Spring warmth gave no clear common change in lilac interval scatter in four US samples in 1961-2009

We extended an earlier test of spring warmth and scatter in the reported time from first leaf to first bloom. Four US region and period models used 5,920 reporting-unit years in 1961-2009. We added nonlinear mean controls and a two-stage influence calculation checked by case-weight changes. Spring slopes for absolute residual intervals ranged from -0.154 to +0.312 days per C. All four adjusted intervals included zero. Linear mean models, flag removal, positive-gap selection and squared-residual checks also gave uncertain slopes. This confirms a broad null result under different methods and sample windows in the same archive. It is not independent biological evidence. Missing visits, selected long records, fitted residuals, coarse monthly weather and approximate cluster uncertainty limit inference. The package includes exact source hashes, joined records, code, numerical checks and a figure. We do not test pollution, daily extremes, species loss or causal climate effects.

phenology · lilac · conditional scatter · spring temperature · observation bias · null result