During the summer of 1976, an electromagnetic propagation experiment was performed on the Olympic Peninsula in northwest Washington. A grounded horizontal electric dipole transmitter, with a peak moment of 1.6 × 105 A m, was used to generate square-wave signals having periods from 10 to 1 sec. Strong signals were received on a 100 m long electrode pair emplaced on the seafloor at a depth of 312 m and at a range of 22 km from the transmitter. The fundamental was received with a 20 to 30 dB signal-to-noise ratio (snr), and many harmonics—14 in one case—were also clearly detected. Calculations based on the assumption that the above-ground near-fields penetrate downward through the ocean to the receiver give results that agree well with measurements at 22 km and 312 m, namely, theory and experiment agree to within a factor of two for all periods transmitted and harmonics detected. Even if a relatively strong signal had reached the ocean through the suboceanic crust, it would have been masked by this over-down mode. Measurements were also made on the seafloor at ranges from 110 to 135 km and a depth of 2500 m, using receiving antennas 540 to 1000 m in length. At these sites, the ocean screened out the over-down mode, and any signal detected would have had to propagate through the crust. Although the equipment was working well, no signals were detected. This negative result indicates that the effective conductivity of the suboceanic path was at least 10−3 mhos/m. However, unknown factors such as the possible existence of a subduction zone on the suboceanic path might well have caused the nondetection.