A tetra-polar electrical impedance device has been constructed to permit study of gastric emptying using a microcomputer. An alternative current of 100 kHz, 4 mM peak-to-peak, is injected through a pair of surface electrodes on the upper abdomen. The voltage variations picked up across the second pair of electrodes correspond to the variations of epigastric impedance in response to the applied current. The low frequency voltage corresponding to gastric emptying and recording noise is then conditioned and digitised at a sampling frequency of 1 Hz. A parallel 8-bit signal is finally converted to standard serial form and sent, in real time, to a microcomputer via a serial port. In a process of off-line analysis, the emptying trace was extracted from measured traces by taking successive Fast Fourier Transform (FFT) of length 64 points and was then fit by three mathematical models: linear, exponential and Weibull. The rate of emptying was calculated in terms of the time needed to achieve 50% emptying (T 1/2) from the best fit model. A clinical experiment was done in 20 healthy volunteers to investigate the reproducibility of the method and compared to a scintigraphic method. Successive measurements on the same subject gave statistically similar results and were statistically independent. No correlation has been observed between impedance and scintigraphic methods.