Spanish-English bilinguals were asked to determine whether a string of letters formed a word in their languages. Three separate conditions were used: two conditions blocked by language and a mixed-language condition. Some of the words were cognates, words with the same spelling and meaning in the two languages. There were no differences between blocked and mixed conditions; cognates were responded to with equal facility in all conditions, and there was an interaction of Cognates and Noncognates by Weak and Strong Language. These results were interpreted as supporting a direct graphemic lexical access model.