40 undergraduates were presented with simple letters made up of patterned line modes. Other subjects were shown the same modes but in the form of parallel unconnected lines. On subsequent recognition tests recognition was superior for the lines presented in the unconnected condition. This was true despite various types of recognition tests designed to control for possible bias in favor of the unconnected condition. These improved procedures support the suggestion that the articulated form itself becomes an added factor in the memory span that hinders retention of its component elements.