As part of the evaluation of a 5-year longitudinal research and development project in individualized language learning, several alternative methods for testing oral English production were tried out. The Bilingual Syntax Measure was selected for adaptation because of the relative effectiveness of its visual component in eliciting responses. Adaptation of the test for native Thai-speaking upper elementary school children included modification of the content of the questions as well as the scoring procedure. A stratified random sample of 100 seventh grade students were tested. Individual tests were tape-recorded, randomized, and prepared for rating. Raters included five native speakers of English and one native Thai-speaking English teacher. Both interrater correlations and internal consistency estimates of reliability were acceptable, while predictive validity correlations with measures of other language skills were highly significant. Content validity is claimed in that the test provides sufficient latitude for responses to go well beyond mere manipulation; questions require factual information about the pictures, inferences regarding causal relationships implied in the pictures, and inferences based on a common external frame of reference. An example of the measure is appended. (Author/AMH)