Premise of the Study – Flowering time may influence pollination success and seed set through a variety of mechanisms, including seasonal changes in total pollinator visitation or the composition and effectiveness of pollinator visitors.

Methods – We investigated mechanisms by which changes in flowering phenology influence pollination and reproductive success of Mertensia ciliata (Boraginaceae). We manipulated flowering onset of potted plants and assessed the frequency and composition of pollinator visitors, as well as seed set. We tested whether floral visitors differed in their effectiveness as pollinators by measuring pollen receipt and seed set resulting from single visits to virgin flowers.

Key Results – Despite a five-fold decrease in pollinator visitation over four weeks, we detected no significant difference in seed set among plants blooming at different times. On a per-visit basis, each bumblebee transferred more conspecific pollen than did a solitary bee or a fly. The proportion of visits by bumblebees increased over the season, countering the decrease in visitation rate so that flowering time had little net effect on seed set.

Conclusions – This work illustrates the need to consider pollinator effectiveness, along with changes in pollinator visitation and species composition to understand the mechanisms by which phenology impacts levels of pollination.

Methodology: Phenology Manipulation

Experimental Design

To test the hypothesis that differences in the timing of flowering alters visitation rate and types of potential pollinator visitors, resulting in lower pollination success and seed set, we manipulated flowering phenology of potted M. ciliata plants. Sixty plants were collected from Rustler’s Gulch in 2012 and overwintered in the ground at RMBL. In 2015, we induced the plants to flower at different times using natural variation in temperature and light found along an elevation gradient in Gunnison National Forest. To inhibit flowering, potted plants were moved to Schofield Pass in early June and placed in a shaded snowbank under a shade-shelter. Each week, 10 randomly selected plants were moved back to RMBL, where the higher light and warmer temperatures at low elevation induced them to flower.

Upon flowering, 10 plants per week were moved from RMBL to a meadow near the original source population. Plants were arranged 30 cm apart into two randomized arrays of five plants, with 2 m between arrays. Arrays were located 50 meters away from unmanipulated M. ciliata populations. Forty plants flowered and were included in the experiment, for a total of four phenology groups spanning four weeks (June 23-July 20).

Pollinator response measurements

Plants in each phenology group were left open to pollination for one week. During that week, we conducted pollinator observations. We tracked pollinator identity and the number of flowers visited during multiple 30-minute observation periods between the hours of 9:00 and 16:00. At the beginning of each observation period, open flowers per potted plant were counted. Visitors were counted as pollinators if they crawled inside the corolla. The order of observations was randomized between the two arrays for a given week. We completed 15 hours of observations per phenology group, except in week four because of lower flower abundance.

The floral abundance among arrays during the first three weeks ranged from 12 to 86, with a mean of 78.27 ± 12 (Mean ± SEM) flowers per array (Fig. 1). The floral abundance in week four was significantly lower than in previous weeks (F3, 39=10.15, P<0.0001) and we observed zero pollinators during our first 15 hours of observations that week. To compensate for the difference in floral display in week four, we increased the number of available flowers by adding cut stems in plant picks to each pot and completed an additional 15 hours of observations. We combined the data from both rounds of observations in our analyses of pollinator visitation rate and pollinator type for week four.

At the end of each week, individual flowers in each phenology group were labeled and bagged with fine mesh jewelry bags (Uline, Pleasant Prairie, WI, USA) to prevent further pollination and loss of seeds. To standardize conditions after pollination exposure, all plants remained in the field until seeds were collected.

Soil moisture and floral trait measurements

Because the goal of this experiment was to assess the how differences in flowering time alone impact pollination and seed set, we also monitored soil moisture and tracked floral traits (corolla width, corolla length, and nectar) that were found to vary with changes in water availability. We monitored the soil moisture in each pot as volumetric water content (VWC) with measurements every week using a 12-cm probe (Campbell Scientific, Edmonton, Alberta, Canada) inserted in the ground at the center of each pot.

For all tagged flowers, we measured corolla width at the opening of the tube and corolla length from the

base of the calyx to a randomly chosen corolla lobe.

Fitness component measurements

To measure deposition of hetero- and conspecific pollen load, stigmas were collected after the corollas fell from the flowers and stigma squashes were made with fuchsin gel (Kearns & Inouye 1993). For an average of 5.2 ± 0.7 flowers per plant, we counted the number of conspecific and heterospecific pollen grains using a compound microscope at 200Í.

Mertensia ciliata may produce up to four one-seeded nutlets per flower. We counted the total number of seeds produced per flower for all labeled flowers (measured as described by Forrest & Thomson 2010). We calculated the average seeds per flower for each potted plant as (number of mature seeds / number of flowers). Mature seeds from tagged flowers were collected in coin envelopes and transported to the University of California, Irvine to be weighed. We calculated mean seed mass for each plant as (mass of collected seeds / number of collected seeds).

References Cited:

Forrest, J. & Thomson, J.D. (2010). Consequences of variation in flowering time within and among individuals of Mertensia fusiformis (Boraginaceae), an early spring wildflower. American Journal of Botany, 97:38-48.

Methodology: Pollinator Effectiveness

To estimate the single-visit pollinator effectiveness of different insect visitors, in 2016 we measured pollen receipt and seed set resulting from single visits to virgin M. ciliata flowers in wild populations at Rustler’s Gulch (June 27-July 12) and Schofield Pass (July 12-29).

Experimental Design

Unopened flowering cymes on individual ramets were bagged with fine mesh jewelry bags to provide a supply of virgin flowers. For each single-visit ramet, two additional ramets in the same clone were bagged to serve as controls. The first control group remained bagged throughout the experiment to serve as a control for self-pollination. Non-production of seeds by plants in the bagged-control group would indicate that despite being self-compatible, M. ciliata flowers are not self-pollinating and therefore require insect pollination. Cymes in the second control group (hereafter open control) were made available to pollinators during observation periods but not observed to be visited, thus serving as a control for missed visits by observers. Each group of three ramets, including the single-visit ramet and two control ramets, belonged to the same M. ciliata clone, and no clone was used for more than one group of three.

Pollinator visit data

Bags were removed from flowering cymes on single-visit and open control ramets during observation periods. We recorded the number of flowers available on single-visit and open control ramets during each observation period. Once a single visit to a single flower was received on the single-visit ramet, the visited flower was marked on the calyx with permanent marker and both the open control and single-visit ramet were re-bagged to prevent further pollination and loss of seeds. For each visitor, we recorded the pollinator identity and number of flowers visited. Only insects that crawled inside of a corolla were counted as visitors. When a single visit was observed to an open control ramet, we re-designated that ramet as a single-visit ramet and marked the visited flower(s).

We monitored 235 marked flowers from 95 ramets that received single visits, as well as 65 open control and 74 bagged control ramets. The most common pollinators, bumblebees (Bombus spp.) and flies (Muscoidea, hereafter flies), accounted for 97.5% of visits, with solitary bees (Osmia spp. 2.1%) and a syrphid fly (Syrphidae 0.4%) making up the rest. We excluded the syrphid fly from our analyses.

Floral trait measurements

For single-visit ramets, we measured corolla width and length of marked flowers, or if marked flowers were withered or had fallen off, we calculated mean corolla width and length from up to five randomly selected flowers on the same ramet. In a few cases, more than one flower was visited on the single-visit ramet. When this happened, we marked the calyx of each visited flower, and used mean trait values (e.g., pollen receipt, seed set, etc.) for these ramets in our analyses.

Fitness component measurements

To measure conspecific and heterospecific pollen receipt, we collected stigmas from marked flowers on single-visit ramets, as well as one randomly selected flower from each open control ramet. Stigmas were collected after the corollas fell from the flowers and stigma squashes were made with fuchsin gel (Kearns & Inouye 1993). For an average of 15.5 ± 3.4 flowers per visitor/control type we counted the number of conspecific and heterospecific pollen grains using a compound microscope at 200x.

For each single-visit ramet, we counted the number of seeds produced per marked flower and, when a ramet had more than one marked flower, calculated mean seed set of marked flowers as (number of mature seeds / number of marked flowers). For each open and bagged control ramet, we calculated mean seed set per flower as (number of mature seeds / number of bagged flowers). Mature seeds were collected in coin envelopes and transported to the University of California, Irvine to be weighed. We calculated mean seed mass per flower as (mass of collected seeds / number of collected seeds). Ramets that failed to set seed because of herbivory or accidental damage were excluded from analyses.

Location of study: Plants used in this experiment were wild M. ciliata populations along Rustler’s Gulch (38° 59.6'N, 107° 0.5’W; 3,009 m. a.s.l.) and Schofield Pass (39°00'54.98'' N, 107° 2'49.40'' W; 3,263 m.a.s.l.).

For all details see:

gallagher&campbell_nonSpatialData_phenologyExperiment.pdf

gallagher&campbell_nonSpatialData_pollinatorEffectiveness.pdf