Greenhouse experiment tidy data file

Experimental Setup Our study was conducted in a greenhouse in Ames, Iowa (42°01'42.7"N 93°38'40.9"W) between May 2018 and October 2019. We used four accessions (LA3123, LA1237, LA2655, LA2983) of S. pimpinellifolium collected from the Galapagos, Northern Ecuador, Northern Peru, and Central Peru and obtained from the Tomato Genetic Resource Center (TGRC) at the University of California Davis (Table S1). We chose these accessions (hereafter populations) to obtain a broad geographic spread and because they have known population sizes and high seed viability. We planted seeds in May 2018. Once seedlings reached maturity about three months after planting, we picked four of the most phenotypically distinct individuals from each population as representative genotypes (total of 16 genotypes) and took cuttings from each to make nine clonal replicates per genotype. Experimental plants were grown under ambient light during the spring, summer, and fall, and supplemented with light in the winter to maintain 12 hours of light per day to more closely match their original tropical conditions. The ambient temperature ranged between 21 and 26°C and we watered plants as needed to maintain moist soil (Fig. S1). We fertilized plants twice weekly to reduce nutrient limitation and sprayed plants with insecticide and miticide monthly to prevent white flies and spider mites (Supplemental Information). Plants were placed across two rooms of a greenhouse, one meter apart in a random arrangement with respect to treatment and population. Because S. pimpinellifolium has indeterminate growth, we pruned plants weekly to maintain a height of approximately one meter. Starting in January 2019, we hand-pollinated all flowers weekly to ensure even pollen distribution, maximum fruit set, and reduced selfing. Pollen was manually collected from all genotypes within a population and then pooled and used for hand-pollination of all flowers in that population. Fruit Removal Treatments Beginning in February 2019, we established three fruit removal treatments – high, moderate, and low – to test how plants respond to the rate of frugivory. We randomly assigned three clonal replicates per genotype to each treatment (Figure 1). In the ‘high’ removal treatment, we removed all ripe fruits three times a week to simulate the high level of frugivory we might expect during peak fruiting season with an intact frugivore community. In the ‘moderate’ treatment, we removed approximately half of the ripe fruits present on a weekly basis. In the ‘low’ removal treatment simulating the absence of frugivores, we did not remove any fruits except the subset we used for data collection once every four weeks, equivalent to less than five percent of the total fruits present at a given time. Fruit removal events over the duration of the study for every individual in the high and moderate treatments are shown in Figure S2. Though fruit yield (as estimated by total removal in high removal treatment) was highly variable across individuals (mean = 85, s = 75), all individuals that produced little to no fruit (yield < 15 fruits) over the entire duration of the study were removed from the analysis (n = 15) and these were distributed approximately evenly across treatments. Data Collection Every four weeks beginning April 2019, we located and marked the pedicels of the two most basal racemes on each plant with a flower ready to receive pollen. We then hand-pollinated the first three flowers on each raceme and tagged them. By marking each raceme and pedicel with nontoxic paint, we tracked ripening time and collected these fruits when ripe for morphological trait measurements. We determined fruits were ripe using a combination of color and softness cues (Grumet et al., 1981). When removing each fruit, we used a twisting motion to separate the fruit from the sepal, rather than removing from the weakened point on the pedicel to mimic removal by frugivores (Douglas J. Levey et al., 2006). We measured the following morphological traits: fruit size, fresh mass, seeds per fruit, pulp dry mass, and seed dry mass. Fresh mass and fruit size were recorded on the same day the fruit was picked. Fruit size was recorded as two lengths, the first from the point of attachment to the most distal end, and the second as the longest length perpendicular to the first. To measure the dry mass of the seeds and pulp, the gelatinous sac around each seed was manually removed and both seeds and fruit pulp were put in a drying oven at 170°C for 48 hours or until reaching a constant mass. We could not collect data on the total fruit set because of the pruning required throughout the experiment.