The factors involved in the study of discrete sources of galactic noise by the Bea interferometer are discussed. Three new forms of Bea interferometer which increase the effectiveness of this technique are described.I. -.INTRODUCTION Owing to the low resolving power of metre wavelength aerial systems of reasonable physical" dimensions, interference methods have been widely used in the study of radio-frequency emission from the Sun and stars. Two forms of interferometer have been employed, the sea interferometer and the two-aerial interferometer. The former depends on interference between the direct ray from a source and the ray reflected from the sea at an aerial situated on the top of a cliff. This system is analogous to Lloyd's mirror in optics. The other, consisting of two aerials spaced many wavelengthS apart along an east-west base line, may be compared to the Michelson interferometer. The relative advantages of the two systems have been previously discussed (Stanley and Slee 1950) but will be briefly reviewed.One inherent advantage of the sea interferometer is that twice the sensitivity is achieved with a single aerial as with two similar aerials in the other system. Further, no interconnecting cables or preamplifiers are required. Probably the most important advantage is due to the "cut-off" of the sea's horizon. The interference pattern commences sharply as a source rises above the horizon, in contrast to a gradual "fading-in" in the two-aerial interferometer. This feature is most useful in resolving two or more close sources.Variable refraction adversely affects the sea interferometer in its use for determining accurate source positions and the curvature of the Earth produces effects which restrict its use for measuring angular widths. The two-aerial interferome~er is not affected by the Earth's curvature and the effects of atmospheric refraction and scintillation are much smaller than in the sea interferometer. The effects of refraction on sea interference measurements of position can be overcome by taking observations over a sufficiently long period and by calibrating the instrument on sources of known position. Scintillations, which are most severe at low angles of incidence, affect the "seeing" of very weak sources.