Summer rain gave no clear repeated link to excess joint butterfly index failure in England and Wales in 1985-2024

We tested whether a rise in summer rain predicted a rise in joint butterfly index failure beyond changes in separate failure rates. We used 22,570 matched site-year pairs for Ringlet and Meadow Brown at 2,586 England and Wales sites. Four annual models covered 1985-2004 and 2005-2024. Rain slopes ranged from -0.04 to +1.20 percentage points per 100 mm increase. All four adjusted intervals included zero. Excess joint failure was positive in 69 of 80 country-years. This descriptive dependence can reflect site differences or shared monitoring limits. The test did not establish a repeated rain mechanism. It does not show no rain effect or estimate population loss. Sparse Welsh samples, country weather, absent visit schedules and source-row selection limit inference. The package contains source rows, weather snapshots, hashes, code, all estimates, a figure and numerical checks.

butterflies · observation bias · joint missingness · rainfall · UKBMS · replication · null result