The feasibility of the magnetic variation field for electromagnetic investigations in the audiomagnetotelluric frequency range is tested at two separate locations in southeastern Scotland in the vicinity of the Southern Upland Fault complex (Lammermuir Fault). The Lammermuir Fault represents a lateral conductivity contrast in that it separates conductive Carboniferous sediments from resistive Ordovician meta-sediments. For one of the experiments and for the first time, a new generation of instruments (S.P.A.M. Mk. III) could be used for data acquisition. The instrument is operated as a network of geophysical sensors, it uses digital data communication between sites and provides on-line data quality control of all connected channels. For the experiment, we installed a 15-channel configuration of three synchronous AMT sites. All recorded data were re-processed off-line, with a newly developed S.P.A.M. Mk.III-compatible data processing package. The software is based on robust statistics and an integrated database helps to organise all relevant information. Despite some problems with noise, the overall quality of the data is satisfying. Both locations reveal strong vertical magnetic field anomalies. The anomalies are resolved convincingly in frequency and space. Contrary to our original assumptions, the observed anomalies at neither location can be modelled using two-dimensional techniques. However, three-dimensional thin sheet models show that strong current channelling into a surface conductor can explain very large induction arrows. Full three-dimensional modelling of the data is necessary to fit both, real and imaginary parts of the observed induction arrows. The modelling results indicate that a SE/NW striking conductive feature at the surface is duplicated at greater depth and dips towards the south. S.P.A.M. Mk.III is ideally suited to collect data of densely spaced sites for such electromagnetic mapping experiments or possibly three-dimensional data sets in future projects. Vertical magnetic field anomalies can be detected over the whole AMT frequency range and they are more robust against noise than the electric fields. A further development of the geomagnetic variation mapping method at high frequencies is therefore highly recommended.