【目的】 探究生境破碎化产生的岛屿面积效应和岛屿边缘效应能否“自下而上（Bottom-up impact）”影响瘿蜂群落结构与分布。【方法】 对不同岛屿面积和岛屿内外样带上白栎Quercus fabri和枹栎Q. serrata的12组叶功能性状进行测定，结合样带寄主植物上瘿蜂虫瘿的数量，分析面积效应与边缘效应对于瘿蜂群落的影响。【结果】 （1）白栎和枹栎在不同面积岛屿内外样带上的分布密度无差异（P>0.05）。（2）白栎的叶氮含量和枹栎的叶绿素含量与岛屿面积呈显著正相关（P<0.05），白栎的叶干重和枹栎的叶厚度与岛屿面积呈显著负相关（P<0.05），但白栎和枹栎的叶功能性状在岛屿边缘与内部之间均未出现显著差异（P>0.05）。（3）白栎与枹栎上的致虫瘿瘿蜂均为纹瘿蜂属Andricus，其中白栎上3种，枹栎上5种。（4）瘿蜂虫瘿数量在岛屿内外样带上没有显著差异（P>0.05），但会随着岛屿面积的增大而变多，与白栎的叶氮含量、叶面积和比叶面积显著正相关（P<0.05），与枹栎的叶绿素、叶面积和比叶面积显著正相关（P<0.05），与白栎和枹栎的叶组织密度呈显著负相关（P<0.05），与二者第一主成分显著正相关（P<0.05）。【结论】 生境破碎化造成的岛屿边缘效应不能明显影响瘿蜂的群落结构与分布，岛屿面积效应通过寄主植物对瘿蜂产生影响；瘿蜂产卵造瘿的偏好与寄主植物对光资源利用能力的强弱存在一定关系。

 [Aim]  To investigate if the island area and island edge effects caused by habitat fragmentation can have a “bottom-up impact” on the distribution and structure of the cynipid gall wasp community. [Methods]  Twelve sets of functional leaf traits for Quercus fabri and Q. serrata were measured from sample belts interori and on the edge of different sized islands in Thousand Island Lake. The influence of the area effect and edge effect on the gall wasp community was analyzed using the combined number of gall wasps on host plants from all the sample belts. [Results]  (1) There was no difference in the distribution densities of Q. fabri or Q. serrata between the interior and edge transects for the different sized islands (P>0.05). (2) The nitrogen content of Q. fabri and the chlorophyll content of Q. serrata showed a significant positive correlation with island area (P<0.05), while the leaf dry weight of Q. fabri and the leaf thickness of Q. serrata were significantly negatively correlated with island area (P<0.05). However, there was no significant difference in leaf functional traits between the island edge and the interior for either species (P>0.05). (3) The gall wasps present on Q. fabri and Q. serrata all belonged to the genus Andricus, of these three species were found on Q. fabri and five species on Q. serrata. (4) There was no difference in the number of gall wasp galls between the interior and edge island sample belts (P>0.05). However, the number of galls increased with island size. Gall wasp number was significantly positively correlated with the leaf nitrogen content, leaf area, and specific leaf area of Q. fabri (P<0.05), and with chlorophyll content, leaf area, and specific leaf area of Q. serrata (P<0.05). In contrast, gall wasp number was significantly negatively correlated with leaf tissue density (P<0.05) and significantly positively correlated with the first principal component for both tree species (P<0.05). [Conclusion]  The island edge effect, resulting from habitat fragmentation, did not have an obvious effect on the community structure or distribution of gall wasps. However, the island area effect influenced gall wasps by affecting their host plants. Gall wasp host preference for egg-laying and gall formation is related to the ability of host plants to utilize light resources.