Submitted by PPG Zoologia (zoologia-pg@pucrs.br) on 2018-05-29T13:56:09Z No. of bitstreams: 1 Disserta??o_NevesMB.pdf: 2866910 bytes, checksum: 6b6ea49def601cec24809d0304a5df5e (MD5) / Approved for entry into archive by Caroline Xavier (caroline.xavier@pucrs.br) on 2018-06-11T19:19:30Z (GMT) No. of bitstreams: 1 Disserta??o_NevesMB.pdf: 2866910 bytes, checksum: 6b6ea49def601cec24809d0304a5df5e (MD5) / Made available in DSpace on 2018-06-11T19:41:34Z (GMT). No. of bitstreams: 1 Disserta??o_NevesMB.pdf: 2866910 bytes, checksum: 6b6ea49def601cec24809d0304a5df5e (MD5) Previous issue date: 2018-04-04 / Coordena??o de Aperfei?oamento de Pessoal de N?vel Superior - CAPES / Mutualistic relationships between plants and animals play a key role in ecological and evolutionary processes. These relationships are seen as complex networks of interaction, which present recurrent structural patterns and are severely threatened by anthropogenic factors. The objective of this work was to evaluate the influence of the isolation caused by urbanization on the interaction networks between plants and floral visitors. Twelve South Brazilian fragments were sampled in the metropolitan area of Porto Alegre, Rio Grande do Sul. Sampling was performed during spring and summer of 2016/17 in 1-hectare circular plots using the focal method (10 minutes observation / plant). We estimated the degree of isolation of the fragments by calculating the area covered by urban matrix in a buffer of 2000 meters around each plot. We performed a qualitative-quantitative sampling of vegetation as part of the explanatory data observed for the network patterns. To construct interaction matrices, we divide visitors into potential pollinators and non-pollinators. We construct quantitative interaction matrices for each fragment/plot by calculating the metrics available in the networklevel function of the bipartite package of the R program. We estimate the correlation between the logarithm of the urbanized area in a radius of 2000 meters and each network metric calculated with the generalized linear model. We sampled 1240 plants (182 species). The total abundance of floral visitors was 1840 specimens (94% potential pollinators). The richness of plant species was greater in areas more isolated by urbanization, which is due to the fact that these areas catch fire more frequently. The richness of visitors was also higher in these areas, which is due to a "bottom up" effect given by the increase in plant richness. The results suggest that networks increase connectance and nestedness (in more urbanized areas) as a way of protection against the disturbance, increasing system stability. In addition, areas further away from urbanization allow a greater level of network specialization, since they have a greater range of niches. We conclude that these findings highlight the importance of investing in the conservation and maintenance of natural areas within large cities, although they may seem isolated in the midst of a dense urban matrix. These areas are still a good haven for wildlife and need to be kept intact. / As rela??es mutual?sticas entre plantas e animais desempenham um papel chave nos processos ecol?gicos e evolutivos. Estas rela??es s?o vistas como redes complexas de intera??o, que apresentam padr?es estruturais recorrentes e est?o severamente amea?adas por a??es antropog?nicas. Este trabalho teve como objetivo avaliar a influ?ncia do isolamento causado pela urbaniza??o nas redes de intera??o entre plantas e visitantes florais. Foram amostrados 12 fragmentos campestres na regi?o metropolitana de Porto Alegre, Rio Grande do Sul. A amostragem foi realizada durante a primavera e o ver?o de 2016/17, em parcelas circulares de 1 hectare atrav?s do m?todo focal (10 minutos de observa??o/planta). Estimamos o grau de isolamento dos fragmentos calculando a ?rea coberta por matriz urbana em uma ?rea tamp?o de 2000 metros em torno de cada parcela. Realizamos um levantamento qualiquantitativo da vegeta??o como vari?vel explanat?ria para os padr?es de redes encontrados. Para construir as matrizes de intera??o dividimos os visitantes em potenciais polinizadores e n?o polinizadores. Constru?mos matrizes quantitativas de intera??o para cada fragmento, calculando as m?tricas dispon?veis na fun??o networklevel do pacote bipartite do programa R. Estimamos a correla??o entre o logaritmo da ?rea urbanizada em um raio de 2000 metros e cada m?trica de rede calculada com modelo linear generalizado. Foram amostradas 1240 plantas (182 esp?cies). A abund?ncia total de visitantes florais foi de 1840 esp?cimes (94% potenciais polinizadores). A riqueza de esp?cies de plantas foi maior em ?reas mais isoladas pela urbaniza??o, o que se deve ao fato destas ?reas pegarem fogo com mais frequ?ncia. A riqueza de visitantes tamb?m foi maior nestas ?reas, o que se deve por um efeito ?bottom up? dado pelo aumento na riqueza de plantas. Os resultados sugerem que as redes aumentam a conect?ncia e aninhamento (em ?reas mais urbanizadas) como uma forma de ?driblar? o dist?rbio, aumentando assim a estabilidade do sistema. Al?m disto, ?reas mais afastadas da urbaniza??o permitem um maior n?vel de especializa??o das redes, pois possuem maior amplitude de nichos. Conclu?mos que essas descobertas destacam a import?ncia do investimento na conserva??o e manuten??o de ?reas naturais nas grandes cidades, embora possam parecer isoladas em meio a uma densa matriz urbana. Estas ?reas ainda s?o um bom ref?gio para a vida selvagem e precisam ser mantidas intactas.