Generalists circulate and specialists persist: the continuum of division of labor in harvester ant colonies

Description This repository contains the dataset and R analysis code associated with the manuscript Generalists circulate and specialists persist: the continuum of division of labor in harvester ant colonies. The study examines how individual workers in colonies of the harvester ant Pogonomyrmex californicus vary along a generalist–specialist behavioral continuum, and whether position along this continuum is associated with task switching, task persistence, spatial coverage, and worker–worker interaction. The paper shows that workers do not fall into discrete generalist and specialist classes, but instead vary continuously in their degree of specialization. Generalist-leaning workers show broader spatial coverage and higher antennation rates, whereas specialist-leaning workers switch tasks less frequently and persist longer once tasks are underway. The included R script is used to perform all analyses reported in the manuscript and to produce all figures. Specifically, the script calculates individual specialization scores, task-switching rates, spatial coverage, antennation rates, probability of resting after task performance, temporal task-engagement metrics, colony-level task budgets, and colony-level division-of-labor indices. It also fits the mixed-effects models, conducts robustness checks using alternative specialization calculations, performs distributional and correlation analyses of specialization scores, and generates the manuscript and supplementary plots. Contents This archive includes: The behavioral dataset used in the analyses. The R script used to reproduce all analyses and figures. Supporting outputs or files needed to run the analysis workflow, where applicable. Software All analyses were conducted in R. The script uses commonly available R packages for data wrangling, visualization, mixed-effects modeling, and distributional analyses, including dplyr, tidyr, stringr, purrr, readr, data.table, ggplot2, lme4, lmerTest, broom.mixed, ggpubr, patchwork, mclust, and diptest.