Journal of Tropical Ecology (2005) 21:541–547. Copyright © 2005 Cambridge University Press doi:10.1017/S026646740500266X Printed in the United Kingdom The cost of nectar replenishment in two epiphytic bromeliads Mariano Ordano and Juan Francisco Ornelas 1 Departamento de Biolog´ ıa Evolutiva, Instituto de Ecolog´ ıa, A.C., Km 2.5 Carretera Antigua a Coatepec No. 351, Congregaci ´ on El Haya, Apdo. Postal 63, Xalapa, Veracruz 91070, M ´ exico (Accepted 17 February 2005) Abstract: Animal-pollinated angiosperm plants that replenish removed nectar invest energy that can entail a reproductive cost. Here we investigated whether or not seed production is affected by replenishing nectar in hummingbird-pollinated Tillandsia multicaulis and T. deppeana (Bromeliaceae) in a montane cloud forest in eastern Mexico. These epiphytic plants respond strongly and positively to repeated nectar removal. The female reproductive cost was assessed in manually, cross-pollinated flowers with or without repeated nectar removal. Seed production from experimental flowers was then contrasted with those naturally exposed to pollination and nectar removal. Tillandsia deppeana set the same number of seeds of the same size regardless of whether or not it had to replenish nectar. Seeds were slightly smaller if the result of natural open pollination than from copious hand crossing. In contrast, T. multicaulis set about half as many seeds when it had to replenish than when it did not, indicating a substantial cost to replenishment. There was no difference in seed number of T. multicaulis between open and hand-augmented pollination. The contrasting results for plants that respond strongly and positively to repeated nectar removal suggest that the female reproductive costs of nectar replenishment can range from costly to beneficial, depending on the conditions of the plants and on the species. Key Words: Bromeliaceae, cloud forest, epiphytes, hummingbirds, Mexico, nectar cost, nectar removal, seed production, Tillandsia INTRODUCTION It is broadly assumed that nectar production has a female reproductive cost. Investment in nectar production can represent up to 30% of the energy devoted by a flower in Asclepias quadrifolia (Pleasants & Chaplin 1983) and 33% of the photosynthates assimilated by A. syriaca plants (Southwick 1984), and caused a significant seed number reduction in plants of Blandfordia nobilis subjected to repeated nectar removal (Pyke 1991). However, nectar investment estimated in Pontederia cordata represented only 3% of the floral tissue and it was 45% higher than expected to assure bee pollination (Harder & Barrett 1992). Nectar production was not correlated with vegetative or reproductive traits in Echium vulgare, and hence no costs of nectar production were detected (Leiss et al. 2004). In Prosopis glandulosa var. torreyana the nectarless morph does not differ from the nectar 1 Corresponding author. Email ornelasj@ecologia.edu.mx. morph in terms of vegetative growth (Golubov et al. 2004), suggesting that the cost of nectar production is negligible or that resources allocated to growth are different from those allocated for reproduction. Nectar replenishment can then range from costly to beneficial depending mainly on the identity of the pollinators and changes in their abundance, and the habitat and breeding system of the plant (McDade & Weeks 2004, Ordano & Ornelas 2004). Nonetheless, the effect of nectar replenishment on female reproduction depends on the level at which reproduction is analysed because resource allocation pathways are different at the branch, individual plant, fruit and seed levels (Obeso 2004). Consequently, there is conflicting evidence in terms of reproductive fitness for the broadly assumed expenses in nectar production. There is also some convincing evidence that nectar reabsorption and plasticity for the scheduling of nectar production represent nectar-saving mechanisms in response to stigmatic pollen deposition or pollen removal (B ´ urquez & Corbet 1991, Luyt & Johnson 2002, Nepi et al. 2001), which may result from selective