The evaluation was conducted by the State Toxicology Laboratory of North Dakota State University and the Insurance Institute for Highway Safety. Volunteers from the North Dakota and Minnesota highway patrols served as subjects and tube readers for the study. A total of eight disposable screening devices from four manufacturers were evaluated. The devices tested were the Alcolyser H, Alcolyser (Iowa type), Alcolyser 100, Becton-Dickinson Device 1, Becton-Dickinson Device 2, the Kitigawa Drunk-O-Tester, the Sober-Meter SM-1, and the Sober-Meter SM-6. All of the disposable screening devices tested are similar in design and operation. They are commonly used for prearrest screening tests to indicate whether the blood alcohol concentration of a suspect is above or below the legal limit. Subjects blow either into a balloon from which the breath sample is passed through the tube containing alcohol-sensitive reagent, or directly through the tube into a balloon. The reagent changes color from yellow-orange to green in proportion to the concentration of alcohol in the breath. Study results indicate that these devices produce high numbers of erroneous findings. Both false positive and false negative readings can be obtained with the tested devices. Footnotes, 11 references, appendixes of State laws and technical data, 12 tables, and 14 figures are provided. (Author abstract modified)