Dataset for: Effects of local open raceme density, patch size, and distance between patches on pollinator behavior responding to floral display size in Salvia nipponica

# Data set and R programs for Murakoshi et al. (2024) Murakoshi, N., Itagaki, T., Oguro, M., and Sakai, S. Effects of floral display size, local open raceme density, patch size, and distance between patches on pollinator behaviour in Salvia nipponica. Scientific Reports 14, 967 (2024). ## Required packages * **car** * **lme4** * **ggplot2** * **interactions** * **egg** ## Data set * **data**: contains data files for the analyses. * **raceme_visit.csv**: data set used for the analysis of raceme visit (Figs. 2-4). Each row represents a record of a raceme for the 3h observation in one day. * **raceme_visit**: number of pollinator visits for the raceme during the 3h of observation. * **open_flower**: number of open flowers of the raceme. * **local_density**: number of open racemes within a circle with radius of 1.5m around the raceme. * **patch_size**: number of ramets in the patch where the raceme was observed. * **nearest_distance**: distance from the nearest patch to the patch where the raceme was observed (m). * **day_id**: ID number of the observation day. * **patch_id**: ID number of the patch where the raceme was observed. * **plant_id**: ID number of the ramet having the raceme. * **raceme_id**: ID number of the raceme. * **year**: year of observation. * **flower_visit.csv**: data set used for the analysis of flower visit (Fig. 5 and 6). Each row represent a record of a flower for the 3h observation in one day. * **flower_visit**: number of visits for the flower during the 3h of observation. * **open_flower**: number of open flowers of the raceme having the flower. * **local_density**: number of open racemes within a circle with radius of 1.5m around the raceme having the flower. * **patch_size**: number of ramets in the patch where the flower was observed. * **nearest_distance**: distance from the nearest patch to the patch where the flower was observed (m). * **day_id**: ID number of the observation day. * **patch_id**: ID number of the patch where the flower was observed. * **plant_id**: ID number of the ramet having the flower. * **raceme_id**: ID number of the raceme having the flower. * **flower_id**: ID of the flower. Thr format is `{raceme id}-{row within a raceme}-{flower ID within a row}`. * **year**: year of observation. * **flower_use.csv**: data set used for the analysis of visited flower ratio (Fig. 7 and 8). Each row represent a record for a visit to a raceme. * **bout**: cumulative number of visits the raceme having the flower received at the time when the record was made. * **visited**: number of visited flowers during the visit. * **not_visited**: number of not visited flowers during the visit. * **open_flower**: number of open flowers for the raceme having the flower at the time of the visit. * **local_density**: number of open racemes within a circle with radius of 1.5m around the raceme. * **patch_size**: number of ramets in the patch where the raceme was observed. * **nearest_distance**: distance from the nearest patch to the patch where the raceme was observed (m). * **day_id**: ID number of the observation day. * **patch_id**: ID number of the patch where the raceme was observed. * **plant_id**: ID number of the ramet having the raceme. * **raceme_id**: ID number of the raceme. * **year**: year of observation. * **seed_maturation.csv**: data set used for the analysis of seed development ratio (Fig. 9). Each row represent a record for a flower. * **mature**: number of mature seeds in the flower. * **aborted**: number of aborted ovules in the flower. * **stem_diameter**: basal stem diameter (mm). * **open_flower**: number of open flowers of the raceme having the flower when the pollinator observation was conducted. * **local_density**: number of open racemes within a circle with radius of 1.5m around the raceme having the flower when the pollinator observation was conducted. * **patch_size**: number of ramets in the patch where the flower was observed when the pollinator observation was conducted. * **nearest_distance**: distance from the nearest patch to the patch where the flower was observed when the pollinator observation was conducted (m). * **day_id**: ID number of the observation day for the pollinator observation. * **patch_id**: ID number of the patch where the flower was observed. * **plant_id**: ID number of the ramet having the flower. * **raceme_id**: ID number of the raceme having the flower. * **flower_id**: ID of the flower. The format is `{raceme id}-{row within a raceme}-{flower ID within a row}`. * **year**: year of observation. * **01_basic_information.rmd**, **01_basic_information.html**: R program used for calculations of statistics described in the paper. * **02_raceme_visit.rmd**, **02_raceme_visit.html**: R program for the analysis of raceme visit (Fig. 2-4). * **03_flower_visit.rmd**, **03_flower_visit.html**: R program for the analysis of flower visit (Fig. 5 and 6). * **04_visited_flowe ...

Pollinator observations Observations of pollinators were carried out on 9 and 7 sunny, warm days during 13 to 27 September 2019 and 16 to 28 September 2020, respectively. On each study day, we observed 4 to 9 racemes of various display sizes growing in patches of various sizes, aiming to observe racemes with 2–8 display sizes from patches of each of the various sizes. Using video cameras (GZ-RX, JVCKENWOOD, Kanagawa, Japan), we recorded pollinator visits to each of the study racemes for 180 minutes during 0900–1300, when pollinator foraging was active. We recorded 4 to 10 racemes on one day using 4–9 video cameras; the numbers of racemes observed was 124 and 56 and the total observation time was 216 h and 168 h in 2019 and 2020, respectively. In 2019, 1 to 6 racemes belonging to the same ramets were simultaneously observed by single video cameras, whereas in 2020, single racemes were observed by single video cameras. In both years, each of the study racemes was observed on only one day. For each raceme, the number of effective visits to its flowers was recorded. We regarded a visit as effective if a pollinator entered a flower and its body touched the anther. Pollinator behaviors approaching flowers but not touching the anthers were not counted, and a revisit to a probed flower within a single bout was counted as a new visit. After observation, we recorded the number of open flowers (display size) for each raceme observed and the number of other racemes having open flowers within a 1.5 m radius from the focal raceme (local open raceme density). A flower was classified as open if its corolla opening was wider than 7 mm, which allows pollinators to enter the flower. For each raceme observed, we counted the number of times a pollinator arrived at the raceme (raceme visits), and the total number of visits to open flowers on that raceme, including multiple visits to the same flower (flower visits) during observation. We also recorded the numbers of open flowers visited or not visited on the raceme during each raceme visit to determine the visited flower ratio among the open flowers in the raceme. Seed development When seeds matured, we counted the number of seeds developed for each flower of each raceme studied to examine the effects of pollinator visits on seed production. We measured the basal stem diameter of the ramet of each raceme as an index of the ramet resource status possibly affecting seed production.

Flowers cluster at various spatial scales, so pollinators use information from multiple scales when foraging in natural plant populations. Little is known about the effects of interactions between scales or their relative strength. We examined bumblebee foraging behaviour in a natural population of Salvia nipponica in 10 and 7 patches in 2019 and 2020, respectively. We recorded within-patch factors (display size of racemes and local open raceme densities) and patch-level factors (patch size and distance from the nearest patch) and analysed their relationships with pollinator behaviour. The numbers of visits per raceme and flower were mainly affected by the interaction of patch size and raceme density; they were higher in locations with lower raceme density in larger patches. The ratio of flowers visited to all open flowers in a raceme during a raceme visit, which relates to a bumblebee’s choice to leave a raceme, was mainly affected by the interaction of display size and local open raceme density; in 2019, it was higher in racemes with smaller display sizes, while in 2020 the strength and direction of the relationship depended on the open raceme density. These results suggest that pollinators relied on the sizes of flower clusters at different spatial scales when visiting and leaving racemes and adjusted their responses to the sizes of flower clusters depending on the distances between clusters. Therefore, it is important to evaluate factors at various spatial scales and their interactions to fully understand pollinator behaviour in natural plant populations.