Visual meteors, due to impinging meteoroids of radius about 1 cm, appear at a rate of a few per hour during non-shower periods. Smaller meteoroids (100 μm – 1 cm) give rise to less bright trails, but are much more abundant. These are usually detected by radars of about 10 m wavelength which, over the past 40 years, have produced a plethora of information concerning mass and height distributions, orbits, etc.

Using such ‘conventional radars’, the peak of the measured height distribution is found at about 95 km, with few meteors detected above 105 km. However, the flux detected is only a few percent of the total flux (a) measured using a large (10 m) optical collector, and (b) expected from a comparison with measurements by satellite impacts and zodiacal light observations (radii < 100 μm). One possibility is that the radars detect few low-velocity (V < ~25 km s-1) meteors since these produce little ionization and thus limit their detectability: the ionizing efficiency of meteors varies as ~ V7/2. In direct opposition, our alternative hypothesis is that the undetected flux is held in a faint high-velocity component which ablates at high altitude. These are not detected by conventional radars because meteor trails have ‘initial widths’ of about 3 m at 105 km; for a radar wavelength of 10 m, components scattered from different regions of the trail therefore destructively interfere, and the probability of detecting any meteor above 105 km is small.

In order to test our hypothesis we have measured the height distribution with a 150 m radar, and we are commencing ancillary observations at 50 m; compared to these wavelengths the initial width is small to at least 140 km. The results show a peak at 105 km with most meteors being above this, significant numbers occurring right up to 140 km. This suggests that the true flux is at least 10 or 20 times that previously deduced, having implications for the number of cornets in the recent past and the balance of material between the smaller bodies in the solar System.