The results of two experiments are reported that involved the use of both a recognition and a cued-recall task. In both experiments, the cue-target relationship was based on either mean ing or rhyme. The only difference between these experiments was rate of presentation (5.0 or 1.5 sec) during study. The results indicate that targets were more likely to be both recognized and recalled when semantic cues appeared at study and in cued recall. Target items were more likely to be only recognized when rhyme cues appeared at study and in cued recall. The results are discussed with respect to retrieval-based recognition (recognition accompanied by recall) and familiarity-based recognition (recognition only). It is suggested that these two types of recogni tion are often inversely related, but they are also reflective of different memory systems. Horton, Pavlick, and Moulin-Julian (in press) have pro posed a dual-process account of recognition memory that is quite similar to earlier accounts, such as those of Man dler (1980) and Jacoby and Dallas (1981). In the Horton et al. proposal, retrieval-based recognition and familiarity based recognition are the two processes involved in rec ognition. Familiarity-based recognition relies on sensory perceptual information, and retrieval-based recognition relies on processes similar to those that influence recall. To this point, this proposal is in conceptual agreement with Mandler (1980) and Jacoby and Dallas (1981). How ever, Horton et al. go on to suggest that retrieval-based recognit ion is an episodic memory process and that in creases in the quality of information in the episodic sys tem lead to increases in both retrieval-based recognition and recall. In contrast, they propose that familiarity-based recognition operates in terms of sensory-perceptual fea tures in semantic memory , much like those involved in pattern recognition. Increases in familiarity-based recog nition occur as the quality of information in the episodic memory system declines . In other words, retrieval-based recognition and familiarity -based recognition tend to be inversely related . The experiments reported by Horton et al. involve the use of a paradigm in which both recognition and recall of target words are measured . In this paradigm , four out comes are possible . Targets may be remembered (i.e ., recognized and recalled), recognized only, recalled only, or forgotten (i.e. , not recognized and not recalled) . The results of two experiments reported by Horton et al. show that increases in the associative relationship between cue and target members of word pairs presented at study are directly related to performance in the remembered out come but are inversely related to performance in the