Grapheme-color synaesthesia, a condition in which achromatic graphemes elicit vivid experiences of color is believed to be a genetically determined trait. We describe a study of 10-year-old twin brothers who are physically identical in appearance but who have considerably different conscious experiences. A phenotypic analysis that measured the consistency of grapheme-color pairings over test-retest confirmed that one twin has grapheme-color synaesthesia and the other twin does not. A genotypic analysis using sixteen microsatellite loci confirmed that the twins are monozygotic. These findings are problematic for previous suggestions that synaesthesia is an X-linked dominant trait. At the very least, the findings show that the penetrance of the genotype for synaesthesia is incomplete and that any view suggesting that synaesthesia is simply an X-linked dominant trait is therefore also incomplete and possibly even incorrect. The findings also negate a previous suggestion, based on a study of female monozygotic twins, that discordance of synaesthesia in identical twins is due to X-inactivation. In general, the findings raise serious questions regarding whether it is possible at this time to establish the genetic contribution to synaesthesia.