Common Eyebright Euphrasia nemorosa

We read each species’ measured specific leaf area (leaf area per unit dry weight) from the Global Spectrum dataset and band it: thin & fast (high SLA), medium, or thick & tough (low SLA). The global median is about 10 mm²/mg.

Honest limits: A species-mean from pooled measurements — individual plants vary with light and site. A broad strategy signal, not a precise per-plant figure.

We place each species on Grime’s competitor–stress-tolerator–ruderal (C–S–R) triangle using the globally-calibrated “StrateFy” method: leaf size drives the competitor score, dense low-area leaves the stress-tolerator score, and thin high-area leaves the ruderal score. The result is a C/S/R percentage mix and one of 19 strategy classes; we show it in plain words and keep the percentages for the curious.

Honest limits: A species-mean strategy from pooled global leaf measurements — a broad ecological signal, not a precise per-plant or per-site value. Derived, never a measured fact.

A recorded categorical fact: each species is tagged C3 (standard), C4 (heat/water-efficient) or CAM (succulent, night-time CO₂ uptake) — or a facultative combination. We only show a trait card for the noteworthy C4/CAM cases; C3 is the unremarkable majority, kept in the data but not surfaced as a card.

We take the plant’s U.S. county range and look up each county’s long-term climate, then summarise: native rainfall = the median annual rainfall across its counties; heat tolerance = the warm end (90th percentile) of average temperature across its range (mirroring how our cold-hardiness floor uses an extreme). Counties + medians blunt the roadside/observer bias that makes point-level climate unreliable. Needs at least five placeable counties.

Honest limits: A realized, sampling-biased niche (where it has been recorded, not its physiological optimum), and county climate is coarse — large Western counties span deserts and mountains. Derived guide, never a measured fact.

McKenzie et al. assigned each species a flower colour with a GPT-4V vision model over iNaturalist photos. We use the confident tier plus a separately-labelled lower-confidence top-up (mostly inconspicuous green/brown flowers), and render it as a small tint — never as an asserted fact, and kept out of the written synopsis.

Honest limits: AI image inference (~87% expert agreement on the confident tier); a decorative, confidence-tiered indicator, not a measured trait.