When the prairie varies: the importance of site characteristics for strategising insect conservation CHRIS LOONEY, 1 BENJAMIN T. CALDWELL 2 and SANFORD D. EIGENBRODE 1 1 Department of Plant, Soil and Entomological Sciences, University of Idaho, Moscow, ID, USA and 2 Department of Environmental Studies, Policy, and Management, University of California, Berkeley, CA, USA Abstract. 1. The bunchgrass prairies of the Palouse region in eastern Washington state and adjacent Idaho have been mostly converted to agriculture in the past century. Prairie habitat currently exists only on small remnants scattered across the landscape. 2. The invertebrate fauna of these habitat remnants is poorly known, both in terms of species diversity and community composition. 3. Baited pitfall traps were used to sample carrion-attracted beetles during June and July of 2003 from high-quality prairie remnants of varying size and soil characteristics. 4. Twenty-three beetle species in five families were found. Species richness and Shan- non-Wiener diversity were not significantly correlated with habitat area or perimeter– area ratio. 5. Beetle abundance and community structure were strongly correlated with soil characteristics. Deeper, loessal soils had greater overall beetle abundance than shal- lower, rocky soils, although some species were more frequently captured in shallower, rocky soil sites. 6. Conservation plans emphasising only remnant size and condition do not suffi- ciently capture variability in insect communities across Palouse prairie remnants. Key words. Bunchgrass, Dermestidae, diversity, fragmentation, Histeridae, Palouse, prairie, Silphidae. Introduction Habitat loss and fragmentation are two of the most prevalent threats to biodiversity worldwide (Noss et al., 1995; Fahrig, 2003). Their impacts are particularly evident in grassland ecosys- tems of the United States, where native biological communities have been replaced by intensive agriculture (Smith, 1981; Ingersoll and Wilson, 1991; Black et al ., 1998). Many grasslands in the United States are so reduced in extent that they are critically endangered, with <2% of their original coverage remaining (Samson and Knopf, 1994; Noss et al., 1995). Typically, native grasslands in agricultural regions only exist as small fragments embedded in a matrix of intensive agriculture (Samson and Knopf, 1994; Noss et al., 1995; Tscharntke et al., 2002a). Species in these ecosystems face increased risk of extinction from the negative effects of small populations and habitat fragmentation. These include increased risk of local extirpation due to stochastic effects (Tscharntke et al., 2002b; Tscharntke and Brandl, 2004), reduced gene-flow (Nei, 1973; Saccheri et al., 1998; Andersen et al., 2004), and the disruption of ecological relationships (Roland and Taylor, 1997; Tscharntke and Brandl, 2004). Conservation priorities in fragmented ecosystems are often heavily influenced by considering the size and shape (or amount of ‘edge’) of habitat remnants (e.g. Washington Department of Natural Resources, 2005), in part because of the well-supported positive relationship between habitat area or edge and species number (reviewed in Rosenzweig, 1995; and Fletcher et al ., 2007). However, these parameters alone are likely insufficient to identify conservation priorities in degraded grassland ecosys- tems. As fertile grasslands were developed less productive par- cels were most likely to escape tillage (Lichthardt and Moseley, 1997). As a result, remaining native grassland patches tend to occur on these least productive soil types, which may not pre- serve the community-level diversity of the original ecosystem. Similar limitations have been noted in the U.S. National Park system, which frequently protects only the least economically valuable native habitat types (Scott et al., 2001; Hansen and Rotella, 2002). Many community types, particularly those of fertile grasslands, are therefore underrepresented in conserva- tion networks (Noss et al., 1995; Wright et al., 2001). The Palouse region of eastern Washington state and adjacent Idaho once supported extensive bunchgrass meadows, or Palouse prairie (Daubenmire, 1942; Tisdale, 1961). The Correspondence: Chris Looney, 1220 Division St. NW, Olympia, WA 98502, USA. E-mail: looneywa@gmail.com Insect Conservation and Diversity (2009) 2, 243–250 doi: 10.1111/j.1752-4598.2009.00061.x Ó 2009 The Authors Journal compilation Ó 2009 The Royal Entomological Society 243