The study of the ionosphere over the sunrise period is necessary for an understanding of the vtiriations in layer structure with time and has been a topic of research of many workers. On the whole these investigations have been restricted to a study of critical frequency variations with relatively short intervals of a few minutes between successive records, of N-h curves deduced from ionograms with long intervals (15 minutes or so) between successive N-h curves or of continuously monitored single frequency reflections. Not one of the three techniques is entirely satisfactory for a detailed study of ionospheric behaviour over sunrise. The first two do not give a sufficiently clear indication of what happens in the initial stages of layer development, while from the third incomplete data is obtained as to what is happening at a specific electron-density level. For this reason a preliminary investigation of the ionosphere over sunrise was made at Rhodes University during August, 1959. The records were obtained at four-and-a-half minute intervals and scaled by the method of KELSO (1952 ). "Many of t he results were inconclusive but it appeared that records would have to be taken at approximately one minute intervals and reduced to N-h curves by a scaling technique which made full allowance for low-level ionization if useful results were to be obtained. An attempt has been made in this thesis to investigate the behaviour of the ionosphere over sunrise more fully than can be done by the three techniques referred to. A number of observed phenomena are also examined. Part I deals with the theoretical background to ionosphere physics in general and describes the equipment, equipment modifications and experimental procedure. Part II presents the results obtained. The records for a largescale travelling disturbance are analysed. Various observed phenomena are described and discussed. A simple method of obtaining production rates from experimental data is described. The implications of the observed variations of production rates with height and time are discussed. Suggestions for further research and improvement of the methods used arc made in Charter 9.