Ecol Appl. 2014;24(8):2122-31. doi: 10.1890/13-2168.1.
Translocation experiments, in which researchers displace animals and then monitor their movements to return home, are commonly used as tools to assess functional connectivity of fragmented landscapes. Such experiments are purported to have important advantages of being time efficient and of standardizing “motivation” to move across individuals. Yet, we lack tests of whether movement behavior of translocated birds reflects natural behavior of unmanipulated birds. We compared the routine movement behavior of a tropical hummingbird, the Green Hermit (Phaethornis guy), to that of experimentally translocated individuals. We tested for differences in site selection patterns during movement at two spatial scales (point and path levels). We also compared movement rates between treatments. Behaviors documented during translocation experiments reflected those observed during routine movements. At the point level, both translocated and non-translocated birds showed similar levels of preference for mature tropical forest. At the path level, step selection functions showed both translocated and non-translocated hummingbirds avoiding movement across non-forested matrix and selecting streams as movement corridors. Movement rates were generally higher during translocation experiments. However, the negative influence of forest cover on movement rates was proportionately similar in translocation and routine movement treatments. We report the first evidence showing that movement behavior of birds during translocation experiments is similar to their natural movement behavior. Therefore, translocation experiments may be reliable tools to address effects of landscape structure on animal movement. We observed consistent selection of landscape elements between translocated and non-translocated birds, indicating that both routine and translocation movement studies lead to similar conclusions regarding the effect of landscape structure and forest composition on functional connectivity. Our observation that hummingbirds avoid non-forest matrix and select riparian corridors also provides a potential mechanism for pollen limitation in fragmented tropical forest.
迁移实验，即研究人员移动动物，然后监测它们的返回运动，通常被用作评估破碎景观功能连接的工具。这种实验据称具有时间效率高和标准化个体移动“动机”的重要优势。然而，我们缺乏对被迁移鸟类的运动行为是否反映未经操纵鸟类自然行为的测试。我们比较了热带蜂鸟绿隐蜂鸟（Phaethornis guy）的常规运动行为和实验性迁移个体的运动行为。我们在两个空间尺度（点和路径水平）上测试了运动期间的选址模式差异。我们还比较了处理之间的移动速度。在迁移实验中记录的行为反映了在常规运动中观察到的行为。在点水平上，被迁移和未被迁移的鸟类对成熟热带森林表现出相似的偏好水平。在路径水平上，步长选择函数表明，被迁移和未被迁移的蜂鸟都避免在无林地基质中移动，并选择溪流作为移动走廊。在迁移实验中，移动速度通常较高。然而，森林覆盖对移动速度的负面影响在迁移和常规运动处理中比例相似。我们报告了第一个证据，表明鸟类在迁移实验中的运动行为与其自然运动行为相似。因此，迁移实验可能是解决景观结构对动物运动影响的可靠工具。我们观察到被迁移和未被迁移的鸟类之间对景观要素的一致选择，表明常规和迁移运动研究都得出了关于景观结构和森林组成对功能连接的影响的类似结论。我们观察到蜂鸟避免非林地基质并选择河岸走廊，这也为破碎热带森林中花粉限制提供了潜在的机制。
