Two-day date delays permit both signs of honeysuckle overlap temperature slopes in four US samples in 1968-1991

We tested how unknown date delay can change the link between spring warmth and shared flowering time in two honeysuckle clones. We reused 1,557 paired site-years from a verified campaign study. We derived sharp bounds for the fitted temperature coefficient while each first or end bloom date can be up to a chosen number of days late. A delay shared by all four dates leaves overlap unchanged. A two-day cap with separate delays permits both coefficient signs in all four region and period samples. A one-day cap permits both signs in the two western samples but keeps both eastern fitted coefficients positive. These ranges hold the observations and temperature design fixed. They exclude sampling uncertainty and do not estimate actual delay. Missing visits, selected complete records and coarse monthly weather prevent causal conclusions.

phenology · honeysuckle · flowering overlap · date delay · partial identification · observation bias