To test the hypothesis that the frequency with which vocalization must be initiated in a given speech segment and the frequency of attendant disfluency are positively related, two passages were constructed. One passage was composed entirely of voiced speech sounds (all-voiced passage). The other contained both voiceless and voiced sounds (combined passage). Thus, in reading the later material, subjects had to effect more “off-on” phonatory adjustments than in the all-voiced selection. Aside from this difference, the passages were closely matched along several other linguistic and phonetic parameters. Fourteen stutterers performed five massed oral readings of each passage. Statistical analyses all showed that there was significantly less stuttering and more rapid adaptation associated with the all-voiced material.