Pollen limitation experiment
We set up a pollination experiment to quantify pollen limitation and its magnitude in each species across our sites. Our experimental setup consisted of 120 plants of each of our two species, grown in a greenhouse from seed, then transplanted into grow bags at each study site (20 plants of each species at each of the six sites). We also had 4 pollen donor plants per species at each site in case we could not collect enough pollen from our experimental plants to cross with all open flowers. Seeds were planted at the WCH greenhouse for S. lycopersicum and the CBG greenhouse for C. pepo on 1 April 2021. Between 29 May and 02 June 2021, seedlings were transplanted into 7-gallon fabric grow bags at each site using organic potting soil (Cowsmo Compost, Cochrane, WI). Plants were fertilized once per month and were scheduled to be watered three times per week, however the growers at each site made decisions about increasing or decreasing watering based on local weather conditions.
For each species, we applied a pollen addition treatment to ten individuals at each site, while the other ten served as unmanipulated, open-pollinated controls (i.e ., ambient pollination). Plants were placed in two to four rows, depending on the space available at each site, and arranged in a paired design, with each pair consisting of one control and one supplemental plant placed side by side, 30 cm apart, to account for potential fine-scale variation within each site (e.g ., shade). Treatments were assigned randomly within each pair. We applied supplemental outcross pollen to the stigmas of all open flowers on pollen-supplemented plants twice weekly, weather permitting, as hand-pollination was not possible during rain. We pollinated C. pepo flowers by collecting pollen from the anthers of open male flowers with a paintbrush and transferring it directly to the stigma of open female flowers on a different plant. We pollinated S. lycopersicum flowers using VegiBee Garden Pollinators (VegiBee, China), vibrating tools that mimic insect sonication. We sonicated each open flower on a supplemental plant, collecting the pollen in a spoon and using it to pollinate all open flowers on another experimental plant by dipping the stigmas into the pollen. At each visit, we counted all open flowers on control and supplemental plants. On supplemental plants, if there were any flowers that had opened and closed since our last visit, causing us to miss the window of opportunity for pollination, we marked these flowers using twist-ties and did not include them in our assessment of pollen limitation.
We harvested fruits twice weekly, counting and harvesting all fruits that were at the stage that local farmers consider ready to harvest. ForS. lycopersicum , this is when the fruits reach a deep orange color and readily come off the vine when tugged gently. For C. pepo , this is when the fruits are between two and four inches in diameter. Half of the fruits harvested from each individual plant were transported to and frozen at the CBG for measures of reproductive success. If an individual produced more than ten fruits in one week, a subset of five fruits were used. We considered both ecologically and agriculturally relevant measures of reproductive success: the proportion of flowers that set fruit, number of seeds produced per fruit, mass of fruit, and size of fruit. We measured fruit size using diameter in the longest direction between the stem end and blossom end. We did not consider size or mass measures for C. pepo because fruits continue to grow larger the longer they stay on the plant, far beyond the optimal size for harvesting, so size and mass reflect when the fruits were harvested. We measured the magnitude of pollen limitation as the difference in reproductive success between control and pollen-supplemented plants for each measure of reproductive success.