No abstract

Seed dispersal is a crucial component of plant population dynamics. Human landscape modifications, such as habitat destruction and fragmentation, can alter the abundance of fruiting plants and animal dispersers, foraging rates, vector movement, and the composition of the disperser community, all of which can singly or in concert affect seed dispersal. Here, we quantify and tease apart the effects of landscape configuration, namely, fragmentation of primary forest and the composition of the surrounding forest matrix, on individual components of seed dispersal of Heliconia acuminata, an Amazonian understory herb. First we identified the effects of landscape configuration on the abundance of fruiting plants and six bird disperser species. Although highly variable in space and time, densities of fruiting plants were similar in continuous forest and fragments. However, the two largest-bodied avian dispersers were less common or absent in small fragments. Second, we determined whether fragmentation affected foraging rates. Fruit removal rates were similar and very high across the landscape, suggesting that Heliconia fruits are a key resource for small frugivores in this landscape. Third, we used radiotelemetry and statistical models to quantify how landscape configuration influences vector movement patterns. Bird dispersers flew farther and faster, and perched longer in primary relative to secondary forests. One species also altered its movement direction in response to habitat boundaries between primary and secondary forests. Finally, we parameterized a simulation model linking data on fruit density and disperser abundance and behavior with empirical estimates of seed retention times to generate seed dispersal patterns in two hypothetical landscapes. Despite clear changes in bird movement in response to landscape configuration, our simulations demonstrate that these differences had negligible effects on dispersal distances. However, small fragments had reduced densities of Turdus albicollis, the largest-bodied disperser and the only one to both regurgitate and defecate seeds. This change in Turdus abundance acted together with lower numbers of fruiting plants in small fragments to decrease the probability of long-distance dispersal events from small patches. These findings emphasize the importance of foraging style for seed dispersal and highlight the primacy of habitat size relative to spatial configuration in preserving biotic interactions.

Reproduction by individuals is typically recorded as count data (e.g., number of fledglings from a nest or inflorescences on a plant) and commonly modeled using Poisson or negative binomial distributions, which assume that variance is greater than or equal to the mean. However, distributions of reproductive effort are often underdispersed (i.e., variance < mean). When used in hypothesis tests, models that ignore underdispersion will be overly conservative and may fail to detect significant patterns. Here we show that generalized Poisson (GP) and Conway‐Maxwell‐Poisson (CMP) distributions are better choices for modeling reproductive effort because they can handle both overdispersion and underdispersion; we provide examples of how ecologists can use GP and CMP distributions in generalized linear models (GLMs) and generalized linear mixed models (GLMMs) to quantify patterns in reproduction. Using a new R package, glmmTMB, we construct GLMMs to investigate how rainfall and population density influence the number of fledglings in the warbler Oreothlypis celata and how flowering rate of Heliconia acuminata differs between fragmented and continuous forest. We also demonstrate how to deal with zero‐inflation, which occurs when there are more zeros than expected in the distribution, e.g., due to complete reproductive failure by some individuals.

Abstract. Studies simultaneously evaluating the importance of safe-site and seed limitation for plant establishment are rare, particularly in human-modified landscapes. We used spatially explicit neighborhood models together with data from 10 0.5-ha mapped census plots in a fragmented landscape spanning 1000 km 2 to (1) evaluate the relative importance of seed production, dispersal, and safe-site limitation for the recruitment of the understory herb Heliconia acuminata; and (2) determine how these processes differ between fragments and continuous forests. Our analyses demonstrated a large degree of variation in seed production, dispersal, and establishment among and within the 10 study plots. Seed production limitation was strong but only at small spatial scales. Average dispersal distance was less than 4 m, leading to severe dispersal limitation at most sites. Overall, safe-site limitation was the most important constraint on seedling establishment. Fragmentation led to a more heterogeneous light environment with negative consequences for seedling establishment but had little effect on seed production or dispersal. These results suggest that the effects of fragmentation on abiotic processes may be more important than the disruption of biotic interactions in driving biodiversity loss in tropical forests, at least for some functional groups. These effects may be common when the matrix surrounding fragments contains enough tree cover to enable movement of dispersers and pollinators.

Recebido em 6/09/2009. Aceito em 15/06//2010 RESUMO -(Variações interanuais na fenologia de uma comunidade arbórea de fl oresta semidecídua no sudeste do Brasil). As comunidades arbóreas de fl orestas tropicais localizadas em regiões com clima sazonal tendem a ajustar a fenologia da queda de folhas e da produção de fl ores e frutos à sazonalidade climática. Neste estudo monitoramos a comunidade arbórea de uma fl oresta semidecidual no período de quatro anos (2001 a 2004), a fi m de analisar a relação entre o número de espécies perdendo folhas, fl orescendo e frutifi cando e a variação anual do comprimento do dia, precipitação e temperatura durante o período de observação. Além disso, avaliamos se essas três fenofases repetem os mesmos padrões em anos sucessivos. Regressões lineares simples indicaram que o número de espécies perdendo folhas foi negativamente relacionado com as três variáveis ambientais para os quatro anos de estudo, sendo o comprimento do dia e a temperatura os melhores preditores para esta fenofase. Houve também relação signifi cativa do comprimento do dia com o número de espécies fl orescendo e frutifi cando nos quatro anos, porém, precipitação e temperatura não foram relacionadas com estas fenofases em todos os anos de estudo. Os testes de estatística circular indicaram forte sazonalidade para queda foliar, com data média durante o mês de julho independente do ano, ao passo que para fl oração e frutifi cação não foi detectada sazonalidade signifi cativa. Como resultado, o número de espécies perdendo folhas foi correlacionado entre todos os anos sucessivos. Entretanto, não houve correlação do número de espécies fl orescendo e frutifi cando em 2003 e 2004, provavelmente devido às altas temperaturas ocorridas neste último ano. Os resultados deste estudo sugerem que a queda de folhas é uma fenofase altamente previsível temporalmente. Ao contrário, a fl oração e a frutifi cação podem variar ano a ano como conseqüência de anomalias climáticas. Essas variações podem ser relevantes para o entendimento dos efeitos em longo prazo de mudanças climáticas sobre as fl orestas sazonais. Palavras-chave: comprimento do dia, fl oresta Atlântica, mudanças climáticas, sazonalidade ABSTRACT -(Interannual variation in the phenology of a tree community in a semideciduous seasonal forest in southeast Brazil). Tropical tree communities in regions with seasonal climate usually adjust leaf-fall, fl owering and fruiting phenology to climate seasonality. Here we monitored a tree community over a 4-year period in order to assess the relationship between the number of species presenting leaf fall, fl ower and fruit and annual variation in day length, precipitation and temperature. Moreover, we asked whether the three phenological variables have the same patterns in successive years. Simple regression analyses showed that the number of species shedding leaves was negatively related to the three environmental variables during the four years, with day length and temperature as the best predictors. There was also a relationship...