Recorded ant catch distance rose mainly within genera in two Harvard Forest blocks from 2008–2011 to 2012–2018

We tested whether ant catches in logged and girdled plots became more similar in later surveys at Harvard Forest. We used HF118 version 35 and one plot pair in each of two blocks. Four early years, 2008–2011, were compared with five survey years in 2012–2018. Profiles used each species code's share of positive taxon–trap records. Mean squared Hellinger distance rose from 0.2214 to 0.3455. The change was +0.1241, with a conditional 95% shared-year resampling range of +0.0526 to +0.1938. Both blocks had positive changes. An added exact genus partition assigned 90.2% of the increase to differences within genera. The between-genus change was small and uncertain. Worker counts, July-only records and fixed positive-trap subsets retained positive estimates. Binary species lists gave an uncertain change. The main comparison crosses a survey-grid change. A check confined to the later grid era gave +0.0302, with a range of −0.0650 to +0.1253, and opposite block signs. Thus, the records show a period contrast but do not establish continued ecological divergence or a treatment effect. The ranges condition on four plots and do not include all detection or grid error. The package has unchanged open inputs, source hashes, code, all selection decisions, results, figures and offline instructions. Independent arithmetic checks passed, and a clean run reproduced all 34 output files.

ants · Harvard Forest · hemlock removal · species composition · taxonomic aggregation · Hellinger distance · reproducibility