There has been increasing worldwide interest in research and development of Rotating Detonation Engines (RDEs) as a propulsion system. A study was conducted to examine detonation behavior in a linearized representation of an RDE. The two-dimensional nature of the phenomenon in such a geometry permits the use of classical detonation visualization techniques. A predetonator was used to generate a steady detonation wave upstream of the test section. The test section was separated from the predetonator by a sealing door such that it could be filled with inert gas replicating the nonreactive properties of the combustion products in an actual RDE. The top wall of the test section contained a linear array of small holes through which premixed stoichiometric hydrogen-oxygen flowed into the inert gas filled test section to form a stratified layer just prior to the arrival of the detonation wave. The test section was equipped with windows to permit high-speed schlieren photography. The soot foil technique was used to capture the detonation cellular structure in the resulting stratified layer. The layer height was varied by changing the time that elapsed from when the hydrogen-oxygen injection started to when the detonation wave arrived at the test section. The minimum layer height required for detonation propagation accommodated 10-11 detonation cells. This is significantly larger than the three detonation cell requirement reported in previous studies carried out with sharp interface homogeneous-mixture stratified layers. An alternate rendition of the test was conducted whereby a finite axial-length stratified layer formed by a buoyancy driven predetonator gravity current. In these tests, the stratified layer occupied roughly half the test section height and the equivalence ratio was used to vary the predetonator detonation cell size. The fuel-rich propagation limit corresponded to an equivalence ratio of 2.5, for which roughly three cells were accommodated by the stratified layer. Tests carried out with nitrogen and carbon dioxide in the test section showed a strong effect of the inert gas, indicating that substantial mixing occurred at the leading edge of the gravity current.