Higher seed number compensates for lower fruit set in deceptive orchids

Floral deception is widespread in orchids, with more than one-third of the species being pollinated this way. The evolutionary success of deceptive orchids is puzzling, as species employing this strategy are thought to have low reproductive success (less flowers yielding fruits) because of low pollination rates. However, direct measurements of total seed production in orchids - which is a better measure of reproductive success - are scarce due to the extremely small size of their seeds. Here, we quantified seed numbers in 1015 fruits belonging to 48 orchid species from the Pannonian ecoregion (central Europe) and obtained fruit set and thousand-seed weight data for these species from the literature. We used phylogenetic comparative methods to test the hypothesis that deceptive species should compensate for their lower fruit set by having either more flowers, larger seeds or more seeds in a fruit. Similarly to previous studies, we found that deceptive orchids have substantially lower fruits-set than nectar-rewarding ones. Also, we found that deceptive species have more seeds in a fruit but not more flowers or larger seeds compared to nectar-rewarding ones. Based on our results, deceptive species compensate for their lower fruit set by having higher seed numbers per fruit. As a consequence, their seed numbers per shoot do not differ from that of nectar-rewarding ones. Together with other benefits of deceptive pollination (e.g. lower energy expenditure due to the lack of nectar production and higher genetic variability due to decreased probability of geitonogamous pollination), our results can explain why deceptive strategies are so widespread in the orchid family.Synthesis. Our results indicate that deceptive orchids can compensate for their lower fruit set by having more (but not larger) seeds in a fruit than rewarding species. These findings highlight possible ways in which plants can increase their reproductive success in face of pollinator limitation. We emphasize that fruit set in itself is an inappropriate measure of the reproductive success of orchids - the total number of seeds per shoot is a much better approximation.