The purpose of this research was to develop and to validate a language aptitude battery for use by the Defense Language Institute (DLI) which offers training in over 50 foreign languages. Predictor and criterion measures were collected on DLI students in 12 lan-guages. Items were selected and grouped through canonical factor analysis of an experimental item pool. Validity was established through prediction of grades. These validities were then compared with those of other tests and predictor variables. The results were that the new battery (DLAB) did produce consistently higher valid-ities than did other predictors. Although differential prediction by language was also investigated, evidence that it can be achieved was not found in this stu...