The Clock-Ratio Census: Every Factor Hunted in the Clock Sector, Graded Against the Floor Its Own Inputs Set

A frequency ratio between two optical atomic clocks is a pure number with an error bar, and the best of them now carry error bars at a few parts in 10^18. No other laboratory in physics hands you a dimensionless factor that clean. This paper spends the discipline of the factor census (paper D1) on that laboratory. The candidate set was closed before grading: six hunts that the clock literature itself runs - the drift of the fine-structure constant alpha in time, the coupling of alpha to place and to gravitational potential, the drift of the proton-to-electron mass ratio mu, Lorentz-violation coefficients in the clock sector, departures from the gravitational redshift clocks are built to expect, and fifth-force couplings of clock transitions to an ultralight scalar field. Each hunt gets its producing test and the ratio that carries it. Each ratio gets its floor printed as a number: the total stated uncertainty of the measurement that carries the hunt, with the systematic budget read for coherence, not just size. The look-elsewhere count is the published ratio population itself, and it is counted in the open. The result is a ledger. Every hunted factor in the sector is null inside its floor, graded clean, at dated values printed in Section 4. One entry stands above its floor, and it is not a factor: the strontium ratios measured in Boulder in 2021 and again in 2025 disagree by about 14 sigma, roughly 1 x 10^-16, which means at least one campaign's stated floor was wrong about itself. A floor that fails is a laboratory result, not new physics, and this census grades it as one. Ledger: six hunts closed in advance - zero factors above their floors - one floor failure, in the strontium ratios, at 14 sigma - every variation limit in the sector null at its dated size. The paper claims no anomaly and no new physics. Several entries are null results, and a census that cannot print a null has printed nothing.