Rain slopes for nine butterfly species had no clear positive repeatability across two periods in Britain, 1980-2024

We tested whether species responses to current summer rain repeat across 1980-2001 and 2002-2024. We used nine butterfly species with complete country indices in England, Scotland and Wales. Separate species models controlled temperature, prior summer weather, time and site coverage. Early-late correlations of rain slopes were -0.064, -0.190 and +0.281. All approximate 95% and three-country bootstrap intervals included zero. Removing site coverage, adding a quadratic trend and changing block length did not give a clear common positive relation. Geographic checks also had wide intervals. Noisy slope estimates can hide stable biological responses. These results do not prove that species sensitivity changed. Country weather, selected species, source index error and observation changes limit inference. The package includes exact source snapshots, hashes, code, annual inputs, all slopes and an offline check.

butterflies · rainfall · weather · repeatability · observation bias · regional comparison