Abstract:Lianas are an important component in tropical rain forests and are often cited as the most obvious physiognomic difference between temperate and tropical forests. Lianas can have profound impacts on the diversity, structure, dynamics and functions of tropical rain forests. However, due to their relative smaller stem size, defining liana individuals are difficult in the field, so lianas are often excluded or ignored in most forest surveys. The relationship between lianas and their host trees remains essentially unknown even in some of the most intensely studied tropical forests. Forest fragmentation (e.g., formation of edges) has been hypothesized to be partly responsible for the increased liana abundance in fragmented tropical rain forests due to their light-demanding nature. Previous studies suggest that lianas often respond positively to increased light induced by disturbance (e.g., windthrows and harvesting) and often increase competition with their host trees which provide the essential support. On the other hand, liana infestations could promote the formation of large treefall gaps by entangling crowns of adjacent trees and the infestations could also create structural stresses on trees and competition for light, moisture and soil nutrients, resulting in reduced tree growth and reproduction. To assess the effects of edges on tree-liana relationship, we chose two forest edges which differed in their formation ages (17 years and 13 years, respectively) in an area of total 0.8 hm2 from edge to forest interior in a 60-year old secondary natural tropical rain forest in Bawangling Nature Reserve, Hainan Island, China. In total, 3250 liana individuals were recorded in the 4 transects perpendicular to forest edges. Each transect was 100×10 m in size and subdivided into 10 10×10 m subplots. Each transect started at the forest edge and continued 100 m into the interior. Liana and tree abundance, diameters (trees ≧1cm diameter at breast height ), and number of trees with lianas were recorded. Our results show that major climbing guilds of lianas (mainstem-twiners, brunch-twiners, tendril-twiners) were more abundant on the edges. Specially: (1) In both forest edges, proportion of trees with lianas changed with distance to forest edge for the smaller trees (dbh 1-5 cm and 5-10 cm), while the larger trees (dbh>20 cm) in the 17-year old edge and the trees dbh 10-20 cm in the 13-year old edge were not affected by forest edge. (2) The infestation of trees by lianas decreased with increasing distance to forest edges. The number of trees with more than three lianas was the lowest and changed significantly in the 17-year old edge, while the number of trees with one liana was the highest and changed significantly in the 13-year old edge. (3) Lianas per tree decreased with increasing distance to the edge in the 17-year old edge. Formation age of edges affected the number of lianas per tree, which decreased sharply in the 17-year old forest edge. We concluded that tree-liana relationships near edges are more complicated than in the interior and many aspects of liana community structure and traits are affected by habitat fragmentation.