LETTER Forest bolsters bird abundance, pest control and coffee yield Daniel S. Karp, 1 * Chase D. Mendenhall, 1 Randi Figueroa Sand  ı, 2 Nicolas Chaumont, 3 Paul R. Ehrlich, 1 Elizabeth A. Hadly 4 and Gretchen C. Daily 5 Abstract Efforts to maximise crop yields are fuelling agricultural intensification, exacerbating the biodiversity crisis. Low-intensity agricultural practices, however, may not sacrifice yields if they support biodiversity-driven ecosystem services. We quantified the value native predators provide to farmers by consuming coffee’s most damaging insect pest, the coffee berry borer beetle (Hypothenemus hampei). Our experiments in Costa Rica showed birds reduced infestation by ~ 50%, bats played a marginal role, and farmland forest cover increased pest removal. We identified borer-consuming bird species by assaying faeces for borer DNA and found higher borer-predator abundances on more forested plantations. Our coarse estimate is that forest patches doubled pest control over 230 km 2 by providing habitat for ~ 55 000 borer-consuming birds. These pest-control services prevented US$75–US$310 ha-year 1 in damage, a benefit per plantation on par with the average annual income of a Costa Rican citizen. Retaining forest and accounting for pest control demonstrates a win–win for biodiversity and coffee farmers. Keywords Agriculture, agro-forestry, bat, bird, conservation biological control, countryside biogeography, ecosystem services, landscape complexity, natural enemies, tropical forest. Ecology Letters (2013) INTRODUCTION Agriculture is intensifying and expanding rapidly, threatening the substantial biodiversity that persists in tropical farming countryside (Tilman et al. 2011; Fischer et al. 2008; Philpott et al. 2008; Perfecto et al. 2009; Karp et al. 2011; Melo et al. 2013). In particular, removal of natural habitat within and around farms fundamentally alters biological communities (Philpott et al. 2008; Perfecto et al. 2009; Mendenhall et al. 2011; Melo et al. 2013). Native predators of crop pests are particularly sensitive to habitat loss because natural habitat provides predators with alternative prey, refuges during distur- bances, and sites for breeding, hibernation and roosting (Landis et al. 2000; Bianchi et al. 2006; Chaplin-Kramer et al. 2011; Jirinec et al. 2011). Therefore, maintaining natural habitat is hypothesised to improve farm yields by supporting predators of crop pests. Whether landowners maintain natural habitat ultimately depends on a suite of trade-offs largely centred on yield and income. Very few studies, however, have quantified the relationships that deter- mine the outcomes of these trade-offs (Ricketts et al. 2004). Oilseed rape is the only crop in which avoided pest damage has been quan- tified in relation to natural habitat presence on farms (Thies & Tscharntke 1999; Thies et al. 2003; Chaplin-Kramer et al. 2011). Complex predator–prey dynamics and research biases towards tem- perate systems make quantifying changes in pest control difficult (Bianchi et al. 2006; Chaplin-Kramer et al. 2011). Resolving these complexities requires analysing the effects of natural habitat adjacent to crops on the predators of crop pests, the pests themselves and predation rates. Relying on just one measure may obscure underly- ing trends. For example, natural habitat may boost both predators and pests, making predator abundance alone a poor proxy for pest control (Bianchi et al. 2006; Chaplin-Kramer et al. 2011). Our investigation focuses on pest-control services provided by native birds and bats to tropical coffee plantations. Coffee has a retail value of ~ US$90 billion, making it one of the most economi- cally important tropical crops (Vega et al. 2003; Jaramillo et al. 2011). Twenty million households make their livelihoods by its cul- tivation, and ~ 10 million ha in more than 50 countries are dedi- cated to its production (Vega et al. 2003). Recently, traditional ‘shade-grown’ coffee practices have been largely abandoned for ‘full-sun’ practices (Perfecto et al. 2009). Shade tree removal and clearing natural habitat adjacent to coffee plantations fundamentally alters biological communities (Philpott et al. 2008; Williams-Guill  en & Perfecto 2010; Karp et al. 2011; Mendenhall et al. 2011). For example, tree removal affects predators of insects and herbivorous insects, some of which may be damaging pests (Perfecto et al. 2004; Williams-Guill  en et al. 2008; Karp et al. 2011). How such changes in biological communities will ultimately affect pests and agricultural yields remains unclear. Coffee’s most damaging insect pest is the coffee berry borer (Hypothenemus hampeii), a ~ 2 mm beetle native to Africa, where cof- fee originated. Recently, the pest has spread to nearly every major coffee-producing country (Burbano et al. 2011; Jaramillo et al. 2011). Borer infestations can be devastating, with harvest losses greater than 75%, and no failsafe method of control (Vega 2004). Pesticides are largely ineffective because the borer resides within the coffee berry’s seed coat and because borers rapidly evolve resistance (Brun et al. 1995). Furthermore, endosulfan (the primary pesticide, a neurotoxin) poses significant human health and environmental risks (Roberts et al. 2007). While biological control with introduced par- asitoids has had only limited success (Damon 2000), native insecti- vores may help reduce infestations (Kellermann et al. 2008; Jaramillo et al. 2010; Johnson et al. 2010; Larsen & Philpott 2010). 1 Department of Biology, Center for Conservation Biology, Stanford University, Stanford, CA, 94305, USA 2 Copal de Agua Buena, Coto Brus, 8257, Costa Rica 3 The Natural Capital Project, Stanford, CA, 94305, USA 4 Department of Biology, Stanford University, Stanford, CA, 94305, USA 5 Department of Biology, Center for Conservation Biology, Woods Institute for the Environment, Stanford University, Stanford, CA, 94305, USA *Correspondence: E-mail: dkarp@stanford.edu © 2013 John Wiley & Sons Ltd/CNRS Ecology Letters, (2013) doi: 10.1111/ele.12173