Michelson–Morley experiment

In nineteenth-century wave theories of light, many physicists supposed that light propagated through a medium called the luminiferous ether. If Earth moved through this medium, researchers expected an experiment comparing light traveling in different directions to reveal a small directional difference in its measured travel time. Albert A. Michelson tested the idea in 1881; he and Edward W. Morley later refined the measurement and reported their 1887 results.

The apparatus was a sensitive interferometer. A beam of light was divided into two paths at right angles, reflected by mirrors, and recombined. The overlapping beams produced interference fringes, whose positions could be measured. The experimenters slowly rotated the apparatus, which was mounted on a heavy stone float supported by mercury, so that the two optical paths changed orientation relative to the expected motion through the ether. Under the simple stationary-ether prediction, this rotation should have changed the relative travel times and shifted the fringes.

Michelson and Morley did not observe the predicted fringe shift at the expected magnitude. Their result, often called a null result, constrained the proposed ether-wind effect and challenged the ether theories they were testing. It did not by itself establish every detail of a new theory of light or prove that all conceivable ether models were impossible. Later theoretical developments, including special relativity, changed how physicists understood space, time and light. Historians caution against reducing that development to a simple claim that this single experiment directly caused Einstein's 1905 theory.

The experiment became a landmark in precision optical measurement and in the history of tests of physical theory. Its design illustrates how an interferometer can reveal tiny differences in optical path length. Michelson interferometers subsequently found uses in other measurements, while the 1887 experiment remains a case study in interpreting a null result: it rules against predictions within tested assumptions and sensitivity, rather than automatically settling every broader theoretical question.