An 80 percent coverage rule left no matched Ringlet and Meadow Brown sites in Wales in 2004-2024

We tested whether low summer rain predicted lower Ringlet growth relative to Meadow Brown at the same monitored sites. We used UKBMS site indices through 2024 and Met Office weather. Each growth record required both species in two consecutive years. A fixed rule required usable records in at least 80 percent of each test period. This retained 55 English sites in 1984-2003 and 66 in 2004-2024, but only one early Welsh site and no later Welsh sites. Prior rain-deficit slopes in the two English samples were +0.123 and +0.043 log units per 100 mm less rain. Both corrected intervals included zero. The one-site early Wales estimate also included zero. A planned four-sample test could not be completed under the coverage rule. Relaxing coverage gave estimates in all four samples, with intervals that included zero. This is a data-support and conditional association result. It does not show that drought is harmless. Missing records, selected sites, coarse weather, detection and short time series limit inference. No pollution or causal climate effect is estimated. The package gives source hashes, paired data, code, checks and offline instructions.

butterflies · rainfall · drought · observation bias · site matching · coverage · UKBMS · null result