The upper Albian Viking Formation in the subsurface of Alberta contains at least four major erosional surfaces, or bounding discontinuities. The highest of these surfaces, Viking Erosion surface 4 (VE4), can be correlated basinwide. There is an easily recognized well log marker in the shales 10-50 m above VE4, known as the Base of Fish Scales (BFS). An isopach map of the BFS-VE4 interval has been constructed using 9987 wells over an area of about 85,800 km 2 . The depositional surface represented by BFS was probably smooth and almost flat during deposition, and the isopach map can be interpreted as showing the topography of the VE4 surface. The surface is stratigraphically deepest in the northeast, and rises southwestward. Its topography is characterized by a broad east-west-trending arch and two linear steps that trend northwest-southeast. The VE4 surface is overlain by a coarse lag a few centimeters to about 5 m thick, with clasts up to 15 cm proximally (SW) and 7 mm distally (NE). Topographic relief on VE4 is at least 40 m. VE4 is interpreted as having formed initially during a major lowering of relative sea level, with modification of the subaerial topography and removal of all evidence of subaerial exposure during the ensuing southwestward marine transgression--thus VE4 is a transgressive surface of erosion. Several well log markers onlap the VE4 surface. They represent coarse to granule sandstones with extremely sharp bases. These tongues of sandstone can be traced up to 20 km basinward from their points of onlap and are interpreted as thin lower shoreface sandstones that formed during minor forced regressions super-imposed on the overall transgression. The initial regression that immediately preceded VE4 implies a fall of relative sea level of at least 40 m, and a northeastward movement of the shoreline of at least 250 km. These two figures imply a basinward gradient of about 0.00016. Structure contour maps of the top of the Mannville Group (Albian) and the top of the Wabamun (Upper Devonian) show smooth unbroken surfaces in the study area, with strikes different from those of the arch and steps. Thus there appear to be no older structural elements that influence Viking trends. The broad east-west-trending arch cannot be identified at Wabamun, Mannville, or older Viking horizons, and had disappeared before the BFS horizon was deposited. It is at least 170 km wide and 6 m high, and formed shortly before the VE4 transgression. Thus there may have been structural control confined to the Viking that influenced changes in relative sea level at the end of Viking deposition.