This study represents several years' work on devising tests to measure children's relative proficiency in two languages or speech varieties. A preliminary oral proficiency test in English and Spanish was administered to 45 bilingual students, of whom 41 took both versions. On the basis of difficulty and significant tendencies, 32 items were selected as the revised version of the vocabulary test and 16 items as the revised version of the grammar test. Five independent variables (grade, sex, years of schooling in the United States, place of birth, and reported language use with family) were used in a univariate analysis of test scores in order to determine whether they contributed significantly to the variance in those scores. In most cases the significant contribution of the independent variables were in the expected direction (e.g., performance in English increased with grade). Modification of the grammar test is required in order to focus students' responses on the intended grammatical structures. Administration of both parts of the proficiency test on grammar to larger and different samples of pupils, and the correlation of test results with other independent variables, are needed to establish the validity of the test. (Author/AM)