Previous iron isotope studies compared Earth's mantle to meteorite groups using three isotope ratios and concluded that Earth's iron matched CI chondrites — primitive meteorites linked to the outer Solar System. This supported models where Earth accreted from material that drifted inward from beyond Jupiter's orbit. The match was clean, the implication dramatic: Earth might be built from distant rubble.
Hopp, Tian, and Kleine added one more measurement. Including 58Fe — a neutron-rich isotope rarely analyzed because it's difficult to measure precisely — separates Earth from CI chondrites decisively. The match vanishes. Earth's iron is isotopically distinct from every known meteorite group, pointing instead to inner Solar System precursors, possibly from bodies in the innermost disk that were never preserved as meteorites. The result reinstates the conventional model: Earth formed by hierarchical growth among local planetesimals and planetary embryos, not from far-flung drift.
One additional isotope ratio inverted the conclusion. Not refined it, not complicated it — reversed it. The three-isotope comparison had enough degrees of freedom to produce an apparent match that was an artifact of incomplete measurement. Adding the fourth variable collapsed the degeneracy.
The through-claim: in any system where you match observations to candidate explanations, the number of measured variables determines how many false matches survive. Three variables can't distinguish what four variables can, and the researcher has no way to know — from within the three-variable analysis — that a fourth variable would break the match. The adequacy of a measurement set is invisible from inside that set. Every clean match achieved with N variables carries the unresolvable question: would N+1 destroy it?