A possible example of a very deep glacial excavation is provided by a distinctive gravity low located at the front of a valley glacier that once flowed into glacial Lake Aniuk (formerly Lake Noatak) in the western Brooks Range. Geologic and geophysical data suggest that sediments or ice filling a glacially excavated valley are the most probable cause of the 30–50 mGal anomaly. Reasonable choices of geometric models and density contrasts indicate that the former excavation is now filled with a buried-ice thickness of 700 m or sediment thicknesses greater than 1 km; comparable anomalies are not known for other glaciated lacustrine valleys. However, many fiords do exceed 1 km in depth, and Crary found one nearly 2 km deep in Antarctica. In studying this fiord, he suggested the probable increased efficiency of excavation directly behind the point where an outlet glacier becomes afloat to form the Ross Ice Shelf and where it thus has a vertical component of motion and a mechanism for debris removal. Floating glacier ice tongues are now rare in the Arctic, but they exist in maritime parts of northern Ellesmere Island and Greenland. Studies of ice movement, environment, and morphology of another large floating glacier tongue in a perennially frozen lake in the Angiussaq Mountains of northern Greenland suggest that Pleistocene Lake Aniuk could have had a similar environment, water temperature, and near-stable water level and that it could have maintained both a floating polar glacier tongue and a perennial ice cover. No direct evidence of efficient excavation was observed in Greenland, but efficient glacial erosion behind a floating polar ice tongue could explain the excavation that caused the Alaskan gravity anomaly.