The magnetotelluric (MT) method was used as one of several geophysical tools to study part of the Idaho Engineering Laboratory (INEL) facility. The purpose of the geophysical study on INEL was to investigate the facility for a possible site to drill a geothermal exploration well. A successful geothermal well would be used to provide hot water for a chemical processing plant. The MT method was employed to map any large-scale structures or conductivity anomalies that might prove interesting as geothermal exploration targets. In addition to the MT data, direct current resistivity soundings, gravity data, aeromagnetic data, and seismic refraction data were obtained in the course of the geophysical study. In the MT survey described in this paper, an additional goal was to provide a better understanding of the electrical units mapped in the regional study of the Snake River Plain (SNRP) by Stanley et al. (1977). It was thought that a widespread conductive layer found beneath surface basalts in the 1977 study could be categorized petrologically by a deep well and additional MT soundings done nearby. Also, INEL is located on the margin of the SNRP, and it was desired to have MT data in the area to study the electical nature of the margin of the plain. The MT sounding interpretations did not indicate any conducticity anomalies or significant structures near the chemical processing plant which could be used to guide the location of the proposed geothermal well to be drilled to a depth of 3 km. The initial interpretation of the MT sounding data was done with one-dimensional models consisting of four or five layers, the minimum number required to fit the data. After the test well (INEL-1) was completed, the electric log was used to guide an improved one-dimensional ID interpretation of the MT sounding data. Profile models derived from the well log provided good agreement with velocity models derived from refraction seismic data. A resolution study using generalized inverse techniques shows that the resolution of resistive layers in the lower part of the MT models is poor, as is the definition of a shallow, altered basalt unit. The only major structure observed on the MT data was the faulted contact between the SNRP and basin and range structures on the west. Modeling of the data near this structure with a two-dimensional computer program showed that the MT data near the fault require a model similar to the seismic refraction models and that structure on a deep crustal conductor is also required.