Ecology, 94(6), 2013, pp. 1307–1316 Ó 2013 by the Ecological Society of America Landscape-scale eco-evolutionary dynamics: Selection by seed predators and fire determine a major reproductive strategy MATT V. TALLUTO 1 AND CRAIG W. BENKMAN Program in Ecology and Department of Zoology and Physiology, University of Wyoming, Laramie, Wyoming 82071 USA Abstract. Recent work in model systems has demonstrated significant effects of rapid evolutionary change on ecological processes (eco-evolutionary dynamics). Fewer studies have addressed whether eco-evolutionary dynamics structure natural ecosystems. We investigated variation in the frequency of serotiny in lodgepole pine (Pinus contorta), a widespread species in which postfire seedling density and ecosystem structure are largely determined by serotiny. Serotiny, the retention of mature seeds in cones in a canopy seed bank, is thought to be an adaptation for stand-replacing fire, but less attention has been paid to the potential selective effects of seed predation on serotiny. We hypothesized that spatial variation in percentage serotiny in lodgepole pine forests results from an eco-evolutionary dynamic where the local level of serotiny depends on the relative strengths of conflicting directional selection from fire (favoring serotiny) and seed predation (favoring cones that open at maturity). We measured percentage serotiny, the abundance of American red squirrels ( Tamiasciurus hudsonicus; the primary pre-dispersal seed predator of lodgepole pine), and several measures of forest structure in Yellowstone National Park, USA. Fire frequency strongly predicted the frequency of serotiny, a pattern that is well-supported in the literature. At sites with high fire frequency (return intervals of ;135–185 years) where fire favors increased serotiny, squirrel abundance was negatively associated with serotiny, suggesting that selection from predation can overwhelm selection from fire when squirrels are abundant. At sites with low fire frequency (return intervals of ;280–310 years), serotiny was nearly universally uncommon (,10%). Finally, forest structure strongly predicted squirrel density independently of serotiny, and serotiny provided no additional explanatory power, suggesting that the correlation is caused by selection against serotiny exerted by squirrels, rather than squirrels responding to variation in percentage serotiny. Key words: American red squirrel; climate change; cross-scale interactions; eco-evolutionary dynamics; fire; lodgepole pine; Pinus contorta; reproductive strategy; seed predation; serotiny; Tamiasciurus hudsonicus; Yellowstone National Park, USA. INTRODUCTION Recently, there has been growing interest in the effects of evolutionary processes on ecological dynamics (eco- evolutionary dynamics; Schoener 2011) and in the potential for intraspecific genetic variation to influence community and ecosystem structure (community and ecosystem genetics; Whitham et al. 2006, Wymore et al. 2011). Theoretical and empirical work has shown that rapid evolution can contribute to ecological change (Hairston et al. 2005), and numerous examples demon- strate that genetic variation in foundation species (i.e., species that have large impacts on ecosystem function and comprise a majority of an ecosystem’s three- dimensional physical structure; Ellison et al. 2005) can cause variation in community and ecosystem structure (Whitham et al. 2006). Nonetheless, it is less clear that these effects are relevant in natural ecosystems, and many existing studies have necessarily been conducted in simplified model ecosystems (Hersch-Green et al. 2011, Schoener 2011). While such studies are important in describing the mechanisms underlying eco-evolutionary dynamics, it is equally crucial to investigate these effects in natural ecosystems. Specifically, unanswered ques- tions include how the strength of evolutionary effects compares to that of ecological drivers in structuring ecosystems, whether these effects occur across broad spatial extents, and how evolutionary and ecological drivers combine to produce complex landscapes. An important question is whether we can identify traits in foundation species that have community- and ecosystem-level consequences, and isolate selective agents driving variation in these traits. Additionally, landscape complexity often arises from multiple spatial- ly structured biotic and abiotic drivers, and biotic interactions in natural communities are often complex (Strauss et al. 2005). It is necessary to demonstrate that eco-evolutionary dynamics are important in the context of this complexity. For example, experimentally dem- Manuscript received 26 November 2012; revised 21 January 2013; accepted 23 January 2013. Corresponding Editor: T. J. Valone. 1 Present address: Dept. 3166, 1000 E. University Ave., Laramie, Wyoming 82071 USA. E-mail: mtalluto@uwyo.edu 1307