Three known examples of coherent emission in radio astronomical sources are reviewed: plasma emission, electron cyclotron maser emission (ECME) and pulsar radio emission.Plasma emission is a multi-stage mechanism with the first stage being generation of Langmuir waves through a streaming instability, and subsequent stages involving partial conversion of the Langmuir turbulence into escaping radiation at the fundamental (F) and second harmonic (H) of the plasma frequency. The early development and subsequent refinements of the theory, motivated by application to solar radio bursts, are reviewed. The driver of the instability is faster electrons outpacing slower electrons, resulting in a positive gradient (df (v )/dv > 0) at the front of the beam. Despite many successes of the theory, there is no widely accepted explanation for type I bursts and various radio continua.The earliest models for ECME were purely theoretical, and the theory was later adapted and applied to Jupiter (DAM), the Earth (AKR), solar spike bursts and flare stars. ECME strongly favors the x mode, whereas plasma emission favors the o mode. Two drivers for ECME are a ring feature (implying df (v)/dv > 0) and a loss-cone feature. Loss-cone driven ECME was initially favored for all applications. The now favored driver for AKR is the ring-feature in a horseshoe distribution, which results from acceleration by a parallel electric on converging magnetic field lines. The driver in DAM and solar and stellar applications is uncertain.The pulsar radio emission mechanism remains an enigma. Ingredients needed in discussing possible mechanisms are reviewed: general properties of pulsars, pulsar electrodynamics, the properties of pulsar plasma and wave dispersion in such plasma. Four specific emission mechanisms (curvature emission, linear acceleration emission, relativistic plasma emission and anomalous Doppler emission) are discussed and it is argued that all encounter difficulties.