Synesthesia is a condition in which an otherwise normal subject "sees" specific colors associated with specific numbers or letters (graphemes). Is this a sensory effect or simply a memory association? Are they simply being "metaphorical" (just as we say cheese tastes "sharp")? We find that synesthetically induced colors can lead to perceptual grouping and pop-out, and are not seen with eccentric viewing, suggesting that the effect is genuinely sensory. In addition, numbers or letters presented in the periphery are difficult to identify if presented in the presence of other, flanking, graphemes ("crowding"). This effect is attenuated by presenting the target graphemes in a different color than the flanking graphemes. We find that synesthetic subjects are significantly better at identifying the target grapheme than controls. Hence, synesthesia is a perceptual effect, as opposed to a memory association, or mere "metaphor." Zeki's V4 lies adjacent to recently discovered areas that process number-graphemes in the fusiform gyrus. We propose that number-color synesthesia arises as a result a mutation that causes excess connections, or a failure of pruning of connections, between V4 and the number area in the fusiform gyrus. Because synesthesia runs in families, we propose that a single gene mutation causes increased neural cross wiring between adjacent brain areas. (Hence the higher incidence of synesthesia in artists and poets who may be more prone to "metaphor"). Synesthesia might also provide a valuable experimental lever for exploring the neural basis of metaphor; e.g., we say "disgusting" and make an expression when we smell feces, a process mediated by the ventromedial frontal. It is a coincidence that social and moral disgust use the same word and expression, also involving ventromedial frontal?