In addition to their natural satellites, the giant planets possess a surprisingly rich population of moonlets, rings, arcs and diffuse haloes of dust1–6, the dynamics of which involves gravitational interactions7 and radiation and magnetic forces8,9. Recent analyses (by A.B., W. Hubbard, F.R. and B.S.) of isolated events observed around Neptune indicate the existence of an incomplete or at least a highly azimuthally variable ring or ‘arc’ around Neptune. Similar phenomena may be detectable around Uranus. In order to look for yet undetected dark matter around Uranus and Neptune, the best signal-to-noise ratio stellar occultations have been reanalysed. To be confirmed, an ‘event’ should be observed with more than one telescope on the same site since the Fresnel scale (see below) is about 2–3 km at the level of Uranus and Neptune, and it should also be observed from widely separated observatories (100–1,000 km) to have an idea of the spatial extent of the occulting object, and to distinguish between satellites and ring-like structures. Moreover, theoretical models of diffracting bars or circular objects provide useful tests to reject some isolated events as being incompatible with real and celestial events10. We report here an investigation of the surroundings of Uranus and Neptune using stellar occulations observed at the European Southern Observatory (ESO) on 22 April 1982 and at the Canada France Hawaii Telescope (CFHT) on 15 June 1983, respectively. The high signal-to-noise ratio of these observations sets stringent limits on the presence of matter around these planets.