The Ricinus incised valley (Turonian–Coniacian Cardium Formation) is at least 50 km long, 10 km wide and up to about 30 m deep. It trends NW–SE, parallel to tectonic trends and shoreline trends in the Cardium Formation. It cuts into shallow‐marine mudstones and sandstones (Raven River Allomember of the Cardium), and is truncated by a transgressive surface of erosion (E5). The valley fill forms a small oil reservoir in the subsurface, but the field is disrupted by thrust faults, and the western margin is poorly constrained because of thrusting and lack of wells. The valley fill consists of interbedded sandstones and mudstones, with both brackish and fully marine trace fossil assemblages. Some of the sandstones are structureless, with individual beds apparently over 4 m in thickness; others are parallel‐laminated with beds up to 3 m thick. Cross‐bedding in sets thicker than 5 cm is extremely rare. The mudstones contain some thin silty and sandy laminations, but are commonly extensively bioturbated with a fully marine trace fauna (robust Helminthopsis, Zoophycos, Planolites, Anconichnus, Skolithos, Ophiomorpha, Thalassinoides, Teichichnus, Asterosoma, Terebellina and Rosselia). The mudstones can be traced as a sheet about 4 m thick along the entire length of the valley; they appear to pinch out against the eastern wall of the valley. Chert‐pebble conglomerates are scattered throughout some wells, but tend to be concentrated near the northern end of the valley. Here, they are associated with sandstones and with fully marine mudstones. The valley was formerly interpreted to have been cut and filled by turbidity currents, but this interpretation is difficult to reconcile with the depositional environments of the rest of the Cardium Alloformation. It is suggested here that the valley was cut by a river during a falling stage of relative sea‐level, and was filled transgressively as an estuary. There is absolutely no evidence for tidal currents, and little indication of waves. The very thick and structureless sandstones suggest pulses of sediment swept into the estuary from the bay head, possibly as density underflows. There is little or no evidence for the reworking of these deposits within the estuary. There are no central basin mudstones (no turbidity maximum), although this could be due to incomplete preservation of the entire length of the estuary. The geometry of the estuary and the facies distribution of the fill suggest that the open sea lay to the north. Chert pebbles moving in the longshore drift system were introduced into the estuary from the north by storms; they were not introduced from the fluvial (southern) end of the estuary. The orientation of the incised valley, parallel to depositional strike for over 50 km, remains a major problem.