Owing to the Japanese language's unique writing system, which consists of phonograms and ideograms, reading impairments of Japanese brain-damaged patients have attracted the interest of many researchers. Past case reports as well as some widely accepted handbooks and textbooks have concluded that a specific aphasia type or lesion site is associated with a particular impairment pattern of phonograms and ideograms in reading. However, the methodology and analytical procedures in previous studies were inadequate for reliable generalizations to be made. First, the test materials were unspecified or inappropriate, or the number of test items was small. Secondly, the conclusions were presented without providing individual performance data to support them Thirdly, in associating patterns of reading impairment with lesion sites, only single cases were reported The present investigation was designed to overcome the omissions of previous studies, and examined the ability to read 46 single phonograms and 46 single ideograms aloud in four groups of sufficiently large numbers of patients; namely, seven pure alexics, 23 Broca aphasics, 13 Wernicke aphasics, and seven patients with alexia and agraphia. Ours are the first data to demonstrate unequivocally no consistent linkage between aphasia type and the patterns of impairment of phonogram and ideogram reading. The impairment patterns were not uniform across patients even in the same aphasia group A majority of the cases in each group showed that phonograms and ideograms were unselectively impaired. However, ideogram reading was more difficult in three cases in the pure alexia and Broca aphasia groups, respectively, and in one case in the Wernicke aphasia group. Phonogram reading was more severely disturbed in four cases among the Broca aphasics and in one case among the patients with alexia with agraphia. An apparent variablity of impaariment patterns characterized the Broca aphasic group. These dyslexic patterns did not appear to correlate with the site and extent of lesions identified by computerized tomography scans. Past reports linking a particular impairment pattern of phonograms and ideogram reading and a specific lesion site were studies of single cases, and their conclusions seem oversimplified. While sensory and motor dysfunctions can usually be neuroanatomically localized in individuals, impairments of certain high cortical functions, such as the reading of phonograms and ideograms, may not be correlated with damage to definite neuroanatomical structures.