For twenty years, the Crab Pulsar's radio signal has shown a striking zebra pattern — evenly spaced bright and dark stripes that no single mechanism could explain. Plasma diffraction alone produces the wrong spacing. Gravitational lensing alone produces the wrong contrast. Medvedev's solution: both lenses operate simultaneously, and the pattern is their interference.
Reader note (2026-11-17): This essay gives no arXiv identifier (its front matter says arxiv: null); the paper I matched to it is arXiv:2602.16955, 'Theory of striped dynamic spectra of the Crab pulsar high-frequency interpulse' by Medvedev (February 2026). Checked against the full text. The core claim matches: the zebra bands are interference maxima from multiple ray paths, produced by gravitational lensing combined with plasma de-lensing. What the paper says about the competing explanations and its own results differs from the essay's account: it says that models generating multiple harmonics predict an incorrect uniform band spacing and that models invoking propagation effects need unrealistic wave phase coherence and source stability; it describes the pulsar system as equivalent to a Young's double-slit setup with fringe visibility of order unity, in contrast to diffraction; and the pattern was first reported in 2007. Its quantitative results are a plasma density profile proportional to r^-3 inferred from the pattern, and a predicted critical frequency between about 42 and 650 GHz above which the bright interference pattern should give way to dimmer, closely spaced diffraction fringes.
The plasma in the pulsar's magnetosphere acts as a defocusing lens, spreading radio waves apart. Gravity acts as a focusing lens, bending them back together. Where the two effects are in phase, you get a bright stripe. Where they're out of phase, dark. The zebra pattern isn't produced by either mechanism — it's produced by the mismatch between them.
This is a general principle hiding in a specific observation: when two optical effects with opposite signs compete, the residual isn't noise — it's structure. The spacing of the stripes encodes the ratio of plasma density to gravitational curvature. Erase either lens and you don't get half the pattern — you get no pattern at all.
The twenty-year delay wasn't because the individual pieces were unknown. Plasma lensing was understood. Gravitational lensing was understood. What was missing was the recognition that neither piece was the explanation — the explanation lived in the space between them. The pattern that looked like it needed a new mechanism actually needed the interference of two old ones.