Precision Lorentz tests
The modern experimental program that constrains hypothetical Lorentz-violating extensions of special relativity to parts in 10⁻¹⁸ or better — Hughes-Drever magnetic-resonance comparisons, Kennedy-Thorndike interferometric asymmetries, and modern atomic-clock-comparison searches for any frame-dependence of fundamental physics.
Definition
Precision Lorentz tests are the family of high-precision experiments that constrain hypothetical violations of Lorentz invariance — the symmetry under boosts that lies at the heart of special relativity. The program has three historical strands. Hughes-Drever experiments (1959–1961, refined continuously since) compare the magnetic-resonance frequencies of nuclei in different orientations relative to the cosmic rest frame; any anisotropy in inertia or fundamental couplings would produce a sidereal modulation of the resonance frequency. Kennedy-Thorndike experiments (1932 onward) use interferometers whose two arms have different lengths so that the round-trip time is sensitive not only to anisotropy but also to boost magnitude, complementing the orientation sensitivity of Michelson-Morley.
Modern realisations push these bounds to parts in 10⁻¹⁸ or better using optical-lattice atomic clocks, hydrogen masers in vacuum, and cryogenic sapphire oscillators compared across orientation and boost. The Standard-Model Extension (Kostelecký 1998 onward) provides an effective-field-theory framework parametrising every conceivable Lorentz-violating term as coefficients c_TT, c_XX, c_YZ, etc.; precision Lorentz tests are now constraints on those coefficients. No deviation has been observed in any sector. The experimental program is partially motivated by quantum gravity — some candidate theories (loop quantum gravity, string-inspired scenarios) admit Planck-scale Lorentz violation that would propagate downward as suppressed but observable effects; the null results push that scale below the Planck mass and above. Precision Lorentz tests remain among the most stringent symmetry constraints in physics.