The hypothesis tested was that stutterers subvocalize more slowly than nonstutterers and that they need more time for the overt production of the fluent parts of their speech. We also investigated whether rate differences could only be observed for those words on which the stutterers expect to stutter. Fifty-nine school children (27 stutterers and 32 nonstutterers) and 19 adults (18 stutterers and 21 nonstutterers) performed a reading task in which a noun was presented together with its definite article. The presentation times of the reading material were controlled by the subjects. Half of the material had to be read silently, the other half orally. In oral reading, only the data from those trials without any indication of disfluencies were used. Dependent variables were presentation times, speech latency, and speech duration. The stutterers' silent presentation times were significantly longer than those of nonstutterers and this difference was significantly greater for children than for adults. In oral reading all stutterers, regardless of age, had longer presentation times, speech latencies, and article durations than the nonstutterers. Some nouns, however, were uttered significantly more rapidly by stutterers than by nonstutterers. These time differences were found to be independent of the stutterers' expectation to stutter. Our results indicate that a strictly motoric explanation of stuttering is inadequate. The data show that the stutterers and nonstutterers differ with respect to the temporal parameters not only during speech execution, but during speech planning as well.