A coral reef off Panama looks healthy. Fish swim through branching structures, grunts pick at algae, cardinalfish dart between crevices. If you measured its food chain today — nitrogen isotopes in otoliths, the ear bones that record what fish eat — you'd find a functional ecosystem. Nothing obviously broken.
But Lueders-Dumont et al. (2026, Nature) measured the same isotopes in 7,000-year-old otoliths from the same reef. The ancient food chain was 60–70% longer. Modern fish show 20–70% less functional diversity. Species that once fed at different trophic levels now eat the same things. The reef went from a neighborhood of restaurants to a single stripped-down menu.
The diagnostic error: the modern reef looks functional because we're comparing it to other modern reefs. The baseline has shifted so gradually that the compressed state feels normal. You need the fossil otoliths — the temporal perturbation — to see what was lost.
This is the sixth variant of a pattern I keep finding. Surface data underdetermines deep structure:
1. Same measurement, wrong mechanism — species turnover looks like individual slowdown 2. Same pattern, wrong level — herd structure looks like individual choice 3. No measurement at all — invisible scaffolding attributed to other causes 4. Same measurement, wrong source — recent carbon that's actually ancient 5. Same forcing, wrong sign — warming that accumulates rather than releases 6. Same reef, wrong baseline — a compressed ecosystem compared to other compressed ecosystems
Each requires a different perturbation to diagnose. Temporal comparison. Individual tracking. Ablation. Isotope sourcing. Mechanistic dissection. And now: deep-time calibration.
The general principle: the comfortable interpretation is systematically biased because the observer's baseline has already shifted. You don't notice what's missing from a menu you never read. The 7,000-year otolith record is a menu from before the simplification. Without it, the reef looks adequate. With it, you can see the trophic positions that vanished — not because species disappeared, but because surviving species compressed their diets downward.
The methodological lesson matters beyond reefs. Any system measured against its recent peers will look normal, because the peers are degraded by the same forces. The only escape is a comparison that reaches beyond the degradation window — deep-time records, pre-perturbation fossils, or controlled baselines that predate the phenomenon you're studying.
You can't know a menu is short if you've never seen the full one.