AbstractAn account is given of investigations on the location of sources of atmospherics in South Africa with visually‐operated cathode‐ray direction‐finders, extending over a period of fifteen months and employing stations in Johannesburg, Durban and Bulawayo.The accuracy with which a thunderstorm source can be located by this method was found to be ±50 km. for storms at distances less than 750 km. At greater distances the accuracy decreases but is usually less than ±100 km. at 1,500 km.Of the land sources 76 per cent. were identified with thunder‐clouds and 18 per cent. with nimbus cloud.Sources over the oceans were found t o be closely associated with areas of low pressure, 89 per cent of such lows having associated sources. The progress of the lows could be followed by this method for several days. Their associated sources in the majority of cases lay in the warm sector from NNE. to ESE. of the centre of the low.The monthly and seasonal distributions of sources over the land agree with the distributions found by Brooks from thunderstorm data, except that over the Indian Ocean the number of atmospherics sources reaches a maximum in the winter.A discussion is given of the possibilities of the device in meteorology in Southern Africa and of difficulties encountered in the practical day‐by‐day use of the system. A commentary follows by three officers of the Rhodesia Meteorological Service.

The wave form of all atmospherics received at night from sources within 2000 km. can be accurately described as a ground pulse followed by a series of sky pulses produced by successive reflexions between the ionosphere and the earth, thirty such reflexions being frequently recorded. The time separation between the peaks of these pulses is determined by the distance travelled and the height of the layer. The primary pulse emitted by the source is usually a single complete oscillation of period ranging from 50 to 400//sec. A t distances greater than 500 km. the ground pulse and the first sky pulse merge owing to the shortness of the time interval between them . Differences of amplitude, form and phase between pulses can arise from differences in angle of emission from the parent lightning channel. The height of the reflecting layer can be determined within ± 1 km. It ranged from 85-5 to 90-5 km. during two winter months, with a mean of 88-0 km. The distances of the sources as found by analysis of the pulse series were corroborated by independent location with cathode-ray direction-finders. The reflexion coefficient of the layer for the pulses of longer period exceeded 0-80. The velocity of the ground pulse where it can be tested is within 0.7 % of that of light.

The photographic method of studying the lightning discharge by means of the Boys camera has the unique advantage of giving direct information concerning events in the discharge in two dimensions of space and one of time and could be extended if necessary to include the third space dimension. The luminous events which it records are, however, secondary processes, and the primary movements of electrical charge which cause them can only be inferred by an application of ideas gained from the laboratory study of the passage of electricity through gases. The direct study of these primary electrical processes involves the observation of the electric field during the discharge by means of a cathoderay oscillograph. Studies of near lightning by this method have recently been reported by Appleton and Chapman (1937). The method gives information concerning the total electric moment of the cloud charges and requires to be compared with the photographic data before its results can be interpreted in terms of the charges themselves and their movements.