Simultaneous Apollo 12 surface and Explorer 35 orbital magnetometer records have been reexamined for intervals suitable for measuring the lunar inductive response in the solar wind and terrestrial magnetosheath. Improvements in the selected time series are obtained by (1) editing of intervals containing extraneous surface field fluctuations caused by plasma interaction with the local crustal magnetic field, and (2) minimization of magnetometer data errors in the Apollo 12 record. Power spectral estimates of the response tangent to the lunar surface, defined in terms of transfer and gain functions, are obtained for the frequency range 10-5 to 10-3 Hz using methods introduced earlier by Wiskerchen and Sonett (1977). Maximum consistency of estimates from different time intervals is found when the analysis is limited to the tangential direction of maximum incident power as measured at Explorer 35 for each interval, i.e., the direction in which the ratio of signal to background noise is greatest. A simplified version of spherically symmetric plasma (SSP) confinement theory is used to interpret the transfer function data via forward model calculations under the assumption that the electrical conductivity profile are narrowest in the depth range 450--1350 km and are more restrictive at these depths than those obtained previously from plasma-confined response data. An upper limit of ≤360 km for the radius of a metallic core with assumed conductivity ≥102 mhos/m is estimated from the one standard deviation error limits of the measurements. Using the bulk conductivity versus temperature model proposed by Huebner et al. (1979), rough limits on the selenotherm are calculated which appear to be in agreement with independent geophysical constraints at depths less than about 800 km.