With mass raised to 0.75, 44 Portal censuses in 2021–2025 do not show a clear change in rodent compensation from 1997–2010

Small rodents can partly replace the function of excluded larger rodents. We tested whether the low compensation reported after 2010 at the Portal Project in Arizona persisted in later data. We used PortalData 6.117.0 and the same four control and four exclosure plots as the original analysis. Full trapping effort on all eight plots left 44 new censuses from June 2021 to December 2025. An individual energy index was body mass raised to 0.75. Compensation was the increase in small-granivore index on exclosures, divided by the kangaroo rat index on controls. We used a ratio of sums and 10,000 bootstrap draws of plots and 12-census time blocks. New-period compensation was 32.6% (nominal 95% interval 1.1 to 59.8). The primary change from July 1997–January 2010 was −18.7 percentage points (−48.2 to +11.4). The change from February 2010–January 2020 was +6.8 points (−13.9 to +28.4). Neither contrast resolved a change. Time resampling alone gave a decrease from the high period (−34.3 to −5.6 points), which shows the importance of plot uncertainty. The total granivore energy index on exclosures was 54.7% of controls (32.8 to 73.6). Bailey's pocket mouse supplied 23.8 of the 32.6 compensation points. With counts, compensation was 51.6%; with fixed species masses it was 32.5%. The evidence supports incomplete replacement under the mass model in these observed censuses. It does not show that compensation was equal between periods. Limits include one site, four plots per treatment, a long sampling gap, imperfect detection, and a mass-based index rather than measured energy flow. The package includes source tables, hashes, code, bootstrap draws, figures, and a plan written before outcome calculation.

functional compensation · Portal Project · rodent exclusion · community function · body mass · bootstrap · replication