The most critical step in maize domestication (Zea mays ssp. mays) was the liberation of the kernel from the hardened, protective casing that envelops the kernel in the maize progenitor, teosinte 1 . This evolutionary step exposed the kernel on the surface of the ear such that it could be readily utilized as a food source by humans. Here, we show that this key event in maize domestication is controlled by a single gene (teosinte glume architecture; tga1) belonging to the SBP-domain family 2 of transcriptional regulators. The factor controlling the phenotypic difference between maize and teosinte maps to a 1 kilobase region within which maize and teosinte show only six fixed differences in their DNA sequences. One of these differences encodes a non-conservative amino acid substitution and may affect protein function, while the other five differences potentially affect gene regulation. Molecular evolution analyses show that this region was the target of selection during maize domestication. Our results demonstrate that modest genetic changes in single genes can induce dramatic changes in phenotype during domestication and evolution.The origin of the maize ear has been considered one of the greatest mysteries in both crop domestication 3 and plant evolution 4 . While a wealth of botanical and genetic information has identified the wild Mexican grass, teosinte (Zea mays ssp. parviglumis), as the direct progenitor of maize, the profound differences in the structure of the maize and teosinte female inflorescences (ears) has posed a challenge to formulating a compelling model for the developmental and genetic steps involved in this evolutionary transition 3 . At the heart of the problem is the fact that teosinte kernels are tightly encased in structures called cupulate fruitcases, while maize kernels are borne uncovered on the surface of the ear (Figure 1a, b). The strength with which the fruitcase envelops the teosinte kernel and the stony nature of this casing far exceed the relatively flimsy and loosely bound chaff that surrounds the kernels of the ancestors of the other domesticated cereals. Indeed, the stony fruitcase of teosinte had been considered such an obstacle to the use of teosinte as a grain that teosinte was dismissed by some as a possible progenitor of maize 5 . It was argued that the genetic steps to free the grain from this casing and thereby convert teosinte into a useful crop were too complex to have arisen under domestication.Each of the 5 to 12 cupulate fruitcases in a teosinte ear is formed from an invaginated internode (cupule) within which the kernel sits, and a glume that covers the opening of the cupule such that the kernel is completely hidden from view (Figure 1b, d). When mature, the teosinte ear