Abstract
In this paper, we investigate the reading performance of a patient previously reported as having an impairment at the level of the graphemic buffer (Kay & Hanley, 1994). Despite his spelling problems, this patient is good at reading familiar words, at reading nonwords, and at written lexical decision. This pattern of performance is discussed with respect to Caramazza, Capasso, and Miceli's (1996) recent claim that damage to the graphemic buffer leads to a problem in reading that will be obvious especially when the patient is asked to read nonwords.