Bloom timing gave no common link to next-year missing lilac reports in four US samples in 1965-2007

We tested whether bloom date predicts a missing first-bloom report in the next year after plant, year and seasonal temperature controls. We used 8,818 observed lilac years in four eastern and western US samples in 1965-2007. Slopes per ten days later bloom were +0.35, -2.41, +0.08 and -8.64 percentage points. Three four-test adjusted intervals included zero. The later western interval was -14.53 to -2.75 points. That estimate changed to +0.65 points when plants needed ten records instead of five. All four intervals include zero with an added sum-of-covariances calculation. The four samples did not give a common link. This does not show random missingness. Missing reports can reflect observation gaps, flowering failure, plant loss or archive changes. Selected plant histories, unknown visits, different regional plant types, coarse weather and few state clusters limit inference. The package gives a raw-data build, a follow-up register, source hashes, code, all results, a figure and an independent coefficient check.

phenology · lilac · observation bias · missing data · reporting loss · temperature