Eight Portal plots did not establish a difference between colony and bait responses to ant removal in 1988–2009

We tested whether colony counts and counts of workers at seed baits gave the same proportional contrast for long-term ant removal. We used PortalData 6.117.0 and eight fixed plots in a rodent-by-ant removal design. The two survey methods shared 19 observed years in 1988–2009. We selected 17 named taxa with comparable colony protocols; 15 occurred in the panel. We averaged years within plots and used plots as the treatment units. The geometric ant-removal response ratio was 0.844 for colonies and 0.594 for bait workers. The primary bait-to-colony ratio of responses was 0.704, with a nominal 95% t interval of 0.445 to 1.112. Thus, the paired test did not establish a difference between methods. All fixed taxon and period checks also had intervals that included equal response. The result concerns one site and only two plots per treatment cell. The intervals assume independent plot errors and equal within-cell variance. Full survey effort is assumed from the methods because no empty-station ledger is available. The package has unchanged source data, provenance, a reusable species panel, all estimates, figures, code, and independent count and numerical checks.

ants · Portal Project · colony counts · bait surveys · experimental ecology · measurement · granivory · reproducibility