Effects of Inbreeding, Outbreeding, and Supplemental Pollen on the Reproduction of a Hummingbird-pollinated Clonal Amazonian Herb Matthias Schleuning 1,2,8 , Mathias Templin 2,3 , Vicky Huama ´n 4 , Giovana P. Vadillo 5,6 , Thomas Becker 1 , Walter Durka 5 , Markus Fischer 3,7 , and Diethart Matthies 1 1 Department of Biology, Plant Ecology, University of Marburg, Karl-von-Frisch Str. 8, D-35043 Marburg, Germany 2 Department of Ecology, Institute of Zoology, University of Mainz, Becherweg 13, D-55128 Mainz, Germany 3 Community Ecology/Botany, Institute of Biochemistry and Biology, University of Potsdam, Maulbeerallee 1, D-14469 Potsdam, Germany 4 Faculty of Biology, University of San Antonio Abad del Cusco, Av. De la Cultura, Nro. 733, Cuzco, Peru 5 Department of Community Ecology (BZF), UFZ-Helmholtz Centre for Environmental Research, Theodor-Lieser Str. 4, D-06120 Halle, Germany 6 Faculty of Biological Sciences, National University of San Marcos, Av. Venezuela Cdra. 34, Lima, Peru 7 Institute of Plant Sciences, University of Bern, Altenbergrain 21, CH-3013 Bern, Switzerland ABSTRACT Understory herbs are an essential part of tropical rain forests, but little is known about factors limiting their reproduction. Many of these herbs are clonal, patchily distributed, and produce large floral displays of nectar-rich 1-d flowers to attract hummingbird pollinators that may transport pollen over long distances. The aim of this study was to investigate the effects of clonality, cross-proximity, and patchy distribution on the reproduction of the hummingbird-pollinated Amazonian herb Heliconia metallica. We experimentally pollinated flowers within populations with self-pollen and with pollen of different diversity, crossed flowers between popu- lations, and added supplemental pollen to ramets growing solitarily or in conspecific patches. Only flowers pollinated early in the morning produced seeds. Selfed flowers produced seeds, but seed number and mass were strongly reduced, suggesting partial sterility and inbreeding depression after selfing. Because of pollen com- petition, flowers produced more seeds after crosses with several than with single donor plants. Crosses between populations mostly resulted in lower seed production than those within populations, suggesting outbreeding depression. Ramets in patches produced fewer seeds than solitary ramets and were more pollen-limited, possibly due to geitonogamy and biparental inbreeding in patches. We conclude that high rates of geitonogamy due to clonality and pollen limitation due to the short receptivity of flowers and patchy distribution constrain the reproduction of this clonal herb. Even in unfragmented rain forests with highly mobile pollinators, out- breeding depression may be a widespread phenomenon in plant reproduction. Abstract in Spanish is available at http://www.blackwell-synergy.com/loi/btp. Key words: clonality; geitonogamy; Heliconiaceae; hummingbird pollination; Peru; pollen limitation; pollination experiment; rain forest understory. UNDERSTORY HERBS ARE AN ESSENTIAL PART OF TROPICAL RAIN FORESTS (Gentry & Emmons 1987), but little is known about factors lim- iting their reproduction (but see Horvitz & Schemske 1988, Bruna et al. 2004). Many Neotropical understory herbs are clonal and produce large floral displays and ample amounts of nectar to attract hummingbird pollinators (Stiles 1979, Kay & Schemske 2003). Genets of these herbs are usually patchily distributed (Kay & Schemske 2003, Schleuning et al. 2008), and patches occur at low densities in the forest understory (e.g., Bruna et al. 2004). Their scattered inflorescences are visited by trap-lining hummingbirds along foraging routes of up to 1 km length (Linhart 1973, Stiles & Wolf 1979). Although plant–pollinator interactions in the forest understory are an interesting model system for tropical ecology, few studies have experimentally investigated the reproduction of hummingbird-pollinated herbs in Neotropical forests (but see Horvitz & Schemske 1988, Bruna et al. 2004). Plant reproduction can be reduced due to limitations in the quantity or quality of the deposited pollen (Aizen & Harder 2007); both are strongly influenced by pollinator behavior (Ghazoul 2005). Low visitation rates of pollinators may limit pollen quan- tity, reduce stigmatic pollen competition (Snow & Spira 1993), and subsequently plant reproduction (quantity limitation; Knight et al. 2005), while selfing and pollen transfer between closely related plants can reduce plant reproduction due to mechanisms of self-incompatibility or inbreeding depression (quality limitation; Burd 1994, Herlihy & Eckert 2002). In clonal plant species, pollen limitation as a consequence of low pollen quality may be particularly strong (Honnay et al. 2006, Scobie & Wilcock 2009). High degrees of geitonogamy in populations of clonal plants can substantially reduce reproduction (Araki et al. 2007, Honnay & Jacquemyn 2008). Although the clonal life form is widespread among rain forest under- story herbs (Villegas 2001, Schleuning et al. 2008), the costs of clonality for the reproduction of these herbs is not known. Received 16 December 2009; revision accepted 8 April 2010. 8 Corresponding author; current address: Biodiversity and Climate Research Centre (BiK-F), Senckenberganlage 25, D-60325 Frankfurt (Main), Germany; e-mail: mschleuning@senckenberg.de BIOTROPICA 43(2): 183–191 2011 10.1111/j.1744-7429.2010.00663.x r 2010 The Author(s) 183 Journal compilation r 2010 by The Association for Tropical Biology and Conservation