This study investigated the ability to produce appropriate voice onset time (VOT) contrasts under conditions of rate modulation in groups of nonfluent aphasic subjects, fluent aphasic subjects, and nonneurological controls. Acoustic analyses of the consonants [b d g p t k] produced in the context of the vowels [i e a o u] at two different rates of speech revealed that normal subjects' VOTs were significantly shorter at the fast rate of speech relative to the slow/normal rate, as expected. In addition, the rate change had a significantly greater effect on voiceless stops as compared to voiced and on velar consonants as compared to labials and alveolars. The nonfluent aphasic patients exhibited a similar pattern except that no differences in magnitude of rate-related changes were found across place of articulation. Further, similar to previous studies, the nonfluent aphasic patients produced voice and voiceless consonants with somewhat overlapping VOT distributions, indicating an impairment in temporal integration in these subjects. Finally, the fluent aphasic patients demonstrated a surprisingly aberrant pattern of results, with VOTs under th fast condition shorter than under the slow, but no differences in magnitude of change across place of articulation or voicing categories. The results are discussed in relation to the nature of speech production deficits in both nonfluent and fluent aphasic patients. Implications for remediation are considered.