Twenty-four magnetotelluric (MT) soundings have been collected in an east-west profile across the center of Long Valley caldera. The average station spacing is approximately 1 km and appears adequate to sample the important features of the upper crustal and deeper resistivity structures. Additional control on the shallowest resistivity is provided by a continuous profile of time domain electromagnetic soundings coincident with the western portion of the MT line. Our MT data set reveals numerous resistivity structures which illuminate the evolution and present state of the Long Valley system. Many of these have been quantified through two-dimensional (2-D) finite element modeling emphasizing the transverse magnetic (TM) mode. Important structural components include low-resistivity layers 0.5-1.5 km deep under the eastern half of the caldera, beneath the axial graben of the resurgent dome, and under the west caldera moat. Most of this layering appears to lie in post-caldera Early Rhyolite tuffs, and the uppermost unwelded Bishop Tuff. These rhyolite units have been observed to be porous and highly altered and to commonly contain Pleistocene intercalated lacustrine clays. The remainder and majority of the Bishop Tuff appears highly resistive. A low resistivity layer also occurs below the axial graben near the base of the Bishop Tuff (1.5 km). Hydrothermal fluids or alteration in precaldera volcanic strata or, less likely, carbonaceous metasediments may be the cause of this. Resistive, probably crystalline basement at high levels is apparent beneath the center of resurgence. Low resistivities are modeled at a depth around 5 km below the entire west moat and central graben and may represent a zone of hydrothermal fluids released from magma crystallization, with potential magmatic contributions at greater depths. The correspondence between this low resistivity and teleseismic delay and low density zones found in other studies is quite striking. A subtle anomaly in the transverse electric (TE) mode impedance is weakly suggestive of a midcrustal conductive axis centered beneath the central graben and resurgent dome. However, it cannot be simulated by two-dimensional transverse electric calculations and requires a full three-dimensional evaluation to ensure that the anomaly does not represent resistivity complexity in just the upper few kilometers. A fundamental, caldera-wide 3-D effect is documented by comparison of observed and computed TE impedance and vertical magnetic field data. The abrupt termination of conductive caldera sediments less than 10 km north and south of our profile greatly depresses the observed TE apparent resistivity and vertical magnetic field relative to the model calculations for periods greater than 0.3 s for the central and eastern caldera. Analysis of the TE mode data also suggests that a similar finite-strike effect lies in the response at periods greater than 3 s due to the mid-crustal west moat conductor. The TM mode measurements are judged to also contain some large-scale departure from the 2-D assumption related to horizontal current gathering from the north and south. This inflates the apparent resistivity and decreases the phase somewhat around 10 s over the central portion of the caldera relative to the 2-D model response. The regional profile of resistivity for the data at hand can be modeled with a 40 ohm-m basal half-space beneath 30 km of crust of 1000 ohm-m or more. Although stations outside the caldera are very desirable to constrain this deep profile better, there is no evidence for a discrete low-resistivity layer deep below Long Valley in contrast to our interpretation in the northeastern Basin and Range.