed dispersal rates were higher in the large fragments, where scatter hoarding was more frequent. Scolytine beetles were the main seed predators and damaged a larger number of seeds in small fragments, but predation by rodents and bruchine beetles was low irrespective of fragment size. As scolytines do not necessarily kill the seeds, low predation by bruchines and rodents, together with its own high productivity, allow A. humilis to be more abundant in small fragments despite the scarcity of its main dispersers. This increased abundance, by its turn, can increase competitive interactions between A. humilis and other plants in small fragments. Thus, abundance patterns of A. humilis are a good example of fragmentation affecting the balance of ecological interactions in a complex way, emphasizing the role of preserving ecological processes for conserving biodiversity in fragmented tropical landscapes. Abstract in Portuguese is available at http://www.blackwell‐synergy.com/loi/btp .     

Not seeing the ocean for the islands: the mediating influence of matrix-based processes on forest fragmentation effects

The pervasive influence of island biogeography theory on forest fragmentation research has often led to a misleading conceptualization of landscapes as areas of forest/habitat and 'non-forest/non-habitat' and an overriding focus on processes within forest remnants at the expense of research in the human-modified matrix. The matrix, however, may be neither uniformly unsuitable as habitat nor serve as a fully–absorbing barrier to the dispersal of forest taxa. In this paper, we present a conceptual model that addresses how forest habitat loss and fragmentation affect biodiversity through reduction of the resource base, subdivision of populations, alterations of species interactions and disturbance regimes, modifications of microclimate and increases in the presence of invasive species and human pressures on remnants. While we acknowledge the importance of changes associated with the forest remnants themselves (e.g. decreased forest area and increased isolation of forest patches), we stress that the extent, intensity and permanence of alterations to the matrix will have an overriding influence on area and isolation effects and emphasize the potential roles of the matrix as not only a barrier but also as habitat, source and conduit. Our intention is to argue for shifting the examination of forest fragmentation effects away from a patch-based perspective focused on factors such as patch area and distance metrics to a landscape mosaic perspective that recognizes the importance of gradients in habitat conditions.     

Ecological impacts of tropical forest fragmentation: how consistent are patterns in species richness and nestedness?

Large areas of tropical forest now exist as remnants scattered across agricultural landscapes, and so understanding the impacts of forest fragmentation is important for biodiversity conservation. We examined species richness and nestedness among tropical forest remnants in birds (meta-analysis of published studies) and insects (field data for fruit-feeding Lepidoptera (butterflies and moths) and ants). Species-area relationships were evident in all four taxa, and avian and insect assemblages in remnants typically were nested subsets of those in larger areas. Avian carnivores and nectarivores and predatory ants were more nested than other guilds, implying that the sequential loss of species was more predictable in these groups, and that fragmentation alters the trophic organization of communities. For butterflies, the ordering of fragments to achieve maximum nestedness was by fragment area, suggesting that differences among fragments were driven mainly by extinction. In contrast for moths, maximum nestedness was achieved by ordering species by wing length; species with longer wings (implying better dispersal) were more likely to occur at all sites, including low diversity sites, suggesting that differences among fragments were driven more strongly by colonization. Although all four taxa exhibited high levels of nestedness, patterns of species turnover were also idiosyncratic, and thus even species-poor sites contributed to landscape-scale biodiversity, particularly for insects.     

Additive partitioning of butterfly diversity in a fragmented landscape: importance of scale and implications for conservation

Aim Most of the Atlantic Forest in Brazil occurs in fragments of various sizes. Previous studies indicate that forest fragmentation affects fruit‐feeding butterflies. Conservation strategies that seek to preserve organisms that are distributed in high‐fragmented biomes need to understand the spatial distribution of these organisms across the landscape. In view of the importance of understanding the fauna of these forest remnants, the objective of the present work is to investigate the extent to which the diversity of this group varies across spatial scales ranging from within‐forest patches to between landscapes. Location South America, south‐eastern Brazil, São Paulo State. Methods We used bait traps to sample fruit feeding butterflies at 50 points in 10 fragments in two different landscapes during a period of 12 months. Total species richness and Shannon index were partitioned additively in diversity at trap level, and beta diversity was calculated among traps, among forest patches, and between landscapes. We used permutation tests to compare these values to the expected ones under the null hypothesis that beta diversity is only a random sampling effect. Results There was significant beta diversity at the smallest scale examined; however, the significance at higher scales depends on the diversity measurement used. Beta diversity with Shannon index was smaller than expected by chance among fragments, whereas species richness was not. Among landscapes, only beta diversity in richness was higher than expected by chance. Main conclusions The results observed occur because there is great variability in species composition among forest patches in the same landscape, changing this diversity even though the communities are formed from the same pool of species. At the largest scale evaluated (between landscapes), these pattern changes and differences in beta diversity in richness were detectable. This difference is probably caused by the presence of rare species. Thus, a conservation strategy that seeks to preserve as many species as possible per unit of area in high‐fragmented biomes should give priority to protecting fragments in different landscapes, rather than more fragments in the same landscape.     

Flying Foxes Prefer to Forage in Farmland in a Tropical Dry Forest Landscape Mosaic in Fiji

To test flying fox adaptations to a habitat mosaic with extreme deforestation, the abundance, habitat choice and feeding behavior of the Pacific flying fox, Pteropus tonganus, were investigated across 16 islands of the Yasawa archipelago, Fiji. The habitat mosaic is formed by 4.3 percent tropical dry forest and 3.3 percent farmland, leaving exotic grasslands and stands of Leucaena leucocephala to overrun the vast majority of land. Pteropus tonganus abundance was high (5757 bats) despite deforestation and hunting. Roosting sites were restricted to native forest fragments. Grasslands and stands of L. leucocephala were completely void of bats at all times. The mean foraging density in farmland was four times higher than in forests and foraging competition was routinely observed in farmland but was extremely rare in forests. The author suggests that during the study, extensive foraging in farmland was supporting the high P. tonganus population. Additionally, the preferential foraging in farmland was responsible for the low foraging densities within forests and dramatically less intraspecies competition for forest resources. Further research is needed on seed dispersal within forests and to test for seasonal variations in bat abundance and feeding.     

Does long‐distance pollen dispersal preclude inbreeding in tropical trees? Fragmentation genetics of Dysoxylum malabaricum in an agro‐forest landscape

Tropical trees often display long‐distance pollen dispersal, even in highly fragmented landscapes. Understanding how patterns of spatial isolation influence pollen dispersal and interact with background patterns of fine‐scale spatial genetic structure (FSGS) is critical for evaluating the genetic consequences of habitat fragmentation. In the endangered tropical timber tree Dysoxylum malabaricum (Meliaceae), we apply eleven microsatellite markers with paternity and parentage analysis to directly estimate historic gene flow and contemporary pollen dispersal across a large area (216 km²) in a highly fragmented agro‐forest landscape. A comparison of genetic diversity and genetic structure in adult and juvenile life stages indicates an increase in differentiation and FSGS over time. Paternity analysis and parentage analysis demonstrate high genetic connectivity across the landscape by pollen dispersal. A comparison between mother trees in forest patches with low and high densities of adult trees shows that the frequency of short‐distance mating increases, as does average kinship among mates in low‐density stands. This indicates that there are potentially negative genetic consequences of low population density associated with forest fragmentation. Single isolated trees, in contrast, frequently receive heterogeneous pollen from distances exceeding 5 km. We discuss the processes leading to the observed patterns of pollen dispersal and the implications of this for conservation management of D. malabaricum and tropical trees more generally.     

Postdispersal Seed Fate of Tropical Montane Trees in an Agricultural Landscape, Southern Costa Rica

Variation in postdispersal seed fate is an important factor driving patterns of forest regeneration. Because most previous studies have not tracked final seed fate and have commonly equated seed removal with predation without considering the possibility of secondary dispersal, little is known about individual seed mortality factors in successional and degraded habitats. This study tracked the postdispersal fate of large-seeded tropical montane trees in abandoned pasture, young secondary forests, and small forest fragments during two consecutive years in an agricultural landscape in southern Costa Rica. The incidence of secondary dispersal by animals, scatterhoarding in particular, and the effects of seed burial on germination were measured. Overall, seeds survived through germination more often in secondary forests with high levels of mortality occurring in abandoned pastures and forest fragments. The majority of seed mortality was caused by rodent predation in forest fragments, insects and fungal pathogens in secondary forests, and a combination of desiccation, insects, and fungal pathogens in pastures. Seeds were frequently secondarily dispersed in larger forest fragments, whereas they were only rarely moved in pastures and secondary forests. Burial tended to improve germination in pastures and was important for an often scatterhoarded species, Otoba novogranatensis, in all habitats. The results of this study suggest that: (1) seed mortality factors differ in response to the type of habitat degradation; (2) large-seeded species have high potential for survival when dispersed to young secondary forests; and (3) seed removal is not a reliable proxy for seed predation, particularly in less degraded forest fragments.