In recnt years, the subtype or uniformity of conduction aphasia has been investigated by several authors. Among them, Lhermitte et al (1980) proposed that kinesthetic aphasia, as described by Luria, was distinguished by the clincal specificity and by the site of lesion in the anterior parietal region of the dominant hemisphere. We also had an opportunity to examine a patient (case 1 ¹⁰⁾) developing conduction aphasia after cerebral infarction confined to the left anterior parietal region, and compared this case neurolinguistically with three other conduction aphasics (case 2, 3, 4) whose lesions were in other areas and with one Broca aphasic.     The results of phonological analysis of speech sound errors (phonemic paraphasic errors), in case 1, were as follows : 1) Errors mainly occurred in consonants, unlike the three other conduction aphasics. This finding was rather similar to that of the Broca aphasic. 2) Concerning the erros types, substitution was seen more frequently than transposition, unlike the other conduction aphasics. This finding also resembled that of the Broca aphasic. However, in case 1, omission was less frequantly seen than in the Broca aphasic. 3) Errors, as with the other conduction aphasics, were distributed at random in terms of the distinctive feature distance, unlike in the Broca aphasic, more than half of whose errors were different from the target sounds only by one distinctive feature. These findings were close similar to those of Lhermitte et al.     The above results suggest that the conduction aphasics with the anterior parietal lesion may be located in a particular position among conduction aphasics. Besides, we think that the impairment of speech output in case 1 without any comprehension disturbance is at a level closer to the phonetic level rather than to the phonemic level.