Having concluded that the transfer of energy from a meteor fireball (bolide) to create the anomalous sounds had to be by means of electromagnetic radiation, my (Colin Keay) primary task was to locate its spectral region. There were no known instances of radio signals originating from a bolide, however large. Previous searches had failed. The only recourse was to eliminate prospective regions of the spectrum one by one. Only for frequencies from approximately 1 hz to 100 khz was there an absence of proof that no radiation was emitted. The fact that this covered the audio frequency range made Elmer Weaver's hypothesis ring true.

The search for possible mechanisms for generating such radiation led me to consideration of processes occurring in nuclear explosions, where an intense burst of radio emission is produced of sufficient intensity to burn out electronic equipment. Reportedly it may be heard as a "click" by soldiers in bunkers near such a blast. A 1965 paper by J.R. Johler and J.C. Morganstern revealed its amplitude peaking at 12 khz (see amplitude/frequency graph). A perturbation of the geomagnetic field by the blast is one of the mechanisms producing the electromagnetic radiation, and this parallels what happens when a meteor explodes and gives rise to an electrophonic "pop", "phut" or "crack" heard by some witnesses.

At this stage mention should be made of the work of two Russians, V.V. Ivanov and Yu. A. Medvedev, who in 1965 published a paper on the electric field effect of large meteors. If meteors enter the atmosphere at a fairly steep angle they disturb the normal geoelectric field, giving rise to a purely electrostatic discharge whose effects may be audible. Brief "swishes" heard from meteors seen directly overhead have occasionally been reported. This too is a matter meriting careful investigation.

However the knotty problem of how a bolide could generate sustained electromagnetic radiation remained. No processes in the plasma surrounding the bolide or the ionization in its trail were known that could do it. Some new mechanism had to be found. Inspiration arose from Fred Hoyle's sunspot theory in which energy is trapped in twisted magnetic fields to create the spots. What if the Earth's magnetic field was similarly trapped in the turbulent plasma trail behind the bolide, and released when the plasma cooled and the ionisation neutralised itself?

Calculations indicated that such a "magnetic spaghetti" could arise in a turbulent bolide trail. Turbulent conditions only exist when the trail is below a certain height in the atmosphere. And for the effect to be sustained for up to ten or more seconds the bolide must arrive in a shallow trajectory. These restrictions provide the underlying basis for Astapovich's empirical conclusions many decades ago that only bolides in low trajectories give rise to reports of electrophonic sounds.

Confirmation that I was on the right track followed soon after publication of this work in the journal SCIENCE in 1980. Three years later another Russian, a noted expert on the subject, V.A. Bronshten, endorsed the concepts in his treatise Physics of Meteoric Phenomena, and in a paper expanding my calculations he showed that a bolide twice as bright as a full moon could generate well over a megawatt of radio power by my "magnetic spaghetti" process.

It was not until 1990 that in Japan I learned of the first recognised detection of radio waves from a meteor fireball. At Nagoya University Dr T Watanabe showed me the records which proved it: his VLF chart recording, a radio spectrogram by Dr T Okada and the photometry from a carefully timed fireball photograph obtained by K Suzuki and his students. Not long afterwards the feat was duplicated by a Canadian team using a video camera. Both teams were fortunate to capture such rare events. Observational proof that large meteors may produce audio frequency electromagnetic radiation has given vital support to the concepts that I earlier developed and, together with essential laboratory studies, has set the emerging science of Geophysical Electrophonics on a secure foundation.