Mean capture mass fell in 1978–2002 but had no clear trend in 2003–2025 on two Portal control plots

We tested whether the lack of a clear energy trend reported for Portal rodents through 2002 extends to 2003–2025. We used PortalData release 6.117.0 and two plots that remained free of direct treatments. The sample contains identified seed-eating rodents, 46 eligible years in 1978–2025, and 7,155 capture records. We kept complete 49-trap censuses, preserved sampled zeros, removed exact duplicate rows, and filled 2.1% of capture masses. We compared log trends with intervals that allow temporal correlation. The 2003–2025 energy index trend was −17.5% per decade, with a 95% interval of −42.3% to +17.8%. This result establishes neither decline nor stability within 10% per decade. With mature quality-control data only, the estimate was −2.0% (−23.7% to +25.9%). Mean capture mass fell by 18.0% per decade in 1978–2002 (−23.4% to −12.2%). Its later trend was +3.3% (−4.3% to +11.5%). The slopes differ after adjustment for four tests (P = 0.00015). Between 1978–1987 and 2016–2025, mean mass fell from 50.60 g to 34.85 g. An exact decomposition assigns −16.39 g to changes in species mix and +0.65 g to changes within species. Two-year block intervals are −22.84 to −9.24 g and −0.68 to +1.54 g. The package provides source data, hashes, a duplicate audit, code, figures and sensitivity checks. This is an extension of known community patterns. Limits include one site, two adjacent plots, missing seasons, recent data quality and unknown capture probability. The energy index is based on mass; it does not measure direct energy flow or establish a cause.

community change · body mass · species composition · Portal Project · granivores · energy index · reproducibility · data audit · null result