Seismic interpretation is often based on the analysis of amplitude anomalies, which depend strongly on the seismic wavelet presented in the data. However, if the wavelet polarity or phase is unknown or fine-scale impedance variations are complex, interpretation of the anomaly can be ambiguous. The Stratton data volume contains a dome-like feature that may be interpreted as the top of a potential gas target, the top of a buried tight reef, or as a thin layer of either higher or lower impedance, depending on the interpreter’s assumption of polarity and phase. This observation provoked our interest in modeling the seismic response of domes using wavelets of differing polarity and phase on stacked data. Because there appears to be only a single event, perhaps the top of an anomalous feature, and not its base, a “gradational” decrease in impedance contrast with depth is included among our models. We have determined that the seismic response from a layer with an impedance contrast decreasing with depth is quite different from that of a layer with constant impedance contrast when the bed thickness exceeds one quarter of the wavelength; that is, a reflection from the base of a “thick” gradational layer is not visible, as expected. We independently determine the polarity and phase of the Stratton data, finding that the surface-based seismic and VSP data are of opposite polarity (European and American, respectively), and concluding that the dome structure represents the top of a gradational thick bed. A model based on a nearby reservoir containing thin gas, oil, and water zones supports this conclusion. This anomaly in the Stratton data appears to represent a hydrocarbon reservoir with thin layers of gas and oil, each with lower impedance than the surrounding beds but with stepwise decreasing contrast over a sufficient thickness to avoid a basal reflection at these wavelengths.