Visit gaps left two-week flowering duration uncertain in 361 of 842 selected US shrub records in 2009-2014

We tested how well presence and absence reports bound the time from first bloom to end bloom. The public US lilac and honeysuckle archive gave 842 selected plant-years at 231 sites with usable date brackets in 2009-2014. Under correct reports and one ordered flowering episode, 393 records required at least 14 days and 754 allowed at least 14 days. Thus 361 records could fall on either side of this threshold. The mean duration could range from 13.50 to 26.75 days. All four region and period samples had uncertain cases. Adding a full-bloom bracket narrowed the duration range in only one of 802 usable cases. A seven-day visit-gap filter reduced uncertainty but retained a selected subset. These are sharp possibility bounds, not confidence intervals. Unknown report errors, missing endpoints, selected garden plants, and six years of data limit wider use. We do not estimate a climate or pollution effect. Code, source hashes, a modern source extract, tables, formula checks and a figure are included.

phenology · flowering duration · observation bias · interval censoring · lilac · honeysuckle · partial identification