Data from: "Divergent evolution of colony-level metabolic scaling in ants"

<b>Summary</b>This dataset has been gathered from the literature to investigate variation in whole-colony metabolic scaling of ant species regarding their ecological and social traits. Whole-colony metabolic rates and masses of ant species were obtained from published data compilations (Shik et al. 2012, 2014) and references therein, complemented with one study not included in the compilations (Mason et al. 2015). Data reported from earlier sources were directly extracted from them. Then, data on trophic level, fungus-farming, foraging complexity, and caste polymorphism of those species were searched for in published data compilations on those traits (see below). The final dataset comprised 53 ant species in 7 subfamilies, sampled across five countries (Belgium, Panama, South Africa, United States, and Venezuela). Variables were species means, with one to 25 colonies sampled per species.This dataset is composed of three files: "Data_ant_data.csv" contains comparative data on 56 ant species; "Data_ant_tree.txt" is a phylogenetic tree for the studied ant species; and "Data_R_script.R" is an R script to reproduce the original analysis and plots.<br><b>Data assembly</b>Metabolic rate was estimated from CO2 produced in respirometry chambers by single colonies which had been collected from the field and reared under controlled laboratory conditions prior to measurement. Specific rearing conditions and acclimation procedures for measurement varied among studies (for details, see Shik et al. 2012, 2014, Mason et al. 2015), but all colonies had food ad libitum and were active, as indicated by their higher mass-specific metabolic rate compared to isolated individuals or groups. For a few species, the largest colonies were too large for the respirometry chamber, so colonies were fragmented into two or three groups each measured separately. For fungus-farming ants, colonies were measured including their fungus farms. Metabolic rates measured at different temperatures were standardized to 25 °C assuming Q10 = 2, and all measurements were expressed in watts. Colonies used for respirometry were also weighed to determine their mass. However, wet mass was reported for some species and dry mass for others. Because we were interested in living colonies, we considered the former. We used species for which both measurements were available to regress wet colony mass (Y) on dry colony mass (X) (both log10-transformed), and then predicted wet colony mass for those species for which only dry mass was reported (log10Y = -2.80 + 1.22 log10X, r² = 0.99, P &lt; 0.001).<br>We also obtained the dry mass proportion of fungus farms comprising the colony. As the frequency distribution of this proportion was highly bimodal (0 vs. &gt;75%), fungus-farming was coded as a dummy variable (1 for “yes”, 0 for no“”).<br>Trophic level was obtained from a global data compilation including 592 species (Drager et al. 2023). The average trophic level of ant species was estimated from nitrogen stable isotope ratios by assuming trophic level increases with the difference between the ratios of the focal ant and of a baseline (a cooccurring primary producer). In ants, trophic level ranges between 1 (herbivores) to 5 (top predators) (Drager et al. 2023). Direct estimates of trophic level were available for 13 of the analyzed species, so we used genera means for the remaining species. This assumes that congeneric species tend to have similar trophic levels, which is supported by the significant phylogenetic signal of this trait (Pagel’s λ = 0.74; Drager et al. 2023). For Nylanderia guatemanlensis, the mean trophic level of its former genus (Paratrechina) was used, as there was no data for Nylanderia.<br>Foraging complexity was assigned based on published data compilations on ant foraging strategies, covering over 400 species (Beckers et al. 1989, Lanan 2014). Ant foraging can be ranked by increasing integration among foragers. For the analyzed species, the following levels can be recognized: (1) solitary foraging: foragers search for food and collect it individually; (2) tandem running: a forager that returns to the nest after successfully finding a resource guides a single nestmate to that resource; (3) group recruitment: returning foragers guide groups of nestmates to detected resources; (4) mass recruitment: returning foragers lay short-lived pheromone trails from the food to the nest, which then guide nestmates to the food; (5) trunk trails: long-lived trails are laid between stable food sources and the nest that both guide foragers and serve as starting points for new trails. Hence, foraging complexity was coded as an ordinal variable: solitary foraging (1), tandem running (2), group recruitment (3), mass recruitment (4), and trunk trail (5).<br>Caste polymorphism was assigned based on a global data compilation of 8890 species (La Richelière et al. 2022). These authors classified each species as either polymorphic (1) or monomorphic (0), defining polymorphism as any interindividual variation in size or head-to-body allometry reported in the primary literature, either continuous or discrete variation. Species for which there were no published data were assigned the same category as congeneric species (La Richelière et al. 2022). Here, 47 species could have their polymorphism directly assigned from the compilation. For the remaining six species (Brachymyrmex sp., Crematogaster sp., Leptothorax unifasciatus, and three unidentified Solenopsis spp.), we obtained the means of their respective genera as an estimate of the probability of being polymorphic given the known proportion of polymorphism in the genus. Species in genera where this probability was larger than 0.5 were classified as polymorphic (1), and monomorphic otherwise (0).<br>Phylogenetic relationships among the analyzed species were determined using a genus-level supertree (AntWiki 2024). Within-genus relationships were represented as polytomies, and branch lengths were standardized to the same value (one). Species names were checked for validity and updated as required according to an on-line database (AntWiki 2024).<br><b>Data analysis</b>To test the role of ecological and social traits in modifying the colony-level metabolic scaling of ants, we used Generalized Least Squares (GLS) models accounting for phylogenetic autocorrelation. Whole-colony metabolic rate was the response variable, whereas colony mass was the main predictor (both log10-transformed). Further, interaction terms between colony mass and each trait were included. The model assumed a phylogenetic autocorrelation structure given by Pagel’s λ. We also assessed the potential for multicollinearity issues by computing all pairwise phylogenetic correlations between predictors. The model was estimated in two turns. First, the full model was fit, and statistical support (P &lt; 0.05) was assessed for the interaction terms with t tests. Then, a model including only supported terms was created to test for independent effects. Model predictive power was measured as the squared correlation between predicted and observed values.<br>To visualize results, partial residuals were plotted against supported predictors, which controls for the remaining terms in the model. To facilitate interpretation of shifts in metabolic scaling involving quantitative predictors, the final model was used to obtain the scaling exponent and its respective 95% confidence interval (CI95%) for groups defined as above and below the mean of that predictor, while keeping each group and all remaining variables at their own means (or their modes for binary variables). Because metabolic measurements of fungus-farming species mixed ant and fungus respiration, we also ran this analysis excluding fungus-growing species to assess the robustness of the results.<br>To determine whether shifts in metabolic scaling correlated with changes in metabolic level, two complementary measures of metabolic level were used: the scaling intercept (predicted metabolic rate where log10-colony mass equals zero, equivalent to 1 g), which standardizes colony mass but is intrinsically correlated with the scaling slope on a log scale; and the midpoint metabolic rate (predicted mass-specific metabolic rate at the midpoint of the log colony mass range), which does not standardize colony mass but is independent of the scaling slope on log scale. For each trait affecting metabolic scaling, 95% confidence intervals were computed for both the intercept and midpoint estimates of contrasting groups of species.<br><b>References</b>AntWiki. (2024). <i>Phylogeny of Formicidae</i>. Available at: https://www.antwiki.org/wiki/Phylogeny_of_FormicidaeBeckers, R., Goss, S., Deneubourg, J. L., &amp; Pasteels, J. M. (1989). Colony size, communication and ant foraging strategy. <i>Psyche</i>, <i>96</i>, 239–256.Drager, K. I., Rivera, M. D., Gibson, J. C., Ruzi, S. A., Hanisch, P. E., Achury, R., &amp; Suarez, A. V. (2023). Testing the predictive value of functional traits in diverse ant communities. <i>Ecology and Evolution</i>, <i>13</i>(4). https://doi.org/10.1002/ece3.10000La Richelière, F., Muñoz, G., Guénard, B., Dunn, R. R., Economo, E. P., Powell, S., Sanders, N. J., Weiser, M. D., Abouheif, E., &amp; Lessard, J. P. (2022). Warm and arid regions of the world are hotspots of superorganism complexity. <i>Proceedings of the Royal Society B: Biological Sciences</i>, <i>289</i>(1968). https://doi.org/10.1098/rspb.2021.1899Lanan, M. (2014). Spatiotemporal resource distribution and foraging strategies of ants (Hymenoptera: Formicidae). <i>Myrmecological News</i>, <i>20</i>, 53–70.Mason, K. S., Kwapich, C. L., &amp; Tschinkel, W. R. (2015). Respiration, worker body size, tempo and activity in whole colonies of ants. <i>Physiological Entomology</i>, <i>40</i>(2), 149–165. https://doi.org/10.1111/phen.12099Shik, J. Z., Hou, C., Kay, A., Kaspari, M., &amp; Gillooly, J. F. (2012). Towards a general life-history model of the superorganism: predicting the survival, growth and reproduction of ant societies. <i>Biology Letters</i>, 1–4. https://doi.org/10.1098/rsbl.2012.0463Shik, J. Z., Santos, J. C., Seal, J. N., Kay, A., Mueller, U. G., &amp; Kaspari, M. (2014). Metabolism and the rise of fungus cultivation by ants. <i>The American Naturalist</i>, <i>184</i>(3), 364–373. https://doi.org/10.1086/677296