Proper collection, handling, and storage of the blood alcohol specimen are essential in medicolegal cases involving the question of sobriety. A standard operating procedure is necessary to ensure maximum reliability. Comments are offered on the advantages of using blood specimens in preference to urine or tissue specimens. The use of a conversion factor to obtain a calculated "presumed blood level" can be dangerous. Cautions and suggestions are offered regarding how and from where the blood should be obtained from a living person and during an autopsy. There are certain time limitations for storage of these blood-alcohol specimens. Each laboratory must establish its own limits for reliable storage, given the conditions in that laboratory. Unexpected and confusing results can lead to an erroneous interpretation if history, circumstances, type of injury, and survival time are not all carefully considered. Several possibilities for error in judgment are discussed.