Nested species subsets of amphibians and reptiles on Neotropical forest islands J. I. Watling 1à , K. Gerow 2 & M. A. Donnelly 1 1Department of Biological Sciences, Florida International University, Miami, FL, USA 2Statistics Department, University of Wyoming, Laramie, WY, USA Keywords amphibians; Bolivia; forest islands; fragmentation; reptiles; savannah. Correspondence James I. Watling, Department of Biological Sciences, Florida International University, University Park, Miami, FL 33199, USA. Tel: +1 305 348 6513; Fax: +1 305 348 1986 Email: watlingj@wustl.edu à Current address: J. I. Watling. Department of Biology, Washington University in St Louis, Campus Box 1137, 1 Brookings Drive, St Louis, MO 63130, USA. Received 19 December 2008; accepted 17 April 2009 doi:10.1111/j.1469-1795.2009.00274.x Abstract Nested species subsets are a common pattern of community assembly character- istic of many types of fragmented landscapes and insular systems. Here we describe nested subset patterns of amphibian and reptile occupancy on 23 forest islands in north-eastern Bolivia. We used observed occupancy patterns to differentiate five distributional guilds: widespread species, rare species, poor colonizers, area- sensitive species and supertramps. Amphibian occurrences were nested along a forest island isolation gradient, and when species from each of the distribution classes were removed from subsequent analyses of nestedness, we found that dispersal-limited poor colonizers were responsible for the association between nestedness and isolation. Amphibians associated with the grassland matrix at the study site showed a nested pattern linked with area, although this pattern did not scale up to all amphibians and could not be unequivocally attributed to any of the distributional guilds we recognized. There were no strong associations between two biological characteristics, body size and relative abundance in the matrix, and the likelihood of occupancy along either forest island area or isolation gradients. The relative importance of isolation in shaping nested patterns of amphibians on these forest islands may be a result of either (1) the greater range in isolation values included in this study compared with many others; (2) the long time since isolation in this landscape, manifesting a footprint of isolation not apparent in more recently fragmented patches; (3) the relatively homogeneous grassland matrix surrounding forest islands that likely provides little refuge for animals moving among forest islands. Introduction Nestedness is a condition in which species distributions occur hierarchically such that the fauna of species-poor patches comprise a perfect subset of the fauna on increas- ingly species-rich patches (Patterson & Atmar, 1986; Wright & Reeves, 1992; Worthen, 1996). Nested species subsets are characteristic of oceanic and land-bridge islands, habitat islands and habitat patches of only a few square meters (Cook, 1995; Hecnar et al., 2002; Mac Nally, Horrocks & Bennett, 2002; Summerville, Veech & Crist, 2002). Four explanations have been suggested to cause nested patterns: passive sampling, nested habitats, differential extinction and differential colonization. Although passive sampling and nested habitats have been implicated as driving forces behind nested patterns in some systems (Calm´e & Desro- chers, 1999; Honnay, Hermy & Coppin, 1999), most studies have focused on extinction and colonization as drivers of nested distribution patterns (Bolger, Alberts & Soul´e, 1991; Cook & Quinn, 1995; Lomolino, 1996). Extinction and colonization, and their relationship with area and isolation, respectively, lie at the heart of the Equilibrium Theory of Island Biogeography (ETIB), which rests on the assump- tions that extinction rates increase with decreasing island area and colonization rates decrease with increasing dis- tance from a ‘mainland’ source of colonists (MacArthur & Wilson, 1967). Although the ETIB was originally proposed to explain distribution patterns on true islands, the theory has since been applied to a variety of island-like habitats, and we use the term ‘island’ in a broad sense to refer to any discrete patch surrounded by an obviously different ‘matrix’ habitat. Most research describing nested subset patterns has focused on distilling the information contained in a species by site matrix into a metric of nestedness (see reviews in Wright et al., 1998; Watling & Donnelly, 2006). Recognizing that important patterns may be identified by describing the responses of individual species to fragmentation (e.g. Gil- bert, 1980; Lomolino, 1999, 2000a,b), some authors have described how individual species conform to or depart from a community-level nested pattern (Martinez-Morales, 2005). The concept of nestedness has been central to the Animal Conservation 12 (2009) 467–476 c  2009 The Authors. Journal compilation c  2009 The Zoological Society of London 467 Animal Conservation. Print ISSN 1367-9430