Pair-based beetle composition tests retained 73 of 643 flood-comparison collections in England and Scotland in 1993-2015

We tested pair-based taxon concentration in flood-marked and unflagged UK beetle traps. A pair measure corrects finite catch size under random sampling, but requires at least two individuals. Of 6,299 usable events in 643 mixed-status collections, 635 events met that count rule. Requiring both flood states among eligible catches left 371 events in 73 collections. Four collection and trap models in England and Scotland in 1993-2003 and 2004-2015 had flood contrasts of 0.236, -0.396, 0.093 and 0.0003. All four adjusted intervals included zero. Residual flood variation was concentrated in few years, especially in England. A two-year variance estimate was degenerate. These results do not establish unchanged composition. Count-based selection, local habitat, capture activity and recording rules prevent a causal flood interpretation. We test no climate trend or pollution effect. The package gives code, aggregate evidence, source hashes and checks; external records stay outside the bundle.

carabid beetles · flooding · observation bias · catch composition · pair concentration · sample selection