2094 Reports Ecology, 85(8), 2004, pp. 2094–2099  2004 by the Ecological Society of America LARVAL PERFORMANCE AND OVIPOSITION SITE PREFERENCE ALONG A PREDATION GRADIENT JOSEF F. RIEGER, 1 CHRISTOPHER A. BINCKLEY, AND WILLIAM J. RESETARITS,JR. 2 Department of Biological Sciences, Old Dominion University, Norfolk, Virginia 23529 USA Abstract. The ability of females to assess habitat quality for offspring can strongly influence individual reproductive success as well as population dynamics and community assembly. However, the specific relationship between oviposition site selection and larval performance is often unclear. Ovipositing females of several species of treefrogs (Hyla) avoid predatory fish. The fish density required to elicit such avoidance and its relationship to larval performance are central to our understanding of how behavior can drive population dynamics and community assembly in complex landscapes. We conducted experiments investigating both oviposition preference and larval performance in pinewoods treefrogs (Hyla femoralis) along a density gradient of predatory fish (0–6 Umbra pygmaea). Female H. femoralis detected and strongly avoided the nonlethal presence of even a single 2-g fish, ovipositing almost exclusively in predator-free controls. In a separate experiment, larval performance largely matched adult preferences; larvae were totally eliminated in all but the lowest fish density. Given a landscape of breeding sites of varying risks, ovipositing H. femoralis behaviorally partition available habitats into those with and without fish, largely matching the associated fitness consequences. Key words: anurans; ephemeral ponds; female preference; fish; habitat selection; Hyla femoralis; ideal free distribution; larval performance; oviposition site selection; predation. INTRODUCTION Oviposition site selection is a form of habitat selec- tion in which a female chooses habitat for her offspring, rather than for herself (Rausher 1983, Resetarits and Wilbur 1989, Blaustein 1999). A female’s choice of where to place her eggs determines the quality of her offspring’s environment and ultimately dictates larval performance and adult fitness, since breeding sites di- rectly affect resource quantity and quality, levels of predation and competition, and the abiotic environment experienced by offspring (Rausher 1983, 1993, Singer 1984, Resetarits and Wilbur 1989, Thompson and Pell- myr 1991, Renwick and Chew 1994, Resetarits 1996, Blaustein 1999, Smith et al. 2000, Murphy 2003a). Habitat selection is presumed to be an adaptive re- sponse to variation in habitat quality (e.g., Fretwell and Lucas 1970, Pulliam and Danielson 1991, Morris 2003), thus theory suggests that adults should choose oviposition sites that maximize offspring performance. Given variation in resources and risk, females should exhibit various oviposition responses depending on the predicted performance of their offspring and their sen- sory capabilities. If larval mortality is a linear decreas- ing function of some factor, females may show a linear dose response. Alternatively, females may demonstrate a threshold response, which may reflect actual offspring performance or derive from detection thresholds in the Manuscript received 18 January 2004; revised 10 March 2004; accepted 15 March 2004. Corresponding Editor: J. Loman. 1 Present address: The Elizabeth River Project, 474 Water St., Suite C103A, Portsmouth, Virginia 23704 USA. 2 Corresponding author. E-mail: wresetar@odu.edu specific cues used to assess habitat quality. The nature and form of performance and preference curves and their degree of correspondence have important impli- cations for how oviposition behavior affects individual fitness, population dynamics, and community assem- bly. Studies of oviposition preference and larval perfor- mance have focused largely on insect–plant relation- ships, where females utilize specific host plants as breeding sites and exhibit a hierarchy in host preference when multiple host plants are present (see e.g., Rausher [1993], Renwick and Chew [1994], and Mayhew [1997] for reviews). Oviposition choices can lead to increased larval growth and survivorship when females avoid low quality host plants, or little relationship may exist between preference and performance, indicating that multiple factors determine this relationship in in- sect-host plant systems (see Thompson and Pellmyr 1991); little data exists for other types of consumer– resource systems. The importance of any specific factor in determining larval performance should manifest itself in the strength of covariance between female preference and larval performance. While factors influencing growth and related measures of performance may certainly contribute to variation in fitness, factors that directly influence offspring mortality should have greater net impacts on fitness and are more likely to drive the evolution of oviposition behavior. Predatory fish simultaneously eliminate and facili- tate numerous taxa in aquatic systems, producing dis- tinct patterns between habitats with and without these