Three hypothetical models of tongue movement of the walrus during suction feeding are examined. These models encompass the entire range of simple tongue retraction movements possible by examining 1) movement of the tongue directly to the rear following the curvature of the palate, 2) to the rear and ventrally in a straight line, and 3) ventrally in a straight line. The percent of muscular force available from the hyoglossus, genioglossus, and styloglossus that could be applied toward retraction as predicted by each model is calculated. The resistance that the tongue would provide during retraction is calculated using projected tongue areas and is combined with the above data from the muscles to provide an estimate of the percent of the total available force that is needed to retract the tongue for each model. A separate examination of the direction of tongue-induced wear striations on the palatal and lingual aspects of the teeth is used to help support or reject the three models.

The model where the tongue is moved directly to the rear is supported by studies of both muscle force and tooth wear. In the mammalian groups that were compared to the walrus, there is a great deal of interspecific variation in movements of the tongue during suction feeding; no two groups can be considered to have identical stereotyped tongue movements.