Exact species-year controls retained less than 30 percent of separate climate information in six floral pigment samples from 1941 to 2017

We tested how much separate ozone and temperature information remains after exact species-year control in 934 exposed-anther herbarium records. Six samples cover Australia, Eurasia and North America in 1941-1985 and 1986-2017. Relative to species-month controls and species-specific linear time trends, exact year controls retain 0-18.04% of residual ozone information and 0-29.48% of residual temperature information. Early Australia retains one common climate axis but cannot separate the two slopes. Later North America retains no climate axis. The other four samples support two slopes, with weak residual support. These are fixed-record design results, not biological effect estimates. Exact year control removes real annual climate differences as well as possible age and collection bias. One nonrandom archive, coarse exposures and small samples limit wider use. Code, a CC0-source panel, source hashes, numerical checks and a figure permit reproduction.

floral pigmentation · ozone · temperature · herbarium · observation bias · identification · species-year controls