Strange as it may seem, there is not universal agreement about the exact nature of seismic reflections. Physicists tend to think of the reflections as caused by the interface between rock types, as a sand-to-shale contact. The problem with this view is that sands and shales interlace in complex ways, both larger and smaller than the seismic wavelength. Many geologists, particularly a group known as seismic stratigraphers, have a different concept. (See Seismic Stratigraphy--Applications to Hydrocarbon Exploration, memoir 26 of the American Association of Petroleum Geologists). They have studied thousands of miles of reflection data along with well logs. They believe a reflection marks a constant geological time horizon. They assert that a long, continuous reflector could represent terrigenous deposition on one end and marine deposition on the other end with a variety of rock types in between. Data interpretation based on this assumption is called chrono-stratigraphy. The view of the seismic stratigrapher seems reasonable enough for areas that are wholly clastic, but when carbonates and other rocks are present, the physicist's view seems more appropriate. For further details, the book of Sheriff  is recommended.