Different stages of the evaluation program were completed by experiments using a Gas Chromatograph Intoximeter, an Alcolmeter pocket model, and Alco-Limiter, an Alcohol Screening Device (ASD), and an Alcolyser 0.80 mg per ml ampules. An in vitro experiment was conducted that evaluated the performance of a Gas Chromatograph Intoximeter in tests using air-alcohol vapor standards covering a wide range of concentration. Detector responses, measured by the simulator technique, the head-space method, and by running tests over a period of time without adjustment for calibration, showed no statistically significant differences. The evaluation of in vivo laboratory studies with human subjects, given known amounts of alcohol ingestion, must carefully differentiate between the different phases of ethyl alcohol metabolism in the human blood when analyzing changes in the breath/alcohol relationship. Sample experimental data showed that the blood/breath alcohol relationship is different at the absortion, distribution, and elimination phases of ethanol metabolism. An experimental illustration of breath alcohol instrument evaluation through in vivo field trials conducted with Alcolyser 0.80 mg. per ml. ampules demonstrated that the resultant qualitative screening tests may be analyzed in two ways: 1) the relationship between stain length and blood alcohol concentration (deduced from the blood alcohol time curve), and 2) the percentage of positive tube readings (fail results) recorded after sorting into specific blood alcohol concentration ranges. Preliminary findings from roadside testing of the Alcolmeter Pocket Model Device by traffic policemen in routine investigations showed a high degree of accuracy, as indicated by a high correlation with duplicate Alcolmeter tests that were made whenever practical. It is concluded that systematic controlled studies can establish the necessary precision for breath alcohol instruments. Nine figures and 22 references are provided.