It is well known that the voice-onset-time (VOT) boundary for synthetic stop consonants shifts towards longer VOT's as place of articulation shifts from front to back, even if all parameters of the signal except VOT and the starting frequencies of F2 and F3 are held constant. Is this a phonetic effect contingent on a decision about the place feature, or is it an auditory effect that is somehow dependent on the spectrum at syllable onset? In order to decide this question, F2 and F3 transitions were varied in a number of steps along a “place continuum,” and the VOT boundary was determined at each step. The function relating the VOT boundary to the place continuum should be flat within place categories and show jumps at the category boundaries if the effect is phonetic, but it should be monotonically increasing if the effect is auditory. Neither hypothesis has been supported by the data so far, which show the function to be generally increasing but strongly nonmonotonic, with a number of seemingly haphazard but statistically reliable peaks and valleys. A radically different pattern (with the same general characteristics) is obtained if the fundamental frequency of the stimuli is changed. Further research on the acoustic basis of the effect and its perceptual and methodological implications will be reported.