Four experiments are reported that examine subjects' ability to form and use images of tones and chords. In Experiments 1 and 3, subjects heard a cue tone or chord and formed an image of a tone or chord one whole step in pitch above the cue. This image was then compared to a probe tone or chord that was either the same as the image in pitch, different from the image in pitch and harmonically closely related, or different and harmonically distantly related. Image formation times and response times were measured. In Experiment 3 a random-tone mask was used to control for possible contribution of the cue in echoic memory. In both experiments tone images were formed faster than chord images, a result consistent with the idea of structural complexity as a determinant of image formation time. Response times and accuracy rates were found to parallel results found in music perception studies, results consistent with the idea of shared mechanisms in the processing of musical images and percepts. Experiments 2 and 4 were control experiments examining the possible influence of demand characteristics and subjects' knowledge. In experiment 2 subjects predicted the results of "an average person" using imagery, and in Experiment 4 subjects completed the task without receiving imagery instructions. The findings ruled out the possibility that demand characteristics and subjects' knowledge were solely responsible for the results of Experiments 1 and 3 and support the argument for the role of imagery.