April-May temperature gave no clear link to the interval from full to end bloom in four US honeysuckle samples in 1968-2008

We tested whether warmer April-May weather was linked to a shorter interval from full bloom to end bloom in two US honeysuckle clones. Four separate network and period models used 6,274 plant-years in selection windows from 1968 to 2008. Models used plant and clone-year effects, March and winter temperature controls, and state and year cluster errors. The four slopes ranged from -0.60 to +0.26 days per degree C. All four adjusted intervals included zero. Weather-grid, plant-trend, flag, and record-length checks did not establish a common shortening result. An exploratory May-only result was negative in the later East but did not repeat in all four samples. Pairing coverage was 89-97%, and the five-year rule retained only 50% of later western pairs. Missing visits, selected plants, sparse late records, different network definitions, coarse weather, and few clusters limit inference. This is a weather association test, not a test of pollution, biodiversity loss, or causal climate-change effects. The package includes paired records, weather extracts, code, source hashes, numerical checks, and reproduction instructions.

phenology · honeysuckle · flowering duration · observation bias · temperature · regional replication · null result