Dataset DOI: 10.5061/dryad.j3tx95xvf

Description of the data and file structure

Dataset Overview

The dataset associated to the manuscript “Influence of worker traits and pollen provisions on bumble bee (Hymenoptera: Apidae) microcolony test performance” compiled the results of four independent bumble bee microcolony experiments. These microcolonies were initiated with five callow worker bumble bees of the species Bombus impatiens, except for the microcolonies initiated with workers of unknown age in experiment four, where our objective was to compare the performance of microcolonies initiated with callow workers or with workers of unknown age side by side (Study_no 4 in file BB_Microcolony Project.csv). Our study specific objectives were to evaluate the contribution of pollen quantity provisioned at the time of nest establishment (defined in our study as the time when the first eggs are laid by the dominant worker), pollen quality, worker size, and worker age on microcolony test performance. We measured the cumulative uptake of sucrose solution and pollen provisions during the 42-day test, as well as recorded the number of surviving workers and produced droned. All the data is compiled in the file ‘BB_Microcolony Project.csv’ and the statistical analysis in file ‘BB_Microcolony manuscript_Statistical Analysis.R’.

Statistical Analysis

Normality of residuals and homoscedasticity were evaluated by conducting Shapiro-Wilk’s and Levene’s tests, respectively. Independent microcolony experiments with two treatment groups (i.e. different amounts of pollen at the nest establishment phase [low, high], microcolonies provisioned with different pollen sources [NC-pollen, CA-pollen], and microcolonies established with workers of different ages [newly emerged, unknown age]) were analyzed with two-sided t-tests, or when assumptions of normality and/or homoscedasticity were not met, with Mann-Whitney U tests. The microcolony experiment with three treatment groups (i.e. microcolonies established with small bees, mixed size bees, large bees), were treated as fixed effects and analyzed with ANOVAs followed by Tukey HSD tests, or when assumptions were not met, with Kruskal-Wallis tests. The four dependent variables analyzed in each microcolony experiment were: 1) cumulative sucrose solution uptake, 2) cumulative uptake of the pollen provision, 3) worker survival, and 4) drone production. The number of workers in a microcolony is limited and fixed, thus the variable was considered discrete and analyzed with Mann-Whitney U tests (experiments with two treatment groups) or with Kruskal-Wallis test (experiment with three treatment groups). The Type I error rate was set at α=0.05 for each statistical test conducted to analyze each dependent variable in the four independent microcolony tests.

Files and variables

File: BB_Microcolony_manuscript_Statistical_Analysis.R

Description: This file contains the code for the statistical analysis and graph (Figure 1)

File: BB_Microcolony_Project.csv

Description: This file contains the complete dataset from four independent bumble bee microcolony experiments.

Variables

• Study_no: The independent microcolony experiment. There were four independent microcolonies to evaluate: 1) Quantity of pollen provisions during nest establishment; 2) Source of pollen; 3) size of callow worker bumble bees to initiate microcolonies; 4) age of worker bumble bees to initiate microcolonies.

• Treatment: The independent variables evaluated in each of the four independent microcolony experiments: 1) Study_no 1 compared the performance of microcolonies in group High_Pollen (cumulative 5-g pollen provisions) with group Low_Pollen (cumulative 2-g pollen provisions); 2) Study_no 2 compared the performance of microcolonies fed through the 42-day test with two different sources of pollen, one from California (CA_Pollen) and another from North Carolina (NC_Pollen); 3) Study_no 3 compared the performance of microcolonies established with callow worker bumble bees categorized as mixed size bees (Mixed_size, average weight: 166 ± 12.0 mg), small bees (Small_size, average weight: 119 ± 4.6 mg), or large bees (Large_size, average weight: 199 ± 15.8 mg) while the were fed with the same pollen source and maintained under the same environmental conditions through the 42-day test; and 4) Study_no 4 compared the performance of microcolonies established with either callow workers (Callow_worker) or with workers of unknown age (Unknown_age).

• Rep: The number of replicates (microcolonies) assigned to each treatment group.

• Pollen: Cumulative uptake of pollen provisions expressed in grams (g). The uptake of pollen provisions was quantified starting at Day 10 and ending at Day 42.

• Sucrose: Cumulative uptake of sucrose solution expressed in grams (g). The uptake of sucrose solution was quantified starting at Day 0 and ending at Day 42.

• Workers: The number of surviving worker bumble bees at the end of the test, on day 42.

• Drones: The cumulative number of adult drones that emerged in each microcolony at the end of the test, on day 42.

Code/software

All statistical analysis were conducted using RStudio version 2024.9.0.375 (Posit team 2024); packages used included “car” (Fox 2019), “dplyr” (Wickham et al. 2023), “ggh4x” (van den Brand 2025), “ggplot2” (Wickham 2016), “ggpubr” (Kassambara 2023), “lme4” (Bates et al. 2015), “lmerTest” (Kuznetsova et al. 2017), “multcomp” (Hothorn et al. 2008), “readr” (Wickham et al. 2024), “survival” (Therneau 2023), “survminer” (Kassambara et al. 2021), “tidyverse” (Wickham et al. 2019).

Each independent bumble bee microcolony experiment was analyzed separately. Data subsets were created for each experiment. Normal distribution and equal variances were evaluated for each continuous variable (Pollen, Sucrose, and Drones) with Shapiro-Wilks and Levene's tests, respectively. Then, the effect of the treatment was evaluated for each of the four measured dependent variables (Pollen, Sucrose, Workers, Drones) with two-sided t-tests if they met assumptions, or with Mann-Whitney U tests if normality and/or equal variances assumptions were not met.