Original article Flowering synchrony and floral display size affect pollination success in a deceit-pollinated tropical orchid Victor Parra-Tabla a,b, *, Carlos F. Vargas c a Departamento de Ecologı´a, Campus de Ciencias Biolo ´gicas y Agropecuarias, Universidad Auto ´noma de Yucata ´n, AP 4-116, Itzimna ´, 97000 Me ´rida, Yucata ´n, Me ´xico b Estacio ´n Biolo ´gica de Don ˜ana, Consejo Superior de Investigaciones Cientı´ficas, Avda. de Marı´a Luisa s/n Pabello ´n del Peru ´ , 41013 Sevilla, Spain c Laboratorio de Variacio ´n y Evolucio ´n, ENCB-IPN, Prolongacio ´n de Carpio y Plan de Ayala s/n, Col. Casco de Santo Toma ´s, CP 11340, Mexico article info Article history: Received 30 September 2005 Accepted 14 February 2007 Published online 6 April 2007 Keywords: Binary logistic regression Deceit-pollination Female success Fluctuating natural selection Frequency-dependent selection Male success Multilevel selection Pollen limitation abstract Due to frequency-dependent negative selection, a strong relationship between reproductive phenology traits and pollination success is expected in deceit-pollinated species. This paper assesses the effects of floral display size on both female (fruit production) and male (pollen removal) pollination success in a population of the deceit-pollinated tropical orchid Myrme- cophila christinae during two consecutive years (1998–1999). Low pollen removal (w9% of total flowers) and fruit production values (w3% of total flowers) were recorded during both years. As expected, binary logistic regressions showed a significant negative effect of floral syn- chrony, and a positive effect of floral display size on both male and female success, although these effects varied across years. Pollination rates in the field and in hand pollinations sug- gest a doubling in pollinator abundance between years. Results suggest that floral display size and flowering synchrony are of adaptive value for M. christinae. However, between- year fluctuations might indicate that reproductive phenology traits in deceit-pollinated species undergo fluctuating selection regimes among years and are probably linked to short-term changes in environmental (abiotic and biotic) conditions. ª 2007 Elsevier Masson SAS. All rights reserved. 1. Introduction Flowering phenology is considered a key characteristic in plant reproductive biology as it influences many mutualistic and antagonistic interactions and thereby impacts on the suc- cess of plant reproduction (Ratchke and Lacey, 1985). Further- more, previous studies have found strong relationships between traits associated with reproductive phenology and fruit and seed production, suggesting that these traits are of adaptive value (e.g. Augspurger, 1981; Dieringer, 1991; Wide ´n, 1991; Ollerton and Diaz, 1995). Nevertheless, pollinator-mediated selection on flowering phenology has been questioned for different reasons. First, doubts about the adaptive value of phenology patterns have been raised in part because of: (a) a lack of a significant rela- tionship between phenology traits and reproductive success * Corresponding author at: Departamento de Ecologı´a, Campus de Ciencias Biolo ´gicas y Agropecuarias, Universidad Auto ´noma de Yucata ´n, A.P. 4-116, Itzimna ´ , 97000 Me ´ rida, Yucata ´ n, Mexico. Fax: þ999 9 42 32 06. E-mail address: ptabla@tunku.uady.mx (V. Parra-Tabla). available at www.sciencedirect.com journal homepage: www.elsevier.com/locate/actoec 1146-609X/$ – see front matter ª 2007 Elsevier Masson SAS. All rights reserved. doi:10.1016/j.actao.2007.02.002 acta oecologica 32 (2007) 26–35