For almost 10 years the KTB superdeep drilling project has offered an excellent field laboratory for adapting seismic techniques to crystalline environments and for testing new ideas for interpreting seismic reflections in terms of lithological or textural properties of metamorphic rock units. The seismic investigations culminated in a three-dimensional (3-D) reflection survey on a 19¿19 km area with the drill site at its center. Interpretation of these data resulted in a detailed, structural model of the German Continental Deep Drilling Program (KTB) location with dominant, steep faults in the upper crust. The 3-D reflection survey was part of a suite of seismic experiments, ranging from wide-angle reflection and refraction profiles to standard vertical seismic profiles (VSP) and more sophisticated surface-to-borehole observations. It was predicted that the drill bit would meet the most prominent, steeply dipping, crustal reflector at a depth of about 6500--7000 m, and indeed, the borehole penetrated a major fault zone in the depth interval between 6850 and 7300 m. This reflector offered the rare opportunity to relate logging results, reflective properties, and geology to observed and modeled data. Post-Variscan thrusting caused cataclastic deformation, with partial, strong alterations within a steeply dipping reverse fault zone. This process generated impedance contrasts within the fault zone on a lateral scale large enough to cause seismic reflections. This was confirmed by borehole measurements along the whole 9.1 km deep KTB profile. The strongest, reflected signals originated from fluid-filled fractures and cataclastic fracture zones rather than from lithological boundaries (i.e., first-order discontinuities between different rock types) or from texture- and/or foliation-induced anisotropy. During the interpretation of seismic data at KTB several lessons were learned: Conventional processing of two-dimensional (2-D) reflection data from a presite survey showed predominantly subhorizontal layering in the upper crust with reflectivity striking in the Variscan direction. Drilling, however, revealed that all rock units are steeply dipping. This confirms that surface common depth point (CDP) seismics strongly enhances subhorizontal reflectivity and may thus produce a very misleading crustal image. Although this was shown for synthetic examples earlier, the KTB provides the experimental proof of how crucial this insight can be.¿ 1997 American Geophysical Union