Developing experimental materials to study metaphor memory is a difficult process, especially in regard to finding a suitable type of control stimulus. The dead metaphors were technically metaphorical but have long since come into common usage; they were thus considered an intermediate type between metaphors and nonmetaphors. The fact that images were reported more often with metaphorical than non-metaphorical sentences suggests that imagery may be a more frequently used mnemonic with metaphors than with more literal language, even though it is not necessarily all that effective. The greater number of counting errors for metaphors may be attributed to differential saliency of the input sentences. The metaphors were novel, while the dead metaphors and nonmetaphors were commonplace. People’s intuitions about metaphors are often highly inaccurate, especially in the direction of thinking them to be more difficult, less informative, and less imageable than they truly are.