The Panama Canal Watershed (PCW) represents a special opportunity for studies related to seed dormancy in the tropics with both applied and basic research objectives. There is a clear need for seed information for nursery planning in restoration projects that involves use of a large number of local species. Moreover, the strong rainfall gradient along the 60 km of continuous lowland forest in the PCW represents an excellent opportunity to understand dormancy and its role as an adaptive trait that evolved in response to environmental factors. This study presents useful seed information for 100 tree species native to the PCW. For each species, it includes collection system, fruiting time, seed mass, seed moisture content, germination, and longevity in storage at 20??C. For the first time, an attempt is made to classify (to class sensu Baskin andamp; Baskin 2004) the class of dormancy in seeds of the PCW. An analysis of the relationship between class of dormancy and seed mass, moisture content, longevity, germination patterns, and seasonality is presented. The relationship between class of seed dormancy and longevity and geographical distribution of species within the watershed based on the rainfall gradient was investigated.

Plants depend on a wide variety of vectors, both biotic and abiotic, to move their seeds to locations away from parent plants. The stages between seed production and seedling establishment in the plant life cycle are under the influence of a complex set of ecological and evolutionary factors (Wang and Smith 2002). In ecological time scales, seed dispersal has been shown to influence both plant population and community dynamics (Howe and Miriti 2004, Seidler and Plotkin 2006). The evolutionary effects are seen in the way these interactions influence the fruiting season, fruit crop size, fruit size, nutritive value of fruit or seed and other strategies adopted by the plant to ensure effective dispersal of their seeds (Howe and Smallwood 1982, Schupp 1993). Most seed dispersal research has focussed on arboreal frugivores such as birds, primates and bats. In comparison, terrestrial frugivores such as ruminants, elephants and rodents are poorly researched. Large-bodied terrestrial frugivores are key to the dispersal of several large-seeded plants and are also capable of dispersing seeds to long distances (> 1 km). Given that large mammals have witnessed tremendous declines in their abundances and ranges in historical times, plants that depend upon them for long distance dispersal might have limited ability to move across fragmented landscapes under changing climates. However, very little is understood about seed dispersal by large mammals or about seed dispersal characteristics of tropical dry forests which support among the highest biomass of large terrestrial mammals. Tropical dry forests are among the most endangered tropical ecosystems (Janzen 1988). It has also been forecast that tropical dry forests in the Indian subcontinent will witness significant changes in temperature and rainfall regimes in the coming decades (Ravindranath et al. 2006). However, compared to tropical moist forests which have received most of the scientific interest, theory, methodology and base line information about seed dispersal in dry forests is poorly developed. To address some of these lacunae, I examined seed dispersal at both population and community-scales in the tropical dry forests of Mudumalai Wildlife Sanctuary from 2005 -2007. Mudumlai is located within a relatively large contiguous stretch of tropical dry forests in southern India and has an intact and abundant frugivore assemblage. In addition, the Mudumalai Forest Dynamics Plot (MFDP), a large 50 hectare plot has been monitored for recruitment, mortality and diameter growth of woody plants for the last twenty years. The plot is an invaluable resource for seed dispersal studies since it provides us an opportunity to link seed dispersal with plant population and community dynamics. In the first part of the thesis (Chapters 3 and 4), I focus on evaluating the effectiveness of ruminants as seed dispersers and then examine factors that influence frugivory by ruminants. This work was carried out at the population level, for the tree Phyllanthus emblica. The fruits of P. emblica are important non-timber forest produce from Asian dry tropics and understanding the quantity of fruit consumed by frugivores has implications for sustainable harvest practices of this resource (Shahabuddin and Prasad 2004). In the second part of the thesis (Chapters 5 and 6), I characterized seed dispersal at the community-scale for woody plants on the MFDP. Community-wide characterization of seed dispersal at sites harboring an intact faunal assemblage is critical to identify plants dependent on endangered dispersers. The second half of the thesis focuses on examining the associations between dispersal mode and fruit characteristics, and also between dispersal mode and spatial patterning of woody plants. Evaluating the effectiveness of ruminants as seed dispersers: Disperser effectiveness is defined as the contribution a disperser makes to the future reproduction of a plant. Evaluating disperser effectiveness involves addressing both quantitative aspects such as frequency of visits, quantity of fruit consumed, diversity of species dispersed, and, qualitative components such as the treatment given in mouth and gut and the suitability of seed deposition sites for germination (Schupp 1993, Dennis and Westcott 2006). Information on quantity and quality of seed dispersal services provided by ruminants is very limited. This is because techniques such as tree watches or fruit traps adapted from avian frugivore studies are ill suited to study terrestrial frugivores, and conventional camera traps provide little quantitative information on the quantity of fruit consumed by frugivores. The quantity of dispersal services provided by ruminants for P. emblica was assessed using a novel camera trap technique. Time-delay was programmed into the camera trap units (4 pictures taken 2 minutes apart) to tell apart frugivores from visitors (by comparing number of fruit remaining in the time-delay picture sequence). During a two year study using this technique, six terrestrial mammals were identified as frugivores of P. emblica. Additionally, seven mammals and one bird species visited fruiting trees but did not consume fallen fruit. Two ruminants, the Indian chevrotain Moschiola indica and chital Axis axis, were the most frequent frugivores of P. emblica and these ruminants accounted for over 95% of fruit removal, while murid rodents accounted for less than 1%. The quality of dispersal services provided by ruminants was examined using a combination of feeding trials (with captive animals) and germination experiments for large-seeded species dispersed by ruminants in tropical dry forests, including P. emblica. Ruminants regurgitated well-protected seeds larger than 5 mm, after retaining them in the rumen for 2–50 h. Though germination success was lowered after rumen retention, it was not too low (> 15%), and would still ensure dispersal if suitable micro-habitats could be found. Given the daily home-range of chital (14–20 ha in tropical dry forests (Mishra 1982)), and their seed retention times (2–50 h), movement of chital during the 1–2 d period when seeds remain in the rumen can potentially disperse seeds to distances over 1 km. Thus, our research shows that ruminants account for most of the fruit removal and are the principal primary dispersers of P. emblica. Factors influencing frugivory by ruminants: Fruit removal, the quantity of fruits consumed by frugivores, is highly variable within plant populations. Understanding factors influencing this selection operated by frugivores helps link frugivore behavior with plant demography. The outcome of such selection operated by frugivores has been examined mostly for interactions involving avian frugivores. The extension of this theory to terrestrial frugivores, which are presented with fruit on the ground, is not well understood. In addition, factors influencing fruit removal have been examined mostly with respect to changing spatial contexts such as crop size and fruiting neighborhoods. The influence of varying temporal contexts on frugivore choice, and their possible demographic and evolutionary consequences is poorly understood. We examined if temporal variation in fruit availability across a hierarchy of nested temporal levels (interannual, intraseasonal, 120 h, 24 h) altered fruit removal of P. emblica on the MFDP. The interactions between P. emblica and its primary disperser (ruminants) was mediated by another frugivore (a primate, langur), which made large quantities of fruit available on the ground to ruminants. The direction and strength of crop size and neighborhood effects on this interaction varied with changing temporal contexts. Fruit availability was higher in the first of the two study years, and at the start of the season in both years. Fruit persistence on trees, determined by primate foraging, was influenced by crop size and conspecific neighborhood densities only in the high fruit availability year. Fruit removal by ruminants was influenced by crop size in both years and neighborhood densities only in the high availability year. In both years, these effects were stronger at the start of the season. Intraseasonal reduction in fruit availability diminished inequalities in fruit removal by ruminants and the influence of crop size and fruiting neighborhoods. Temporal asymmetry in frugivore-mediated selection could reduce the potential for co-evolution between frugivores and plants by diluting selective pressures. Langurs determined the length of time fruit persisted on individual trees, and, also influenced the quantity and duration of fruit availability for the primary dispersers, ruminants. Such inter-dependencies formed between disparate animal consumers add additional levels of complexity to plant–frugivore mutualistic networks and can have potential reproductive consequences for specific individuals within populations. Seed dispersal within a tropical dry forest community: Megafauna are considered to be important seed dispersers, but their contribution has mostly been examined in isolation, with limited reference to other frugivores within communities. Community-wide characterization of seed dispersal at sites harboring an intact fauna is essential to evaluate the role of megafauna in dispersal, examine dispersal syndromes and assess the vulnerability of plant communities to disperser declines and climate change. Dispersal modes were inventoried for 92% of species and fruit traits for 84% of the woody plant community on the MFDP. Fruit–frugivore interactions were identified using a combination of camera traps, tree watches and opportunistic observations. Though 60% of the species were animal-dispersed, mechanically-dispersed species were more dominant and constituted nearly 70% of the stand. Bird-dispersed species were very rare and constituted less than 10% of the stand cumulatively. In comparison, bird-dispersed species account for ~ 50% of the stand in moist forests of tropical Asia. Three broad dispersal modes were identified – mechanical, mammal and bird-mammal, and these displayed strong associations with fruit traits. Among the 11 frequently-observed frugivore groups, the three largest frugivores, deer, bear and elephant, together dispersed 40% of the community. There was low overlap in fruit diets between these three megafaunal groups, though diets of deer and bear overlapped with different groups of small and medium-sized frugivores. Deer and bear displayed significant associations with fruit traits, while elephant did not. Among the fleshy fruits, large fruits and large-seeded fruits had fewer dispersers compared to smaller, multi-seeded fruits. Although the largest fleshy fruits and seeds were often dispersed by megafauna, few were exclusively dependent on megafauna. The three megafaunal groups were the only potential long distance dispersers for 25% of the community. For a better understanding of the impacts of megafaunal extinctions, even rare events of dispersal by megafauna should be considered since these could prove important for plants lacking other long distance dispersers. Given that these megafauna and several medium-sized frugivores are rare in degraded areas of tropical Asia (Corlett 2007, Karanth et al. 2010), one fourth of the plant community at Mudumalai could have limited ability to disperse across fragmented landscapes in response to changing climates. Seed dispersal and spatial patterns in tropical dry forests: To address the relevance of seed dispersal to plant community structure, the linkages between dispersal mode and spatial aggregation of trees were examined for the woody plant community on the MFDP. For plants, seed arrival patterns vary depending upon their dispersal mode and form the primary matrix upon which further forces shape spatial patterns. Spatial patterns were examined via the pair correlation function (statistic) and statistics derived from cluster models (Matern process). Considering only trees that had 10 individuals (48 species), it was found that tropical dry forest trees exhibited extensive spatial aggregation. The degree of spatial aggregation at Mudumalai was remarkably similar to wet forests in tropical Asia. Species with limited-dispersal (mechanically-dispersed species) were more aggregated than those with more frequent long-distance dispersal (animal-dispersed species) at smaller spatial scales (< 50 m). Animal-dispersed species had larger cluster radius, with bird-dispersed species having the largest radii. These results are in concordance with Siedler & Plotkin (2006) and Li et al. (2009) who find that mechanically-dispersed species are more aggregated than animal-dispersed species. These results demonstrate that seed dispersal has a strong influence on the spatial patterning of plant communities. Conclusion: At the population level, our results imply that P. emblica is mainly dependent on ruminants for seed dispersal. In tropical dry forest sites such as Mudumalai, up to 15% of the species are dispersed by ruminants, with langurs mediating the ruminant-plant interaction similar to P. emblica. The observed fruit removal patterns for P. emblica reiterate the fact that all fruiting individuals in a population are not equally attractive to frugivores, especially when fruits are abundant. Alterations of such complex fruit-frugivore interactions through human-mediated changes of fruit or frugivore densities (through fruit harvests or hunting of large mammalian frugivores such as ruminants), can feed back into the seed dispersal loop and affect several downstream ecological and evolutionary processes. At the community-scale, the dispersal profile of woody plants in Mudumalai’s tropical dry forests differed considerably from tropical moist sites, with a greater representation of mammal and mechanically-dispersed species. However, the linkages between seed dispersal and spatial patterns of trees at Mudumalai are similar to tropical moist forests. Extensions of the models developed from this site harbouring an intact faunal assemblage to other tropical dry forests would enable rapid characterization of dispersal profiles and identification of plant species dependent on endangered fauna for seed dispersal that should be prioritized for restoration programs.

Studies of seed dynamics and seedling establishment at the community level in savanna ecosystems are scant, particularly in the tropical savannas of Central Brazil. The Brazilian savannas (cerrado) have the largest diversity of plant species among Neotropical savannas. Cerrado vegetation exhibits consistent changes in tree density and tree size along shallow topographic gradients. Vegetation types differ from closed savannas with high tree density, and a nearly continuous canopy cover in the uppermost portions, to open savannas with scattered short trees in the low portions of the topographic gradients. Whether or not dynamics of seeds and seedlings of woody species are consistent with variations in tree density and tree canopy cover across the cerrado landscape has not been determined, but could potentially influence such spatial patterns of tree density and canopy cover. The main objective of this study was to evaluate seed dynamics (i.e., seed rain, soil seed banks, and seed characteristics) and seedling establishment of woody species in three major cerrado vegetation types (closed, intermediate and open savannas) which differ in tree canopy cover and tree height. These vegetation types are located along topographic gradients of approximately 30 m in elevation and 1000 m in length. I conducted field work from June 2005 to August 2008 at the Instituto Brasileiro de Geografia e Estatística (IBGE) reserve, a field experimental station located 35 Km south of Brasilia, Brazil (15˚ 56' S, 47˚ 63' W, altitude 1100 m). I established permanent plots in each of the three vegetation types to assess seed rain, seed limitation, seed predation, and seed removal. I also quantified density and composition of the soil seed bank in each vegetation type and studied seed characteristics (i.e., kind of dormancy, moisture content, longevity) of 14 common cerrado woody species. In the field, I quantified establishment of woody seedlings in fire-protected savannas as well as in accidentally burned and frequently burned savannas. In the greenhouse I performed experiments to evaluate the effect of litter cover and light level on seedling emergence of 9 common cerrado woody species. Finally, I evaluated the effect of pulses of heat on seed germination of 5 common cerrado woody species under controlled conditions. The overall results of this study show that regardless of tree canopy cover, cerrado vegetation types are limited in seeds and seedlings of woody species, but these limitations are higher in open than in closed savannas. Most woody species do not form persistent soil seed banks along cerrado vegetation types because their seeds are short-lived, predated and quickly removed from the ground. In addition, dispersal season and dormancy appear to control timing of germination at the onset of the rainy season to ensure survival of seedlings in the field. In the absence of fire, establishment of woody seedlings is higher in closed than in open savannas because canopy cover mitigates the stressful environmental conditions of open savannas and thus facilitates establishment of woody seedlings. As a result, closed savannas are likely to maintain higher density of woody elements over time than open savannas. In burned savannas, however, species composition of woody and herbaceous seedlings is significantly affected because establishment of seedlings of woody and herbaceous species decreases but vegetative reproduction increases. This dissertation study indicates that interactions of early life-history stages of woody species with environmental and disturbance factors appear to play a more significant role in maintaining woody cover variations along cerrado topographic gradients than previously recognized.

Este estudo teve como objetivo determinar o efeito de distúrbios na forma de práticas de manejo sobre a dinâmica do banco de sementes do solo (BSS) e da vegetação residente em uma pastagem nativa do sul do Brasil, bem como do seu potencial de evitar que Eragrostis plana Nees (capim-annoni-2) nele se estabeleça. O experimento foi conduzido na Estância Upacaraí, município de Dom Pedrito, Estado do Rio Grande do Sul, Brasil, em área sem a presença do capim-annoni-2, cujo solo é classificado como Chernosolo Argilúvico Órtico Vértico, unidade de mapeamento Ponche Verde. As coordenadas geográficas centrais da área experimental são 31º09’38”S e 54º57’29”W, em altitude de 140 m. De acordo com a classificação de Köppen, o clima é Cfa. O estudo discute nos Capítulos 2, 3 e 4 os efeitos dos seguintes tratamentos: 1) regime de manejo da pastagem: exclusão (X) e pastejos rotativo (R) e contínuo (C); 2) níveis iniciais de distúrbio da pastagem: pastagem alta – 10-12 cm (A), pastagem baixa - < 5 cm (B) e pastagem baixa com exposição do solo (E); 3) regime de adubação: sem adubo (S), fósforo (P) e nitrogênio (N). O delineamento experimental foi em parcelas sub-subdivididas em blocos completos, compreendendo três repetições. Os tratamentos regime de manejo, níveis iniciais de distúrbio do solo e regime de adubação foram alocados nas parcelas, sub-parcelas e sub-subparcelas, respectivamente. Em 54 subsubparcelas (sub-parcelas B e E), semeou-se três espécies cultivadas de inverno (Trifolium repens L. (trevo-branco), Lotus corniculatus L. (cornichão São Gabriel) e Lolium multiflorum Lam. (azevém-anual)) e E. plana (capim-annoni-2). Nas demais 27 sub-subparcelas (subparcelas A), semeou-se somente capim-annoni-2. As amostras do BSS (Capítulo 2) foram coletadas em três datas: 4-3-2004, antes da aplicação dos tratamentos, e em 22-4-2005 e 6-5- 2006, após a aplicação dos mesmos. Nas três datas avaliadas, foram registradas 33 famílias e 137 espécies no BSS. Trevo-branco foi registrado apenas no BSS2, cornichão São Gabriel foi registrado nos três BSSs e azevém-anual e capim-annoni-2 foram registrados nos BSS2 e BSS3. Dentre as 137 espécies, 17 eram alóctones e 39 ruderais. As ruderais Centunculus minimus, Anagallis arvensis e Gamochaeta spicata foram registradas em todas as unidades amostrais dos três BSSs. O regime de manejo da pastagem e níveis iniciais de distúrbio do solo influenciaram a composição florística e o tamanho do banco de sementes do solo. Em geral, os maiores BSS de capim-annoni-2 registrados na terceira avaliação (BSS3), estiveram associados com as condições ambientais que preservaram a viabilidade das sementes, tais como os regimes de manejo exclusão e pastejo rotativo. Ao contrário, os bancos menores estão associados com o regime de pastejo contínuo, no qual as sementes foram expostas a uma condição ambiental que favoreceu a germinação mas também sua deterioração. Os levantamentos florísticos (Capítulo 3) foram realizados em sete épocas: outono-2004, antes da aplicação dos tratamentos, primavera-2004, outonos e primaveras de 2005 e 2006, e outono- 2007. Foram registradas 27 famílias e 141 espécies residentes + 4 alóctones semeadas. Das 145 espécies, 36 eram ruderais. O regime de manejo do campo e os níveis iniciais de distúrbio do solo influenciaram a composição florística da comunidade. A trajetória da composição florística das comunidades submetidas ao nível inicial de distúrbio pastagem baixa com exposição de solo é distinta daquela registrada nos níveis iniciais de distúrbio do solo pastagem alta e baixa. A exclusão reduziu a riqueza de espécies. No Capítulo 4, os níveis iniciais de distúrbio do solo e o regime de manejo influenciaram o estabelecimento do capimannoni- 2. O estabelecimento desta espécie foi favorecido pela intensidade do distúrbio do solo e menor altura da pastagem decorrente do regime pastejo contínuo. Ao contrário, a pastagem alta, associada ao pastejo rotativo, evitou o estabelecimento do capim-annoni-2. / The objective of this study was to evaluate the effects of disturbance imposed by management practices on the dynamics of the soil seed bank (SSB) in a natural grassland in Southern Brazil, as well as its potential in avoiding the invasion of Eragrostis plana Nees (capim-annoni-2). The experiment was carried out at Upacaraí Farm, in Dom Pedrito town, Rio Grande do Sul State, Brazil, in an area free of capim-annoni-2, in a Chernosolo Argilúvico Órtico Vértico soil type, Ponche Verde unit. The geographic coordinates at the study area are 31º09’38”S and 54º57’29”W, with 140m of altitude. According to Köppen the climate type is Cfa. The study in Chapters 2, 3 and 4 discusses the effects of the following treatments: 1) grassland management regime: exclusion (X) and rotational (R) and continuous (C) grazing; 2) initial levels of soil disturbance: high grassland - 10-12 cm height (A), low grassland - < 5 cm height (B) and low grassland with soil exposed (E); 3) fertilization regime: without fertilization (S), phosphorous (P) and nitrogen (N). The experimental design was a split-splitplot type in complete blocks. Three replicates were used. The grassland management regime, initial levels of soil disturbance and fertilization regime treatments were allocated in plots, split-plots and split-splitplots, respectively. The 54 split-splitplots (B and E split-plots) were sowed three winter cultivated species (Trifolium repens L. (white clover), Lotus corniculatus L. (birdsfoot trefoil São Gabriel) and Lolium multiflorum Lam. (annual rye grass)) and E. plana (capim-annoni-2). The other 27 split-splitplots (A split-plots) were sowed only capim-annoni-2. The samples of the SSB (Chapter 2) were collected in three dates: on March 4, 2004, before the application of treatments, and on April 22, 2005 and on May 6, 2006, after that application. In the three evaluation dates were recorded 33 families and 137 species in the SSB. White clover was registered only in the SSB2, birdsfoot trefoil São Gabriel was registered in three SSBs and annual rye grass and capim-annoni-2 were registered in the SSB2 and SSB3. Amongst that 137 species, 17 were aloctones and 39 pioneers. The pioneers Centunculus minimus, Anagallis arvensis and Gamochaeta spicata were registered in all surveying units of the three soil seed banks. The grassland management regime and initial levels of soil disturbance influenced the floristic composition and the soil seed bank size. In general, higher soil seed banks of capim-annoni-2 registered in the third evaluation (SSB3) were correlated with environmental conditions that preserve the seed viability such as grassland management regimes exclusion and rotative grazing. However, lower soil seed banks were associated with the grassland management continuous grazing, in which seeds were estimulated to germinate but also to deteriorate. The floristic surveying (Chapter 3) were done in seven seasons: autumn-2004, before treatment applications, spring-2004, autumns and springs-2005 and 2006, and autumn-2007. Twenty seven families and 141 resident species plus the 4 sowed ones were registered. Amongst 145 species, 36 were pioneers. The grassland management regime and initial levels of soil disturbance influenced the floristic composition of the community. The floristic composition trajectory of the community under the level of disturbance low grassland with soil exposed is distinct from the trajectories of the initial soil disturbance level grassland high and low. The exclusion regime reduced the species richness. In the Chapter 4, the initial levels of soil disturbance and grassland management regime influenced the establishment of capim-annoni-2. Its recruitment was favored by intensity of the soil disturbance and lower grassland height determined by continuous grazing. On the contrary, low intensity of the soil disturbance and higher grassland height, associated to the rotative grazing, avoided the establishment of capim-annoni-2.

Este estudo teve como objetivo determinar o efeito de distúrbios na forma de práticas de manejo sobre a dinâmica do banco de sementes do solo (BSS) e da vegetação residente em uma pastagem nativa do sul do Brasil, bem como do seu potencial de evitar que Eragrostis plana Nees (capim-annoni-2) nele se estabeleça. O experimento foi conduzido na Estância Upacaraí, município de Dom Pedrito, Estado do Rio Grande do Sul, Brasil, em área sem a presença do capim-annoni-2, cujo solo é classificado como Chernosolo Argilúvico Órtico Vértico, unidade de mapeamento Ponche Verde. As coordenadas geográficas centrais da área experimental são 31º09’38”S e 54º57’29”W, em altitude de 140 m. De acordo com a classificação de Köppen, o clima é Cfa. O estudo discute nos Capítulos 2, 3 e 4 os efeitos dos seguintes tratamentos: 1) regime de manejo da pastagem: exclusão (X) e pastejos rotativo (R) e contínuo (C); 2) níveis iniciais de distúrbio da pastagem: pastagem alta – 10-12 cm (A), pastagem baixa - < 5 cm (B) e pastagem baixa com exposição do solo (E); 3) regime de adubação: sem adubo (S), fósforo (P) e nitrogênio (N). O delineamento experimental foi em parcelas sub-subdivididas em blocos completos, compreendendo três repetições. Os tratamentos regime de manejo, níveis iniciais de distúrbio do solo e regime de adubação foram alocados nas parcelas, sub-parcelas e sub-subparcelas, respectivamente. Em 54 subsubparcelas (sub-parcelas B e E), semeou-se três espécies cultivadas de inverno (Trifolium repens L. (trevo-branco), Lotus corniculatus L. (cornichão São Gabriel) e Lolium multiflorum Lam. (azevém-anual)) e E. plana (capim-annoni-2). Nas demais 27 sub-subparcelas (subparcelas A), semeou-se somente capim-annoni-2. As amostras do BSS (Capítulo 2) foram coletadas em três datas: 4-3-2004, antes da aplicação dos tratamentos, e em 22-4-2005 e 6-5- 2006, após a aplicação dos mesmos. Nas três datas avaliadas, foram registradas 33 famílias e 137 espécies no BSS. Trevo-branco foi registrado apenas no BSS2, cornichão São Gabriel foi registrado nos três BSSs e azevém-anual e capim-annoni-2 foram registrados nos BSS2 e BSS3. Dentre as 137 espécies, 17 eram alóctones e 39 ruderais. As ruderais Centunculus minimus, Anagallis arvensis e Gamochaeta spicata foram registradas em todas as unidades amostrais dos três BSSs. O regime de manejo da pastagem e níveis iniciais de distúrbio do solo influenciaram a composição florística e o tamanho do banco de sementes do solo. Em geral, os maiores BSS de capim-annoni-2 registrados na terceira avaliação (BSS3), estiveram associados com as condições ambientais que preservaram a viabilidade das sementes, tais como os regimes de manejo exclusão e pastejo rotativo. Ao contrário, os bancos menores estão associados com o regime de pastejo contínuo, no qual as sementes foram expostas a uma condição ambiental que favoreceu a germinação mas também sua deterioração. Os levantamentos florísticos (Capítulo 3) foram realizados em sete épocas: outono-2004, antes da aplicação dos tratamentos, primavera-2004, outonos e primaveras de 2005 e 2006, e outono- 2007. Foram registradas 27 famílias e 141 espécies residentes + 4 alóctones semeadas. Das 145 espécies, 36 eram ruderais. O regime de manejo do campo e os níveis iniciais de distúrbio do solo influenciaram a composição florística da comunidade. A trajetória da composição florística das comunidades submetidas ao nível inicial de distúrbio pastagem baixa com exposição de solo é distinta daquela registrada nos níveis iniciais de distúrbio do solo pastagem alta e baixa. A exclusão reduziu a riqueza de espécies. No Capítulo 4, os níveis iniciais de distúrbio do solo e o regime de manejo influenciaram o estabelecimento do capimannoni- 2. O estabelecimento desta espécie foi favorecido pela intensidade do distúrbio do solo e menor altura da pastagem decorrente do regime pastejo contínuo. Ao contrário, a pastagem alta, associada ao pastejo rotativo, evitou o estabelecimento do capim-annoni-2. / The objective of this study was to evaluate the effects of disturbance imposed by management practices on the dynamics of the soil seed bank (SSB) in a natural grassland in Southern Brazil, as well as its potential in avoiding the invasion of Eragrostis plana Nees (capim-annoni-2). The experiment was carried out at Upacaraí Farm, in Dom Pedrito town, Rio Grande do Sul State, Brazil, in an area free of capim-annoni-2, in a Chernosolo Argilúvico Órtico Vértico soil type, Ponche Verde unit. The geographic coordinates at the study area are 31º09’38”S and 54º57’29”W, with 140m of altitude. According to Köppen the climate type is Cfa. The study in Chapters 2, 3 and 4 discusses the effects of the following treatments: 1) grassland management regime: exclusion (X) and rotational (R) and continuous (C) grazing; 2) initial levels of soil disturbance: high grassland - 10-12 cm height (A), low grassland - < 5 cm height (B) and low grassland with soil exposed (E); 3) fertilization regime: without fertilization (S), phosphorous (P) and nitrogen (N). The experimental design was a split-splitplot type in complete blocks. Three replicates were used. The grassland management regime, initial levels of soil disturbance and fertilization regime treatments were allocated in plots, split-plots and split-splitplots, respectively. The 54 split-splitplots (B and E split-plots) were sowed three winter cultivated species (Trifolium repens L. (white clover), Lotus corniculatus L. (birdsfoot trefoil São Gabriel) and Lolium multiflorum Lam. (annual rye grass)) and E. plana (capim-annoni-2). The other 27 split-splitplots (A split-plots) were sowed only capim-annoni-2. The samples of the SSB (Chapter 2) were collected in three dates: on March 4, 2004, before the application of treatments, and on April 22, 2005 and on May 6, 2006, after that application. In the three evaluation dates were recorded 33 families and 137 species in the SSB. White clover was registered only in the SSB2, birdsfoot trefoil São Gabriel was registered in three SSBs and annual rye grass and capim-annoni-2 were registered in the SSB2 and SSB3. Amongst that 137 species, 17 were aloctones and 39 pioneers. The pioneers Centunculus minimus, Anagallis arvensis and Gamochaeta spicata were registered in all surveying units of the three soil seed banks. The grassland management regime and initial levels of soil disturbance influenced the floristic composition and the soil seed bank size. In general, higher soil seed banks of capim-annoni-2 registered in the third evaluation (SSB3) were correlated with environmental conditions that preserve the seed viability such as grassland management regimes exclusion and rotative grazing. However, lower soil seed banks were associated with the grassland management continuous grazing, in which seeds were estimulated to germinate but also to deteriorate. The floristic surveying (Chapter 3) were done in seven seasons: autumn-2004, before treatment applications, spring-2004, autumns and springs-2005 and 2006, and autumn-2007. Twenty seven families and 141 resident species plus the 4 sowed ones were registered. Amongst 145 species, 36 were pioneers. The grassland management regime and initial levels of soil disturbance influenced the floristic composition of the community. The floristic composition trajectory of the community under the level of disturbance low grassland with soil exposed is distinct from the trajectories of the initial soil disturbance level grassland high and low. The exclusion regime reduced the species richness. In the Chapter 4, the initial levels of soil disturbance and grassland management regime influenced the establishment of capim-annoni-2. Its recruitment was favored by intensity of the soil disturbance and lower grassland height determined by continuous grazing. On the contrary, low intensity of the soil disturbance and higher grassland height, associated to the rotative grazing, avoided the establishment of capim-annoni-2.

Este estudo teve como objetivo determinar o efeito de distúrbios na forma de práticas de manejo sobre a dinâmica do banco de sementes do solo (BSS) e da vegetação residente em uma pastagem nativa do sul do Brasil, bem como do seu potencial de evitar que Eragrostis plana Nees (capim-annoni-2) nele se estabeleça. O experimento foi conduzido na Estância Upacaraí, município de Dom Pedrito, Estado do Rio Grande do Sul, Brasil, em área sem a presença do capim-annoni-2, cujo solo é classificado como Chernosolo Argilúvico Órtico Vértico, unidade de mapeamento Ponche Verde. As coordenadas geográficas centrais da área experimental são 31º09’38”S e 54º57’29”W, em altitude de 140 m. De acordo com a classificação de Köppen, o clima é Cfa. O estudo discute nos Capítulos 2, 3 e 4 os efeitos dos seguintes tratamentos: 1) regime de manejo da pastagem: exclusão (X) e pastejos rotativo (R) e contínuo (C); 2) níveis iniciais de distúrbio da pastagem: pastagem alta – 10-12 cm (A), pastagem baixa - < 5 cm (B) e pastagem baixa com exposição do solo (E); 3) regime de adubação: sem adubo (S), fósforo (P) e nitrogênio (N). O delineamento experimental foi em parcelas sub-subdivididas em blocos completos, compreendendo três repetições. Os tratamentos regime de manejo, níveis iniciais de distúrbio do solo e regime de adubação foram alocados nas parcelas, sub-parcelas e sub-subparcelas, respectivamente. Em 54 subsubparcelas (sub-parcelas B e E), semeou-se três espécies cultivadas de inverno (Trifolium repens L. (trevo-branco), Lotus corniculatus L. (cornichão São Gabriel) e Lolium multiflorum Lam. (azevém-anual)) e E. plana (capim-annoni-2). Nas demais 27 sub-subparcelas (subparcelas A), semeou-se somente capim-annoni-2. As amostras do BSS (Capítulo 2) foram coletadas em três datas: 4-3-2004, antes da aplicação dos tratamentos, e em 22-4-2005 e 6-5- 2006, após a aplicação dos mesmos. Nas três datas avaliadas, foram registradas 33 famílias e 137 espécies no BSS. Trevo-branco foi registrado apenas no BSS2, cornichão São Gabriel foi registrado nos três BSSs e azevém-anual e capim-annoni-2 foram registrados nos BSS2 e BSS3. Dentre as 137 espécies, 17 eram alóctones e 39 ruderais. As ruderais Centunculus minimus, Anagallis arvensis e Gamochaeta spicata foram registradas em todas as unidades amostrais dos três BSSs. O regime de manejo da pastagem e níveis iniciais de distúrbio do solo influenciaram a composição florística e o tamanho do banco de sementes do solo. Em geral, os maiores BSS de capim-annoni-2 registrados na terceira avaliação (BSS3), estiveram associados com as condições ambientais que preservaram a viabilidade das sementes, tais como os regimes de manejo exclusão e pastejo rotativo. Ao contrário, os bancos menores estão associados com o regime de pastejo contínuo, no qual as sementes foram expostas a uma condição ambiental que favoreceu a germinação mas também sua deterioração. Os levantamentos florísticos (Capítulo 3) foram realizados em sete épocas: outono-2004, antes da aplicação dos tratamentos, primavera-2004, outonos e primaveras de 2005 e 2006, e outono- 2007. Foram registradas 27 famílias e 141 espécies residentes + 4 alóctones semeadas. Das 145 espécies, 36 eram ruderais. O regime de manejo do campo e os níveis iniciais de distúrbio do solo influenciaram a composição florística da comunidade. A trajetória da composição florística das comunidades submetidas ao nível inicial de distúrbio pastagem baixa com exposição de solo é distinta daquela registrada nos níveis iniciais de distúrbio do solo pastagem alta e baixa. A exclusão reduziu a riqueza de espécies. No Capítulo 4, os níveis iniciais de distúrbio do solo e o regime de manejo influenciaram o estabelecimento do capimannoni- 2. O estabelecimento desta espécie foi favorecido pela intensidade do distúrbio do solo e menor altura da pastagem decorrente do regime pastejo contínuo. Ao contrário, a pastagem alta, associada ao pastejo rotativo, evitou o estabelecimento do capim-annoni-2. / The objective of this study was to evaluate the effects of disturbance imposed by management practices on the dynamics of the soil seed bank (SSB) in a natural grassland in Southern Brazil, as well as its potential in avoiding the invasion of Eragrostis plana Nees (capim-annoni-2). The experiment was carried out at Upacaraí Farm, in Dom Pedrito town, Rio Grande do Sul State, Brazil, in an area free of capim-annoni-2, in a Chernosolo Argilúvico Órtico Vértico soil type, Ponche Verde unit. The geographic coordinates at the study area are 31º09’38”S and 54º57’29”W, with 140m of altitude. According to Köppen the climate type is Cfa. The study in Chapters 2, 3 and 4 discusses the effects of the following treatments: 1) grassland management regime: exclusion (X) and rotational (R) and continuous (C) grazing; 2) initial levels of soil disturbance: high grassland - 10-12 cm height (A), low grassland - < 5 cm height (B) and low grassland with soil exposed (E); 3) fertilization regime: without fertilization (S), phosphorous (P) and nitrogen (N). The experimental design was a split-splitplot type in complete blocks. Three replicates were used. The grassland management regime, initial levels of soil disturbance and fertilization regime treatments were allocated in plots, split-plots and split-splitplots, respectively. The 54 split-splitplots (B and E split-plots) were sowed three winter cultivated species (Trifolium repens L. (white clover), Lotus corniculatus L. (birdsfoot trefoil São Gabriel) and Lolium multiflorum Lam. (annual rye grass)) and E. plana (capim-annoni-2). The other 27 split-splitplots (A split-plots) were sowed only capim-annoni-2. The samples of the SSB (Chapter 2) were collected in three dates: on March 4, 2004, before the application of treatments, and on April 22, 2005 and on May 6, 2006, after that application. In the three evaluation dates were recorded 33 families and 137 species in the SSB. White clover was registered only in the SSB2, birdsfoot trefoil São Gabriel was registered in three SSBs and annual rye grass and capim-annoni-2 were registered in the SSB2 and SSB3. Amongst that 137 species, 17 were aloctones and 39 pioneers. The pioneers Centunculus minimus, Anagallis arvensis and Gamochaeta spicata were registered in all surveying units of the three soil seed banks. The grassland management regime and initial levels of soil disturbance influenced the floristic composition and the soil seed bank size. In general, higher soil seed banks of capim-annoni-2 registered in the third evaluation (SSB3) were correlated with environmental conditions that preserve the seed viability such as grassland management regimes exclusion and rotative grazing. However, lower soil seed banks were associated with the grassland management continuous grazing, in which seeds were estimulated to germinate but also to deteriorate. The floristic surveying (Chapter 3) were done in seven seasons: autumn-2004, before treatment applications, spring-2004, autumns and springs-2005 and 2006, and autumn-2007. Twenty seven families and 141 resident species plus the 4 sowed ones were registered. Amongst 145 species, 36 were pioneers. The grassland management regime and initial levels of soil disturbance influenced the floristic composition of the community. The floristic composition trajectory of the community under the level of disturbance low grassland with soil exposed is distinct from the trajectories of the initial soil disturbance level grassland high and low. The exclusion regime reduced the species richness. In the Chapter 4, the initial levels of soil disturbance and grassland management regime influenced the establishment of capim-annoni-2. Its recruitment was favored by intensity of the soil disturbance and lower grassland height determined by continuous grazing. On the contrary, low intensity of the soil disturbance and higher grassland height, associated to the rotative grazing, avoided the establishment of capim-annoni-2.