Two experiments examined the effects on recall of encoding and retrieval "depth" (the extent to which subjects process the semantic as well as the phonetic and orthographic attributes of verbal material), encoding-retrieval cue compatability, and subject versus experimenter generation of cues. In the first experiment, 117 undergraduates, divided into three groups, saw and heard a list of 72 target words. One group was told to write down an associate for each word--one a rhyme, and one just to learn the list for later recall. Rhyming or associate cues were then given on a recall test. Results indicated that encoding depth has an impact on memory performance over and above encoding-retrieval cue compatability. Semantic processing seemed to produce cued recall superior to phonetic processing even if the cues provided at recall were incompatible with those encountered during encoding. Recall was still maximized by compatible cues. Using 84 subjects and the same 72 target words, the second experiment added the question of subject versus experimenter generation cues to the questions about encoding depth and compatibility. Results of this experiment confirmed the importance of encoding depth while also indicating that subjects perform better when they generate their own cues. (JL)