The main features of the Gunn effect can be accounted for by the transferred electron model of Ridley and Watkins, which predicts a bulk differential negative resistance and subsequent domain formation if electrons can be transferred sufficiently rapidly from the lowest conduction band minimum to lower-mobility subsidiary minima. Experimental results for n-GaAs in verification of this model are presented. In the longer samples the current-time Waveform consists of sharp spikes separated by flat Valleys, as expected from the motion of domains. The voltage across the domains is found to scale with sample length as predicted; the value of the electric field inside the domain is estimated to be

V/cm, while the field outside is about 1500 V/cm. Gunn effect oscillations have also been observed in n-CdTe; resistance vs hydrostatic pressure experiments: show that the transferred electron model is a reasonable explanation for this material as well. Finally, it will be shown that the absence of an instability in n-InSb and n-InAs is also consistent with the model. However, the occurrence of stable amplification and some of the properties of the oscillators near threshold seem difficult to understand in terms of a bulk differential negative resistance model. A possible explanation based on transferred electron effects in the positive resistance region will be discussed.