Sensory storage of fragmentary displays.

A partial-report version of the Eriksen-Collins paradigm was devised to test iconic storage. The test lay in presenting arrays of 12 partial letters, one of which was randomly chosen for completion. Additional conditions checked whether the completed letters were legible and to what extent the completion fragments were guessable. Test sequences included dark or light inter-stimulus intervals and 4 different delay times, which had no significant effects. Analysis showed that subjects were no better able to report a synthesis of the first and second fragments than to guess the letters from separate fragments. A second experiment demonstrated some integration of successive single-letter presentations, although subjects apparently do not have iconic access to full arrays of letter fragments.