The relationship was explored between the ability to understand nonverbal metaphoric similarity and the ability to produce verbal metaphors to talk about that similarity. Subjects, 45 college students, were shown filmed scenes that depicted both metaphoric and literal similarity. Half of the scenes were of stationary objects and half of moving objects (events). In the latter case, for the metaphoric pairs, the same style of motion was used to link two unlike objects (a spinning ballerina and a spinning top). The subjects rated both the literal and the metaphoric pairs for the similarity, novelty, and comprehensibility of the pairing. They then wrote original metaphors as descriptions of the pairs and rated those metaphors for ease and quality of production. Results showed that the subjects rated the literal pairs as more similar, less novel, and more comprehensible than the metaphoric pairs of scenes. The intercorrelations of all ratings for metaphoric pairs showed that similarity, novelty, and comprehensibility correlated significantly with the ease of metaphor production, which in turn was correlated significantly with the rated quality of production. Similarity and novelty influenced the ease of apprehending the metaphoric relationship that underlies the descriptive metaphor. (FL)