The aim of the study is to evaluate and further develop a non-invasive method for the measurement of intragastric movements. Ferrimagnetic particles (gamma-Fe2O3) are ingested within a standard test meal (pancakes), and given a permanent magnetisation by an applied homogeneous magnetic field. Magnetic measurements are performed outside the stomach, using fluxgate magnetometers anterior and posterior, in seven periods of 12 minutes each. Any movements caused by grinding and mixing in the ingesta would result in a misalignment of the magnetic moments, and hence in a decay of the external magnetic field. Twelve healthy male volunteers, aged 26.7 +/- 2.7 years (mean +/- SD), are studied. After an initial phase of low and almost linear flux density decay (32 +/- 19%, from minute 10 to minute 22 postprandially), the mean rotation and decay gradually increases (67 +/- 21%, from minute 100 to 112), and the decay becomes bi-exponential. It is hypothesised that the bi-exponential nature of the decay corresponds to the relative fractions, solid and semifluid, of gastric content. The decay curves often show superimposed 3 cycles min-1 variations, which correlate well with the basic electrical rhythm of the stomach measured by electrogastrography (EGG). It is concluded that these variations are induced by gastric contractions.