The oceanic lithosphere is an extremely efficient waveguide for high-frequency seismic energy. In particular, the propagation of the regional to teleseismic oceanic Pn and Sn phases is largely controlled by properties of the oceanic plates. The shallow velocity gradient in the sub-Moho lithosphere results in a nearly linear travel time curve for these oceanic phases and an onset velocity near the material velocity of the uppermost mantle. The confinement of Pn/Sn to the lithosphere imposes a constraint on the maximum range that a normally refracted wave can be observed. The rapid disappearance of Sn and the discontinuous drop in Pn/Sn group velocity beyond a critical distance, dependent upon the local thickness of the lithosphere, are interpreted as a shadowing effect of the low Q asthenosphere. Wave number integration was used to compute complete synthetic seismograms for a model of oceanic lithosphere. The results were compared to data collected during the 1983 Ngendei Seismic Experiment in the southwest Pacific. The Pn/Sn coda is successfully modeled as a sum of leaky organ-pipe modes in the sediment layer and oceanic water column. While scattering is present to some degree, it is not required to explain the long duration and complicated nature of the Pn/Sn wave trains. The presence of extremely high frequencies in Pn/Sn phases and the greater efficiency of Sn than Pn propagation are interpreted in terms of an absorption band rheology. A shorter high-frequency relaxation time for P waves than for S waves results in a rheology with the property that Qα>Qβ at low frequency while Qβ>Qα at high frequency, consistent with the teleseismic Pn/Sn observations. The absorption band model is to viewed as only an approximation to the true frequency dependence of Q in the oceanic lithosphere for which analytic expressions for the material dispersion have been developed. ¿American Geophysical Union 1987