Skip to main content
Dryad logo

Extensive pollinator sharing does not promote character displacement in two orchid congeners

Citation

Joffard, Nina (2022), Extensive pollinator sharing does not promote character displacement in two orchid congeners, Dryad, Dataset, https://doi.org/10.5061/dryad.9cnp5hqkp

Abstract

Pollinator sharing between close relatives can be costly, which can promote pollination niche partitioning and floral divergence. This should be reflected by a higher species divergence in sympatry than in allopatry.

We tested this hypothesis in two orchid congeners with overlapping distributions and flowering times. We characterized floral traits and pollination niches and quantified pollen limitation in 15 pure and mixed populations, and we measured phenotypic selection on floral traits and performed controlled crosses in one mixed site.

Most floral traits differed between species, yet pollinator sharing was extensive. Only the timing of scent emission diverged more in mixed than in pure sites, and this was not mirrored by the timing of pollinator visitation. We did not detect divergent selection on floral traits. Seed production was pollen limited in most populations, but not more severely in mixed than in pure sites. Interspecific crosses produced the same or a higher proportion of viable seeds than intraspecific ones.

The two orchid species attract the same pollinator species despite showing divergent floral traits. Yet, this does not promote character displacement, implying a low cost of pollinator sharing. Our results highlight the importance of characterizing both traits and ecological niches in character displacement studies.

Methods

To assess whether divergence in floral traits and pollination niches between Gymnadenia conopsea and G. densiflora was higher in mixed than in pure sites, in 2019 we measured floral traits (phenology, morphology and floral scent) in 15 populations, located in six pure G. conopsea sites, three pure G. densiflora sites, and three mixed sites, where populations of both species were sampled. In 2020, we characterized pollination niches and quantified pollen limitation in 10 of those 15 populations (located in three pure G. conopsea sites, two pure G. densiflora sites, and three mixed sites, where populations of both species were sampled except in Hörninge, where too few G. densiflora plants were detected this year). Finally, in one population (Ismantorp), we performed controlled crosses to evaluate the strength of post-pollination barriers between the two species.

Phenology data: We visited each population every two days to record flowering start (i.e. day on which the first flower had opened) and end (i.e. day on which the last flower wilted; not recorded in G. densiflora). Note that no phenology data could be collected in Ryd.

Morpholoy data: On each plant, we measured one of the lowermost flowers for corolla length, corolla width and spur length (distance from corolla to spur tip) to the nearest 0.1 mm using a digital caliper. At the end of the reproductive season, we measured plant height and we counted the number of flowers.

Floral scent data: We extracted floral scents emitted by each plant twice - once during the day and once during the following night - using dynamic headspace adsorption. Scent samples were analyzed by Gas Chromatography-Mass Spectrometry. We estimated emission rates of individual compounds using the peak area of the internal standard and the total emission rate by summing emission rates of individual compounds. The total emission rate was then divided by the number of open flowers to estimate the flower emission rate. Scent composition was studied using relative amounts of individual compounds (expressed in percentages of the blend).

Pollinator data: We recorded diurnal and nocturnal pollinator visits in each of the ten populations for four to five days and nights during the flowering peak. For each pollinator visit scored on the videos, we identified the pollinator species and we recorded the duration of the visit and the number of flowers probed by the pollinator. We then calculated the pollinator visitation rate by dividing the number of visits that a plant received on a given day or night by the recording time. In addition, we patrolled for three days and nights in each population to catch flower visitors. Insects that were observed feeding from Gymnadenia flowers were caught with a hand net, killed by freezing and identified to the species level and the number of pollinia they carried was recorded.

Pollen supplementation experiment: In ten populations, we randomly assigned 30 plants to an open- or a supplementally hand-pollinated treatment (N=15 per treatment). We visited each population every two days and performed hand pollinations continuously as flowers opened using pollinia from unmarked plants. At the end of the reproductive season, we counted the number of fruits produced by each plant and we weighed three fruits per plant to estimate mean fruit mass. We then estimated female reproductive success as: number of fruits × mean fruit mass (i.e. total fruit mass).

Crossing experiment: In one popualtion, we selected 30 plants (15 of each species) and randomly assigned three flowers per plant to each of the following treatments: pollination using either (i) self-pollen, (ii) outcross, conspecific pollen, or (iii) heterospecific pollen. Plants were bagged before and after the crosses to exclude pollinators. After fruit collection, seeds were observed under a dissecting microscope as described in Söderquist et al. (2020) to estimate seed viability as the proportion of seeds bearing an embryo.

Statistical analyses applied to these datasets are detailed in the related article.

Usage Notes

Sites:

ALB: Alböke (pure G. conopsea site)

GRA: Gråborg (mixed site)

HOR: Hörninge (mixed site)

ISM: Ismantorp (mixed site)

OST: Österskog (pure G. densiflora site)

RYD: Ryd (pure G. conopsea site)

SAN: Sandby borg (pure G. conopsea site)

SEG: Segerstad (pure G. conopsea site)

STO: Störlinge (pure G. conopsea site)

TOR: Torpmossen (pure G. densiflora site)

TRI: Triberga (pure G. conopsea site)

VAN: Vanserum (pure G. densiflora site)