R-code and data from: Contrasting effects of a lethal liquid bait on reducing counts of foraging Myrmica rubra and Solenopsis invicta ants
Description
This dataset and R-code are a supplement to a submitted article titled: Contrasting effects of a lethal liquid bait on reducing counts of foraging Myrmica rubra and Solenopsis invicta ants Article abstract: Liquid ant baits show promise for invasive ant management but bait uptake by ants depends on bait characteristics, such as food components, the type and dose of lethal agent, and the bait matrix (e.g., solid, liquid, gel). We recently reported that field deployment of our liquid ant bait [an aqueous solution of sucrose (4.55% w/v), essential amino acids (1% w/v), and the lethal agent boric acid (1%)] reduced counts of foraging workers of the invasive red ant, Myrmica rubra, likely due to colony decline. Here we tested the effect of a modified bait (sucrose content 45.5% and larger bait reservoir) to enhance bait consumption by ants, and to quickly reduce worker counts of M. rubra and of the invasive red imported fire ant, Solenopsis invicta. In two-choice laboratory bioassays, worker ants of M. rubra and Lasius niger (black garden ant) consumed similar proportions of the original ant bait and the modified ant bait, whereas workers of Camponotus modoc (Western carpenter ants) consumed less of the modified bait. In field trials on M. rubra, forager ant counts in the treatment plot – but not the control plot – significantly declined by day 6 post-treatment and then remained low for the duration of the experiment. However, in field trials on S. invicta, forager ant counts were lower in the treatment plot than in the control plot but they did not decline over the course of the experiment, indicating that bait consumption had a limited effect on curtailing forager activity or colony size. The modified ant bait seems to be more suitable for some ant species than others. Whereas the bait was effective in reducing forager counts, and thus likely colony size, of M. rubra, further studies are needed to test the bait’s effectiveness on other invasive pest ants, particularly those that are well-known liquid feeders such as the Argentine ant. We have uploaded our data and scripts as an R studio project. Code used to wrangle data, analyze data and generate plots can be accessed in the project folder by opening the project file. The project contains: ---Data Data files of laboratory experiments to assess ant consumption in lab feeding trials, and worker counts in field baiting trials. ---Output Plots and csv files generated from data analysis. ---Script Scripts of R code used to wrangle data, conduct analyses, and generate outputs.
Files
Steps to reproduce
The full description of the methodology can be found in the paper. Here we describe them in brief, primarily in relation to how the data were collected and used. Data were analyzed in R and R studio. In the script "2025_01_16 Lab feeding trials_individual species_modified vs original" we compared proportional consumption on bait formulations in three ant species. We measured the change in mass on bait tubes fed on by ant colonies ("wi" minus "wf") and simultaneously tubes used to measure passive water evaporation ("wi.evap" minus "wf.evap"). Final consumption values ("Consum") were obtained by subtracting the mass loss due to evaporation from the mass loss of tubes fed on by ants ("w" minus "w.evap"). Consumption values were converted into proportional consumption on a per colony basis, dividing the consumption for each bait treatment by the total consumption for each colony ("prop_consum"). Proportional consumption was then modelled with beta distributed GLMM using the glmmTMB function from the glmmTMB package. Model fit was assessed using the simulateResiduals function from the DHARMa package to generate DHARMa residuals and then plotted to inspect residual patterns. Statistical significance was evaluated using likelihood ratio rest by comparing the full model to an intercept only model using the anova function. In the script "2023_05_29 Field mortality trial" we evaluated counts of two invasive ant species ("EFA" and "RIFA") in response to two bait treatments - a non-lethal control bait and a lethal treatment bait. Separate plots were established for non-lethal and lethal baits with multiple colonies per plot. Counts of ants at individual colonies ("count") were taken once or twice per week at multiple colonies ("colony" or "colony_id") for each species per treatment ("trmt" or "treatment"). We modelled counts with count day and treatment as categorical predictors as well as their interaction. As measurements were repeatedly taken at each colony and that treatment and control plots were separated, we included colonies as a random effect nested within each plot ("site"). Models were fit using the glmmTMB function. EFA models were fit with a negative binomial distribution and RIFA models with a tweedie distribution. Model fit was assessed using the simulateResiduals function from the DHARMa package to generate DHARMa residuals and then plotted to inspect residual patterns. Statistical significance was evaluated using likelihood ratio rest by comparing the full model to reduced models using the anova function.