February-March warmth had negative butterfly-minus-swallow sighting slopes in four Japan samples in 1953-2020

We tested whether February-March warmth changed the first butterfly sighting date relative to the first swallow sighting at the same Japan station. Four region and period models used 4,077 paired station-years at 69 stations in 1953-2020. Models controlled station and report-mark pairs, January temperature and a linear time trend. All four gap slopes were negative, from -1.70 to -2.22 days per C. In southern Japan, the early slope was -1.70, with a four-test interval of -2.78 to -0.62. The late slope was -2.00, with an interval of -3.57 to -0.44. Both northern corrected intervals included zero. Station trends, equal station weights and adjacent-year report checks kept both southern intervals negative. All four corrected intervals included zero with shared year effects. A five-year time-block check made the late southern interval include zero. The pattern agrees with a stronger local warmth response in recorded butterfly dates, but does not identify a biological cause. Unknown search effort, abundance, missing reports, coarse weather and weather windows that extend past sightings limit inference. We do not test pollution, true arrival or emergence, harmful mismatch, or causal climate-change effects. The package includes source hashes, the full panel, all models, code, figures and independent numerical checks.

phenology · temperature · observation bias · Japan · butterfly · barn swallow · paired records · ecological timing