From about 1930 on various scattered notes appeared from time to time, pointing out certain radio effects that probably resulted from meteors. However, the real birth of the serious observation of meteors by radio took place on the night of October 9–10, 1946, when the Giacobinid meteor shower returned in considerable strength. This was observed with re-built World-War II radar equipment in both England and the U.S.A., and the results achieved left no doubt concerning the value of the new technique in meteoric astronomy. Figure 1 illustrates two historic examples of meteor echoes recorded in England on this occasion (Appleton and Naismith, 1947; Hey et al., 1947). Since this date the field has expanded rapidly. Useful general summaries may be found in a number of monographs, e.g., Lovell (1954), McKinley (1961).

In recent years, in addition to back-scatter radio methods of meteor studies, the oblique-scattering technique, earlier utilized for purposes of meteor radio-communication, has been more widely used. This has produced results, not only because of the availability of large quantities of experimental data obtained on numerous radio-communication lines, but also due to certain advantages of this ‘forward-scatter’ technique.

Intensive development of the radar method of investigating meteors and connected phenomena in the Earth's atmosphere required studying interaction of radio waves with ionized meteor trails. Approximate solutions of the problem of radio reflection by ionized meteor trails have been obtained in a number of papers (Feinstein, 1950; Herlofson, 1951; Kaiser and Closs, 1952; Manning, 1953, 1963; Eshleman, 1955; Brysk, 1958, 1959).

The analysis of two factors influencing radar-meteor echo durations leads to the conclusion, that (1) the echo duration depends considerably on the train position with respect to the sensitivity contours of the radar; (2) the mean echo duration changes with respect to the radiant motion of a meteor shower. As a consequence of the factors mentioned above, the magnitude function, or the mass distribution function depends significantly on the observational conditions, as well as on the choice of the range and time intervals of the investigated sample.

An analysis of meteor radar rates, meteor end-point heights and radar-echo amplitudes observed over the period 1953–66 indicates a long-term variation in these quantities with highest values being recorded in 1963 near solar minimum. The phenomenon is explained in terms of a solar-controlled variation in air density at the 90- to 110-km level, with highest densities occurring in 1963. This provides an explanation for the anomalous increase in meteor radar rates observed in 1963. The analysis indicates that air density at the meteor burn-out level varies inversely as solar flux. The density variation is thus opposite in phase to that at higher altitudes as deduced from satellite drag measurements.

A program of simultaneous observation of meteors was organized in the period between August 25 and September 3, 1962 at the Ondřejov Observatory. Experienced amateur astronomers, 37 in number, took part in the visual and telescopic observation of meteors according to the following scheme: four visual observers surveyed an area, the centre of which coincided with the axis of the antenna diagram of the Ondřejov meteor radar: elevation 45°, south azimuth. 24 observers used 10 × 80 binoculars with a field of view of 7·3° diameter and kept under control three zones in the altitudes of 30°, 45°, and 60° above the horizon. In each zone, there were four azimuths of the centres of the fields of view, namely 349·5°, 356·5°, 3·5° and 10·5°; each of these areas being observed by two observers. All standard data of meteor statistics were recorded. Time recording was doubled: recorders read the time indicated by light signals on a chronometer, and independently the time was indicated on a film moving in synchronization with the radar recording film. Thus an accuracy in time recording of ±0·5 sec was guaranteed. Unfortunately, the motor of the time camera sometimes caused interference of variable intensity unfavourably affecting the radar data, particularly the radar hourly rates.

Meteor luminosity is frequently used for the computation of meteor mass. This procedure is a rough approximation only if the spectral distribution of the light energy and its changes during the meteor flight are not known. A good-quality photograph of a meteor spectrum includes much more information on the meteor phenomenon than any other method can provide. There is of course a stronger limitation by meteor brightness than for direct meteor photography. Detailed spectral analysis is available for meteors 4 or 5 magnitudes brighter than for those photographed directly. This could be the reason that meteor spectroscopy is not applied to a desirable extent for the study of meteor phenomenon.

The atomic multiplets and molecular band systems identified in six categories of upper-air sources are studied. These total 510 multiplets of 38 atoms and ions, and 54 band systems of 27 molecules. The average excitation decreases as one progresses from the sources resulting from the influx of the faster and smaller particles to those produced by the slower and larger objects.

The effects on photographic meteor spectra produced by an occulting shutter and by the geometric orientation of the trail with respect to the camera are considered. A Perseid meteor photographed with both an occulted and unocculted spectrograph shows that large changes in the relative intensity of spectral lines can be caused by a shutter. The strong wake spectrum of this meteor is studied and a classification of multiplets according to duration and wake intensity is presented. It is emphazised that a strong wake is produced by a coasting gas rather than by particles or droplets. The gas may coast an appreciable distance during the radiating lifetime of some multiplets.

The spectrum of a bright Quadrantid meteor is used to show that long-duration multiplets are strongly enhanced when the orientation of the trail lengthens the effective path length of the wake column, i.e., the meteor is observed near its radiant. The abundance of calcium appears to be above average in this meteor.

Multiple photographs of the λ5577 line of neutral oxygen have been secured showing a decay time in excess of 1·5 sec. The decay is more rapid at lower heights and it appears that the mechanism of excitation to the metastable level operates for an interval of about 0·1 sec.

As more accurate photometry of meteor spectra is achieved it will be essential to give careful attention to all of these effects.

Our program of meteor research has been changed since November, 1963, by introducing spectral photography at the Dodaira branch station of the Tokyo Astronomical Observatory, in addition to the direct photography at the Mitaka Station. Observations at these two stations are basic, and they have been supplemented in case of need by observations at the temporary station at Tatebayashi, situated about 40 km Northeast of Dodaira. Double- or triple-station observations were carried out mainly during the active period of the major streams, such as Quadrantids, Perseids, Leonids or Geminids, with the purpose of the dynamical and physical studies of meteors. The coordinates of these three stations are as given in Table 1.

Five meteoric spectra have been studied that were obtained during the period mid-June to mid-December, 1966. They include spectra of a Perseid, a Geminid, and three sporadic meteors. Four were photographed on Royal Pan emulsion without a filter; the Geminid on Tri X Aerecon with a GG11 filter.

Lines in the spectrum of one of the sporadic meteors indicate that two meteoroidal fragments were involved which diverged with increasing atmospheric penetration and which flared at different altitudes, indicating that the flares were not the result of atmospheric stratification.

The second sporadic and the Perseid were photographed about 2½ hours apart. The sodium D line is strong in the Perseid but does not appear in the sporadic spectrum. If present, it is blended with the diagonal sequence Δv = + 4 of the first positive group of the neutral nitrogen molecule, which appears in both spectra. A line questionably identified as the forbidden line of oxygen at 5577 Å appears faintly in the sporadic. The strength of the N2 sequence in these two spectra, and the absence of the D line in the sporadic, are very unusual. The proximity in time of appearance of these two spectra, and the absence of the N2 bands from the other two sporadic spectra, suggest that a temporary atmospheric condition may be responsible for the strength of the nitrogen radiation.

The Geminid has as its principal feature the Δv = + 3 sequence of the first positive group of N2. We have not previously observed this band in meteors as low in velocity as the Geminids.

On the basis of photographic observations made by the method of instantaneous exposure, the wake and terminal blending phenomena were investigated. The first of these phenomena is explained by fragmentation, and the second by the recombination of meteor atoms and ions. The successive disintegrations of the original meteor body were also studied.

The heights of sporadic telescopic meteors were observed from three stations, distant from each other 9·1, 6·0 and 3·2 km respectively, during the meteor expedition in the summer of 1960 (Kohoutek, 1967). Three teams of four observers gathered individually data on 1187 meteors, together with 1864 drawings, during the period of 7 nights, August 15–27. Altogether, 189 pairs observed by at least two stations were identified from the complete material. The heights of the respective beginning (b), centre (c) or end (e) points of 159 pairs were determined by means of graphic-numerical analysis.

I wish to stress during the 20 minutes available to me that a very open-minded approach is required in making assumptions for a physical theory of meteors.

It is shown that the Earth's magnetic field can have significant effects on diffusion only above about 95 km altitude. It is further predicted that diffusion can be markedly inhibited by the field only if the meteor train axis makes an angle of less than 1° with the field direction. Experimental support for this view is presented.

Small meteor particles are considered, which produce meteors with magnitudes between +3m and +10m to +12m, and for which the thermal conductivity may be assumed infinite. Fragmentation and air cap effects are neglected. The derivation is made taking into account radiation losses and temperature variation during the evaporation time. The temperature variation of the meteoroid is defined in this case by the equation:

The probability distributions of echo durations as measured on a 33 Mc/sec horizontally beamed radar at Kühlungsborn, G.D.R. (54°07′N, 11°46′E) indicate significant differences between echo durations during night- and day-time. These differences are attributed to attachment during night, whereas the strong radiation detachment of electrons during day-time leaves effectively no negative ions. The nocturnal attachment rate found from the difference in the probability distributions is higher than the rates deduced from the aeronomical processes hitherto known in the meteor height range. The probable attachment processes, the rate coefficients and particle densities are listed as far as known.

A model of the head echo is examined, in which, as a source of ionization and recombination, an interaction layer of impinging atmospheric molecules (O2) mixing with reflected ones is assumed. Dissociative recombination is considered as the most probable process. If we suppose that the interaction layer is that arising in nearly free molecular flow with no shock wave present, we obtain, using the observed heights of head echoes, dimensions of meteoroids in the range 10–1 cm for absolute magnitudes between −10 and −2. It is shown that such a layer may be used to explain the high initial recombination coefficient of the order of 10−5 cm3/sec. Using the ionization collision cross-sections measured by Utterback (1963) we obtain for the mean extent of the interaction layer a value 1–100 m for heights 80–130 km, in accordance with radar observations.