Two ocean bottom magnetometers have been deployed at sites on either side of the Juan de Fuca Ridge, 20 and 40 km from the crest, near latitude 47¿N. Twelve days of data were recorded simultaneously at the two sites, and these data have been used to determine the electrical conductivity structure beneath the region. The vertical field variations at both stations are of very low amplitude in the period range 0.25 to 8 hours, with no detectable phase reversal across the ridge. Electric fields in the ocean layer have been estimated from the attenuation of sea floor horizontal fields with respect to Victoria, and the resulting impedance spectra inverted to find resistivity as a function of depth using the technique of Oldenburg (1979). Resistivity models which fit the data within the specified bounds are characterized by a comparatively resistive crust and upper lithosphere overlying a low resistivity zone between depths of 55 and 75 km which in turn appears to be underlain by a higher resistivity section. The average resistivities are used to estimate temperature and melt fraction on the basis of the effective medium theory of Shankland and Waff (1976) for a basalt melt fraction within an olivine matrix. For the low resistivity zone, and depending on the water content of the melt (0 to 10%), the estimated temperature range is 1400¿C to 1250¿C and the corresponding melt fraction 0.05 to 0.09. The data suggest that the melt fraction decreases below 80 km. The conductivity profile from magnetotelluric data on 72 Ma ocean floor in the North Central Pacific (Filloux, 1977) has similar characteristics to the 1 Ma old conductivity structure near the Juan de Fuca Ridge, but with the low resistivity layer at depths of 140 to 220 km. In both cases, the resistivity minimum should be equated with the region of maximum partial melt within the asthenosphere, rather than with the base of the lithosphere. The increase in depth to this zone with increasing lithosphere age is in general agreement with plate tectonic concepts.