Wild bumblebees use both absolute and relative evaluation when foraging

Hemingway, Claire1 ; Pimplikar, Smruti2 ; Muth, Felicity2

Published Jan 02, 2024 on Dryad. https://doi.org/10.5061/dryad.gb5mkkwwx

Data files

Jan 02, 2024 version files 231.02 KB

Abstract

Foraging theory assumes that animals assess value based on objective payoffs, however, animals often evaluate rewards comparatively, forming expectations based on recent experience. This form of evaluation may be particularly relevant for nectar foragers such as bumblebees, where individuals can visit thousands of flowers daily that vary in nectar quality. While many animals, including bees, demonstrate reference-based evaluation in experimental contexts, it is unclear whether this occurs in the wild. Here we asked how daily experience with wildflower nectar influenced wild bumblebees’ reward evaluation. We measured the daily nectar concentration of bee-visited wildflowers (Penstemon spp.), before presenting foragers with conspecific flowers filled with a range of artificial nectar concentrations. We recorded bees’ acceptance of artificial nectar, probability of subsequent visits to flowers on the same plant, and residence time. While bees had a minimum threshold of nectar acceptability that was unaffected by experience, when there was higher-concentration environmental nectar, they were less likely to accept lower-quality rewards on manipulated plants. Bees also visited more flowers and stayed longer on plants with higher-concentration nectar. This study shows evidence for both absolute and reference-based evaluation in wild bees and points towards differences between bees’ behavior in lab- and wild-foraging contexts.

README: Absolute and reference-based evaluation in wild bumblebees

https://doi.org/10.5061/dryad.gb5mkkwwx

There are four annotated datasets: Fielddata.xlsx, field_data_accept.xlsx, field_data_visit.xlsx, nectar_data.xlsx

Description of the data and file structure

In Fielddata.xlsx, acceptance and visitation are coded as proportional data. This dataset was used to generate several of the main figures.

Table of contents for field_data_visit

Column A: bee identity - factor
Column B: day number of experiment - continuous
Column C: date of data collection - factor
Column D: location of data collection - factor
Column E: time interval for data collection - continuous
Column F: time of day of data collection - continuous
Column G: experimental treatment concentration (%) - continuous
Column H: number of florets visited on previous flower - continuous
Column I: average number of florets visited on previous flowers - continuous
Column J: number of florets visited including revisits - continuous
Column K: number of florets filled - continuous
Column L: number of florets visited - continuous
Column M: number of florets visited including revisits - continuous
Column N: number of florets probed - continuous
Column O: number of empty florets probed - continuous
Column P: number of florets where nectar was consumed - continuous
Column Q: residence time on flower (s) - continuous
Column R: species of next flower visited - factor
Column S: average daily nectar concentration (%) continuous
Column T: species identification (factor)
Column U: inter-tegular distance (mm) -continuous

Blank spaces in columns T and U indicate missing data for bee id or size. Columns V through X are the table of contents.

In the two behavioral datasets, field_data_accept.xlsx and field_data_visit.xlsx, each behavioral response coded as a binary variable with a separate row corresponding to each behavioral observation.

Table of contents for field_data_accept

Column A: bee identity - factor
Column B: day number of experiment - continuous
Column C: date of data collection - factor
Column D: location of data collection - factor
Column E: time interval for data collection - continuous
Column F: time of day of data collection - continuous
Column G: experimental treatment concentration (%) - continuous
Column H: visit number - continuous
Column I: acceptance - binary 0/1
Column J: number of florets in which nectar was consumed - continuous
Column K: average daily nectar concentration (%) - continuous
Column L: species identification - factor
Column M: inter-tegular distance (mm) - continuous

Table of contents for field_data_visit

Column A: bee identity - factor
Column B: day number of experiment - continuous
Column C: date of data collection - factor
Column D: location of data collection - factor
Column E: time interval for data collection - continuous
Column F: time of day of data collection - continuous
Column G: experimental treatment concentration (%) - continuous
Column H: number of florets filled - continuous
Column I: number or filled florets visited - continuous
Column J: number of florets visited including revisits to previously emptied florets - continuous
Column K: visit number - continuous
Column L: visit response - binary 0/1
Column M: average daily nectar concentration - continuous
Column N: species identification - factor
Column O: inter-tegular distance (mm) - continuous

Blank spaces in column L and M indicate missing data for bee id or size. Columns N through P are the table of contents.

The last dataset, nectar_data.xlsx, contains daily nectar measurements for each day of the experiment at both locations. All datasets should be annotated with columns labeled.

Column A: date of data collection - factor
Column B: day of data collection - continuous
Column C: hour of data collection - continuous
Column D: nectar concentration measurement (%) - continuous
Column E: location of data collection - factor
Columns F through H are the table of contents.

Links to other publicly accessible locations of the data:

https://doi.org/10.6084/m9.figshare.21802704.v4

Code/Software

The R markdown file, Hemingway_et al.Rmd can be used to run all reported analyses and generate all published figures.

Methods

We collected data in June 2021 at two sites in montane meadow and mixed conifer forests. The first site was in the Humboldt-Toiyabe National Forest (Dog Valley Meadow, California), which was lower elevation and bloomed earlier. The second site was in Tahoe National Forest (Van Norden, California). Experiments were conducted using the four most abundant species of bumblebee (n=180): Bombus vosnesenskii (n=88), B. vandykei (n=16), B. centralis (n=52), B. vancouverensis nearcticus (n=5). We focused on two flower species within the genus Penstemon: P. rydbergii at Dog Valley and P. heterodoxus at Van Norden. Each testing day, we calculated the average concentration of nectar in Penstemon in the study meadow, before offering foraging bumblebees Penstemon plants containing artificial nectar that varied across a range of concentrations. This allowed us to determine whether bees’ acceptance of nectar rewards varied based on their short-term foraging experience. To calculate wildflower nectar concentration, we collected nectar each day from 15-30 plants using 1μl microcapillary tubes. These plants were bagged 24 hours earlier (when unopened) using white mesh organza drawstring bags (10 × 15 cm) to prevent visits from other insects prior to nectar extraction. Once nectar was extracted from all flowers on a plant, we cut plants at the base of the stem for use in behavioral experiments, using different plants for each bee. Following nectar extraction, we filled all previously-emptied flowers on each plant with 1 ml of a known sucrose solution of one of six concentrations (w/w): 0, 1, 5, 10, 25, and 50%. Once the bee landed on the manipulated plant, we recorded several behavioral responses. Namely, we stated the time that a bee landed on the plant, how many flowers she visited (typically via walking between flowers, but in one case the bee flew off the plant and then returned to another flower), whether she accepted or rejected the sucrose solution in each flower, and the time that she departed from the plant.