Pollinator visitation
We quantified the rate of pollinator visitation to each of our focal species using a combination of in-person observations and video. Observations were conducted between 0900 and 1800 hours during non-rainy weather. For each observation period, we selected a group of flowers from 1-3 plants for which we could observe each flower simultaneously. For S. lycopersicum , this ranged from 4-88 flowers per observation session, and for C. pepo , 1-5 flowers. For video observations, GoPro Hero or Hero4 cameras (GoPro, San Mateo, CA) were stationed 15–45 cm in front of open flowers and recorded videos that were later scored for visits. Human observations were conducted in 10-minute observation sessions. Video observations ranged in length from 25 to 55 minutes, depending on the camera used. We conducted 3–6 observation sessions weekly for each species when flowers were open for a total of 405–869 minutes per species per site across the growing season, for a total of 4170 minutes for S. lycopersicum and 3708 minutes for C. pepo . Because human observations and video recordings observed different numbers of flowers for different lengths of time, we standardized our observations to a visitation rate. We calculated visitation rates for each observation period as the number of visits per flower per minute. For S. lycopersicum , we considered a visit to be any instance where a floral visitor landed and remained on the anther cone of a flower for more than one second, in order to exclude visitors that were not actually sonicating the flowers. ForC. pepo , we considered a visit to be any instance where a floral visitor came in contact with the anthers or stigma of a flower. We identified visitors to the lowest taxonomic level possible in the field or from video resolution, often to the species level for bumblebees (Bombus spp.), European honeybees (Apis mellifera ), and squash bees (Peponapis pruinosa ).