Collector-label controls retained less than 15 percent of time variation in four adjusted herbarium samples from 1941 to 2017

We tested repeat collector-label support in a public flower pigment archive. All 283 Australian source records lacked collector names. Four Eurasian and North American samples had 905 complete known-label records in 1941-1985 and 1986-2017. After taxon, month and location controls, added taxon-by-name-prefix effects retained 0 to 14.7 percent of residual time variation. Full-label controls retained 0 to 1.4 percent. A simpler taxon-only base retained 2.0 to 10.5 percent under name-prefix controls. These are exact design properties, not biological effect estimates. Collector names are unverified, collection numbers can split labels, prefixes can merge people, and missing records are selective. The results show limited support for one collector-control test. They do not show that collectors caused pigment change or that climate had no effect. Code, aggregate results, source hashes and numerical checks permit reproduction.

herbarium · collector labels · observation bias · floral pigmentation · climate change · ozone · identifiability