No description available.

This study investigated the utility to vertebrates of upland linear riparian rainforest fragments on the Atherton and Evelyn Tablelands in the Australian Wet Tropics region, north Queensland. Similar linear fragments were selected, that varied in forest age and their connectivity to large areas of continuous forest:- (connected primary (N=6), isolated primary (N=5), connected secondary (N=6) and isolated secondary (N=7)). Primary sites had either never been cleared or only subject to selective logging, while secondary forest had been completely cleared and allowed to regenerate for at least 30 years. These linear fragments were contrasted with riparian sites within continuous forest sites (N=6 to 7), which were situated in State Forest or National Parks, and sites within the cleared matrix (pasture, N=6). Vertebrates surveyed were birds, ground-dwelling mammals and reptiles, particularly leaf-litter skinks. All surveys were conducted between September and December in 2001 and/or 2000. Chapter 2 investigates the effects of forest age, isolation and structural vegetation features on bird assemblages within linear riparian fragments of rainforest. Bird surveys and structural vegetation assessments were conducted within connected and isolated primary and secondary linear fragments, and compared with those of continuous forest habitat (N=6) and pasture. There were strong effects of forest age; all three types of primary rainforest had higher values than secondary rainforest for most measured attributes of vegetation structure (including canopy height and cover; and frequency of large-diameter trees, lianes, epiphytes, strangler figs; and woody debris), but lower frequencies of tree ferns and thorny scramblers. Sites within primary rainforest also had a greater frequency of many bird species across different guilds of habitat, feeding and movement. Assemblages of rainforest-dependent birds showed an effect of isolation, although its strength was less than that of forest age. Isolated fragments of primary rainforest differed significantly from continuous primary rainforest in their rainforest-dependent bird species assemblages (and had lower species richness), and isolated fragments of secondary rainforest differed from those that were connected. There was a significant association between the species composition of rainforest birds and some measured vegetation parameters across all sites, but not within primary or secondary sites. Vegetation differences did not explain the lowered frequency of several species in isolated fragments. Limited dispersal seems unlikely to be a main cause, and causal processes probably vary among species. Specialist rainforest species endemic to the Wet Tropics region showed stronger responses to present-day rainforest age and fragmentation than those not endemic. Variation in nest depredation levels associated with rainforest fragmentation (edge effects) is examined in Chapter 3. Artificial nests were placed in the forest understorey at seven edge sites where continuous forest adjoined pasture, seven interiors (about one kilometre from the edge), and six primary linear riparian forest remnants (50-100 m wide) that were connected to continuous forest. Four nest types were compared, representing different combinations of two factors; height (ground, shrub) and shape (open, domed). At each site, four nests of each type, containing one quail egg and two model plasticine eggs, were interspersed about 15 m apart within a 160 m transect. Predators were identified from marks on the plasticine eggs. The overall depredation rate was 66.5% of 320 nests' contents damaged over a three-day period. Large rodents, especially the rat Uromys caudimaculatus, and birds, especially the spotted catbird Ailuroedus melanotis, were the main predators. Mammals comprised 56.5% and birds 31.0% of identified predators, with 12.5% of unknown identity. The depredation rate did not vary among site-types, or between open and domed nests, and there were no statistically significant interactions. Nest height strongly affected depredation rates by particular types of predator; depredation rates by mammals were highest at ground nests, whereas attacks by birds were most frequent at shrub nests. These effects counterbalanced so that overall there was little net effect of nest height. Mammals accounted for 78.4% of depredated ground nests and birds for at least 47.4% of shrub nests (and possibly up to 70.1%). The main predators were species characteristic of rainforest, rather than habitat generalists, open-country or edge specialists. For birds that nest in the tropical rainforest understorey of the study region, it is unlikely that edges and linear remnants presently function as ecological population sinks due to mortality associated with increased nest depredation. The use of linear riparian remnants by small ground-dwelling mammals and reptiles (mainly leaf litter skinks), is reported in Chapter 4. Site types were continuous rainforest, connected and isolated linear fragments of both uncleared primary rainforest and secondary regrowth rainforest. Mammals were also surveyed in pasture sites. Neither reptile species richness nor abundance varied significantly among site types. Although mammal species richness varied significantly between site types, with isolated primary sites containing highest species richness, overall mammal abundance did not differ significantly among site types. Pasture sites differed significantly from all rainforest sites in their mammal species composition, and were dominated by the introduced house mouse (Mus musculus). This species was absent from all rainforest sites, which were characterised by moderate abundances of bush rat/Cape York rat Rattus fuscipes/leucopus, fawn-footed melomys Melomys cervinipes and giant white-tailed rat Uromys caudimaculatus. None of these species varied significantly in abundance among site types, although the giant white-tailed rat showed a trend (P=0.09) for reduced abundance in isolated secondary sites. A single reptile species, the prickly forest skink Gnypetoscincus queenslandiae, occurred in sufficient numbers for individual analysis, and its abundance varied significantly among the forested site types, being less abundant in all linear fragments than in continuous forest sites. The utility of linear riparian rainforest for vertebrates appears to be species-specific and involves many factors. However, overall, species endemic to the Wet Tropics (which are hence of the highest conservation significance) appear to be the most sensitive to fragmentation. These species were most likely to show altered abundances or frequencies of occurrence due to isolation, forest age, and habitat linearity. The ecology of species within this group warrants further investigation within fragmented and non-fragmented regions of the Tablelands. For many other vertebrates examined in this study, there appears to be sufficient functional connectedness between remnants on the Tablelands to minimise the effects of fragmentation. Nevertheless, the lower density of many of these species in pasture may indicate that their long-term persistence within the fragmented rainforest areas could benefit from the maintenance or establishment of habitat linkages. Certainly, if the current rainforest vegetation cover were further reduced, or if the land use in the matrix became more intensive, the establishment of specific habitat linkages could become more important as existing dispersal routes could be lost. It also appears that nest depredation levels are unlikely to limit the value of linear rainforest remnants and other small rainforest remnants as breeding habitat for birds (at least for understorey-nesting species), relative to more intact rainforest, in the study region.

Submitted by Maria de Lourdes Mariano (lmariano@ufscar.br) on 2017-01-12T17:03:09Z No. of bitstreams: 1 VICENTIN_Alessandra_2015.pdf: 7092733 bytes, checksum: bebdbfc8311d3934becdc4804be64968 (MD5) / Approved for entry into archive by Maria de Lourdes Mariano (lmariano@ufscar.br) on 2017-01-12T17:03:22Z (GMT) No. of bitstreams: 1 VICENTIN_Alessandra_2015.pdf: 7092733 bytes, checksum: bebdbfc8311d3934becdc4804be64968 (MD5) / Approved for entry into archive by Maria de Lourdes Mariano (lmariano@ufscar.br) on 2017-01-12T17:03:29Z (GMT) No. of bitstreams: 1 VICENTIN_Alessandra_2015.pdf: 7092733 bytes, checksum: bebdbfc8311d3934becdc4804be64968 (MD5) / Made available in DSpace on 2017-01-12T17:03:37Z (GMT). No. of bitstreams: 1 VICENTIN_Alessandra_2015.pdf: 7092733 bytes, checksum: bebdbfc8311d3934becdc4804be64968 (MD5) Previous issue date: 2015-06-19 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) / The Atlantic Forest is considered hotspot for biodiversity conservation, due to its high biological diversity, high endemism and high threat of degradation. This scenario has raised concern about finding information to support recovery projects of degraded areas. This study aimed to evaluate the growth, coverage percentage in the soil and the survival of species of the Atlantic Forest, from the planting seedlings in different treatments, present in a particular area in the recovery process, in the region Vale do Ribeira, located in Sete Barras, Brazil. In 2004, the owner made the removal of vegetation for grazing deployment with brachiaria. In 2009, UFSCar was sought to implement the project of "experimental recovery of degraded areas" where seedlings were planted, divided into spacing of 1.5 x 2 m (dense treatment) and 3 x 2 m (conventional treatment). In this study, from July 2013 was carried out monitoring of the area to assess the growth and survival of planted seedlings as well as the percentage of shading and grasses. Five surveys were conducted (July 2013, October 2013, January 2014, April 2014 and July 2014) to evaluate the survival and growth (height, stem diameter and crown diameter) of the seedlings. 30 months after planting, the 220 individuals initially raised in this study, the survival was 85.2% in the dense treatment and 84.8% in the conventional treatment, so mortality was around 15% for both treatments. The species with the highest average growth in height were Croton floribundus, Cytharexyllum solanaceum, Casearia gossypiosperma, Croton urucurana, Colubrina glandulosa, Lafoensia pacari, Inga marginata e Myrsine guianensis. The species with the highest average growth in stem diameter were Acnistus arborescens, Croton urucurana, Apeiba tibourbou, Croton floribundus, Inga marginata, Campomanesia xanthocarpa, Colubrina glandulosa and Lafoensia pacari. The species with the highest average growth in crown diameter were Croton urucurana, Croton floribundus, Cytharexyllum solanaceum, Acacia polyphyla, Colubrina glandulosa, Inga marginata, Lafoensia pacari and Rhamnidium elaeocarpum. As for the percentage coverage in the soil and grasses, there was significant difference in the dense and conventional treatment compared to witness. It is concluded that the above species are indicated for use in future restoration projects in areas of Rain Forest. Both treatments showed greatest shading (in comparison with witness) at the end of the monitoring period, thereby enabling a reduction of exotic grasses. However, in dense treatment was no shading and greater control of grasses. Thus, it can be concluded that at higher cost of implementation and the similar results, the dense treatment was not recommended thickened in the region. / O bioma Mata Atlântica é considerado um hotspot para conservação da biodiversidade, por apresentar grande diversidade biológica, alto grau de endemismo e elevada ameaça de degradação. Este cenário tem despertado preocupação com relação à busca de informações para subsidiar projetos de recuperação das áreas degradadas. Este trabalho teve como objetivo avaliar o crescimento, porcentagem de cobertura no solo e a sobrevivência de espécies da Mata Atlântica, provenientes do plantio de mudas em diferentes tratamentos, presentes em uma área particular em processo de recuperação, na região Vale do Ribeira, Sete Barras-SP. No ano de 2004, o proprietário fez a supressão da vegetação para implantação de pasto com braquiária. No ano de 2009, a UFSCar foi procurada para implantar o projeto de “Recuperação experimental de áreas degradadas”, onde foram plantadas mudas, distribuídas em espaçamentos de 1,5x 2 m (tratamento adensado) e 3 x 2 m (tratamento convencional). Neste estudo, a partir de julho de 2013 foi realizado o monitoramento da área para avaliar o crescimento e a sobrevivência das mudas plantadas assim como a porcentagem de sombreamento e gramíneas. Foram realizados cinco levantamentos (julho de 2013, outubro de 2013, janeiro de 2014, abril de 2014 e julho de 2014) para avaliar a sobrevivência e o crescimento (altura, diâmetro do colo e diâmetro da copa) das mudas. Após 30 meses do plantio, dos 220 indivíduos levantados inicialmente neste estudo, a sobrevivência foi de 85,2% no tratamento adensado e 84,8% no tratamento convencional, ou seja, a mortalidade ficou em torno de 15% para ambos os tratamentos. As espécies que apresentaram as maiores médias de crescimento em altura foram Croton floribundus, Cytharexyllum solanaceum, Casearia gossypiosperma, Croton urucurana, Colubrina glandulosa, Lafoensia pacari, Inga marginata e Myrsine guianensis. As espécies que apresentaram maior crescimento em diâmetro do colo foram Acnistus arborescens, Croton urucurana, Apeiba tibourbou, Croton floribundus, Inga marginata, Campomanesia xanthocarpa, Colubrina glandulosa e Lafoensia pacari. As espécies que apresentaram maior crescimento em diâmetro da copa foram Croton urucurana, Croton floribundus, Cytharexyllum solanaceum, Acacia polyphyla, Colubrina glandulosa, Inga marginata, Lafoensia pacari e Rhamnidium elaeocarpum. Quanto à porcentagem de cobertura no solo e gramíneas, houve diferença significativa no tratamento adensado e convencional em comparação com testemunho. Conclui-se que as espécies acima são indicadas para serem utilizadas em futuros projetos de restauração em área de Floresta Ombrófila Densa. Ambos os tratamentos apresentaram maior sombreamento (em comparação com testemunho) no final do período de monitoramento, possibilitando assim a diminuição das gramíneas exóticas. No entanto, no tratamento adensado não houve maior sombreamento e controle de gramíneas. Dessa forma, pode-se concluir que pelo maior custo de implantação e pelos resultados similares, o tratamento adensado não seria recomendado na região.

Despite ecological, anthropological, and archaeological debate surrounding their desirability as habitats for human occupation, tropical rainforests have received relatively little attention in discussions of Homo sapiens' Pleistocene dispersal. Sri Lanka has yielded some of the earliest dated fossil and material culture evidence (c. 38-35,000 cal. years BP) for our species in a modern rainforest context beyond Africa. Nevertheless, assertions in Sri Lanka, and elsewhere, regarding early human rainforest reliance have been largely based on coarse or 'off-site' palaeoenvironmental records, and the overall role of these environments in human subsistence strategies has remained uncertain. This study applies stable carbon and oxygen isotope analysis to Sri Lankan human fossil, and associated faunal, tooth enamel dated to between 36-29,000 and 3,000 cal. years BP, in order to directly test human rainforest resource reliance, reconstruct a stable isotope ecology, and develop 'on-site' palaeoenvironmental records for Late Pleistocene-Holocene Sri Lankan rainforest foragers. Stable carbon and oxygen isotope analysis of modern Sri Lankan primates, and stable carbon, oxygen, and nitrogen isotope analysis of modern plant samples from the Polonnaruwa Nature Sanctuary, are also performed to investigate the ecology of Sri Lankan primates on which Late Pleistocene-Holocene forager subsistence strategies were focused. The results demonstrate that Homo sapiens relied on rainforest resources in Sri Lanka from c. 36-29,000 cal. years BP until the Iron Age c. 3 cal. years BP, even when open environments, and their corresponding resources, were available. This remains the case through periods of evident environmental change at the Last Glacial Maximum and even upon the arrival of agriculture in the island's tropical forests. The primate stable isotope data prove difficult to interpret as ecological niche separation in the absence of observation data. Nonetheless, humans were evidently able to not only use but also rapidly specialise in the exploitation of South Asia's rainforests.

Expanding anthropogenic development within the tropical forest biome is driving the loss of an irreplaceable global resource. Mega-diverse tropical forests are vital for regulating the global carbon cycle, and are essential for climate change mitigation. Today, over half of the world’s remaining tropical forest is degraded or regenerating secondary forest. Tropical forests are becoming increasingly fragmented through the expansion of agriculture and roads. Landscape-scale flooding of terrestrial habitats caused by dam construction is an emerging driver of habitat loss and fragmentation. Much attention has been paid to the long-term impacts of tropical forest fragmentation for biodiversity, ecosystem functioning, and carbon emissions. Most of our understanding of the impacts associated with habitat fragmentation originates from systems in which the habitat matrix surrounding remnant forest patches is another, albeit low quality, terrestrial habitat. However, dam-induced habitat fragmentation results in remnant terrestrial biological communities becoming isolated on islands within a water matrix. A water matrix presents the worst-case scenario for remnant habitat fragments. In Chapter 2 I synthesise the results of numerous studies reporting the responses of taxonomic groups to isolation on reservoir land-bridge islands, and uncover a globally-applicable pattern of extinction debt acting upon remnant biological communities on reservoir islands. All islands, regardless of taxonomic group, habitat type, or island area lose species as island isolation time increases. Moreover, I show that contrary to existing ecological theory, once terrestrial habitat becomes isolated within a water matrix, it is effectively too isolated for species losses to be buffered by metapopulation dynamics. Dam development is rapidly expanding in the largest remaining tract of intact tropical forest, the Amazon Basin. In Chapters 3 and 4 I study the Balbina mega-dam system in the central Brazilian Amazon. Here, I use detailed field inventories of trees and lianas on islands and in continuous mainland habitat to determine the impact of landscape-scale habitat fragmentation caused by reservoir creation on these taxonomic groups. I find that islands maintain tree communities at significantly lower densities, richness and diversity compared to continuous forest. Furthermore, tree communities on islands exhibit compositional divergence from those found in mainland continuous forest. Island tree assemblages are dominated by low-wood density species, and may be on a trajectory towards communities characteristic of early successional forests with reduced carbon storage capacity. In contrast, liana assemblages remain compositionally intact and are becoming increasingly dominant relative to trees. Thus, lianas appear robust to many of the negative impacts associated with landscape-scale habitat fragmentation. As insular tree communities continue to degrade through area- and edge-effects, lianas may become a key feature of this archipelagic landscape due to their competitive advantage over trees in disturbed forest habitats. Lianas significantly inhibit tree recruitment and carbon storage. Thus, findings from Chapters 3 and 4 provide strong evidence for additional, and currently unaccounted-for biodiversity and carbon impacts associated with tropical dams. As development of tropical forest regions increases, there is an urgent need to reconcile the need for resources with the need for ecosystem service provision, such as carbon storage, particularly as we attempt to mitigate the impacts of rising atmospheric carbon. Recent studies have shown that secondary tropical forests have the potential to rapidly uptake atmospheric carbon, and act as a powerful tool in climate change mitigation policy. Broad-scale estimates of secondary forest carbon uptake are currently based on above-ground biomass alone. In Chapter 5 I present carbon stock estimates of additional tropical forest carbon pools - soil and dead woody biomass - in secondary forests ranging from 40-120 years. I find that soil fertility (nitrogen concentration) is key in determining carbon storage in secondary forests, and that the stability of carbon stocks held in dead woody biomass increases with secondary forest stand age. I highlight the need to integrate detailed site-specific information into broad-scale predictive models of secondary tropical forest carbon sequestration. This thesis links ecological theory and landscape-scale field inventories, to provide new understanding of the long-term costs of tropical forest fragmentation for biodiversity conservation and carbon storage, and provides further evidence of the important role secondary tropical forests may play in carbon sequestration and climate change mitigation.

In the Pacific Northwest (PNW), concern about the impacts of climate and land cover change on water resources, flood-generating processes, and ecosystem dynamics emphasize the need for a mechanistic understanding of the interactions between forest canopies and hydrological processes. A detailed measurement and modeling program during the 1999 and 2000 hydrologic years characterized hydrological conditions and processes in a 500-600 year old Douglas fir-western hemlock seasonal temperate rainforest. The measurement program included sub-canopy arrays of radiometers, tipping bucket rain gauges, and soil temperature and moisture probes, to supplement a vertical temperature and humidity profile within the forest canopy. Analysis of the precipitation interception characteristics of the canopy indicated that the mean direct throughfall proportion was 0.36, and the mean saturation storage was 3.3 mm. Evaporation from small storms insufficient to saturate the canopy comprised 19% of the net interception loss, and canopy drying and evaporation during rainfall accounted for 47% and 33% of the net loss, respectively. Results of the measurement program were used to modify the Simultaneous Heat and Water (SHAW) model for forested systems. Changes to the model include improved representation of interception dynamics, stomatal conductance, and within-canopy energy transfer processes. The model effectively simulated canopy air and vapor density profiles, snowcover processes, throughfall, soil water content profiles, shallow soil temperatures, and transpiration fluxes for both a calibration period and for an uncalibrated year. Soil warming at bare locations was delayed until most of the snowcover ablated due to the large heat sink associated with the residual snow patches. During the summer, simulated evapotranspiration decreased from a maximum monthly mean of 2.17 mm day����� in July to 1.34 mm day����� in September, as a result of declining soil moisture and net radiation. Our results indicate that a relatively simple parameterization of the SHAW model for the vegetation canopy can accurately simulate seasonal hydrologic fluxes in this environment. Application and validation of the model in other forest systems will establish similarities and differences in the interactions of vegetation and hydrology, and assess the sensitivity of other systems to natural and anthropogenic perturbations. / Graduation date: 2002

This study reports forest structure, regeneration and the soil properties from unlogged and logged forest in the Bulungan Research Forest, Malinau District, East Kalimantan, Indonesia. Four sites were compared by using four 1-ha replicate plots in each of primary forest (PF), 5, 10 and 30-yr old logged forest (LF-5, LF-10, LF- 30). The tree species composition differ among forest types, as it was shown that the mean value of similarity indices for all pairs were 0.215 (for the Jaccard index) and 0.353 (for the Sorensen index). The low values for similarities among forest types were most probably caused by low numbers of species shared between each forest type. Both correlation values, r = 0.023 for Jaccard index and r = 0.031 for Sorensen index, showed no strong correlation between the similarity index (C) and the distance between forest types. This supports the use of a chronosequence approach. A total of 914 tree species with ³ 10 cm dbh were recorded from 223 genera and 65 families. There were no significant differences in mean species numbers (166 – 180/ha) among treatments. Mean density of species was lower in LF-5 and LF-10 (501/ha) than in PF or LF-30 (605/ha and 577/ha); similarly to mean basal area (LF-5, 28.5 m2/ha; LF-10, 32.6 m2/ha) vs. PF (45.8 m2/ha) and LF-30 (46.9 m2/ha). Dead wood on the forest floor was significantly higher in LF-10 (75 m3/ha) than in the other treatments. Seedlings (< 2 cm dbh) of 1,022 species were recorded from 408 genera and 111 families. The mean number of tree seedling species ranged between 170-206; the mean density of seedlings was about two-fold lower in LF-10 (2790/ha) than in the other treatments. Saplings (>2 – 9.9 cm dbh) of 802 species belonged to 241 genera and 65 families. There was a high variability in species richness across treatments (89 – 191/ha), but not in stem numbers. The Dipterocarpaceae family was dominant in all treatments, followed by the Euphorbiaceae. The soils were acidic, low in nutrients and had low to very low fertility. Both primary and logged forest areas are marginal or not suitable for sustained production of plantation crops. Logging caused soil compaction in LF-30. Although in terms of number of species and trees, amount of BA, number of saplings and seedlings LF-30 appeared to have satisfied prescriptions for a second harvest, ecologically the forest is far from mature. The Indonesian Selective Cutting and Replanting (TPTI) system may need to be revised to a 35 – 45 year cycle to ensure long-term forest productivity in terms of not only timber but other goods and ecosystem services, the value of which are never quantified in monetary terms, but can be higher than the timber revenue.

One of the primary challenges facing researchers and practitioners in their efforts to address issues of poverty and environment is the need to deepen our understanding of the logic that guides local people's decisions over resource use, particularly among the rural poor whose livelihoods depend on fragile and dynamic environments. This study seeks to identify the set of factors that influences how rainforest people respond to abrupt natural disturbances and resource scarcity through changes in livelihood and resource management practices in two rural poor communities of the Peruvian Amazon. Data were gathered through in-depth survey interviews (n=95 households) between June and December 2003 in the Amerindian community of Arica Viejo (Ucayali River) and the mestizo (ribereno) community of Roca Fuerte (Maranon River). The results reveal that socioeconomic characteristics such as forest experience and knowledge, and access to agricultural land explain striking differences among households in livelihood responses to environmental change, particularly concerning resource use behavior, resilience to disturbance, and the propensity to adopt sustainable resource management strategies.

No description available.

A growing body of work reveals that animal-mediated pollination is negatively affected by anthropogenic disturbance. Landscape-scale disturbance results in two often inter-related processes: (1) habitat loss, and (2) disruptions of habitat configuration (i.e. fragmentation). Understanding the relative effects of such processes is critical in designing effective management strategies to limit pollination and pollinator decline. I reviewed existing published work from 1989 to 2009 and found that only six of 303 studies separated the effects of habitat loss from fragmentation. I provide a synthesis of the current landscape, behavioral, and pollination ecology literature in order to present preliminary multiple working hypotheses to explain how these two landscape processes might independently influence pollination dynamics (Chapter 2). Despite the potential importance of independent effects of habitat fragmentation, effects on pollination remain largely untested. Studies designed to disentangle the independent effects of habitat loss and fragmentation are essential for gaining insight into landscape-mediated pollination declines. I also found that the field of landscape pollination ecology could benefit from quantification of the matrix, landscape functional connectivity, and pollinator movement behavior. To test the hypothesis that pollinator movement can be influenced by landscape configuration, I translocated radio-tagged hummingbirds across agricultural and forested landscapes near Las Cruces, Costa Rica (Chapter 3). I found return paths were on average more direct in forested than in agricultural landscapes. In addition, movement paths chosen in agricultural landscapes were more forested than the most direct route suggesting that hummingbirds avoided crossing open areas when possible. To determine if differences in pollinator movement translated to differences in plant reproduction, I tested the relative importance of landscape composition versus configuration on the reproductive success of Heliconia tortuosa, a hummingbird-pollinated forest herb (Chapter 4). I used a stratified random sampling design to select sites across orthogonal gradients in patch size, amount of forest, and elevation. I tested four landscape change hypotheses (i.e., local, landscape composition, landscape fragmentation, and fragmentation threshold). I found that Heliconia reproduction supported both the local site and landscape fragmentation hypotheses. Seed set increased with increasing forest patch size independent of amount of forest in the surrounding landscape. I also found that increasing patch size positively influenced the relative abundance of pollinators. The observed differences in seed set likely resulted from differences in hummingbird movements (Chapter 3) and/or abundance under different landscape configurations. / Graduation date: 2013