THE PROPOSITION THAT THERE IS A CORRELATION BETWEEN A NORMAL INDIVIDUAL'S BAC AND HIS OR HER EXTENT OF NEUROLOGICAL IMPAIRMENT IS GENERALLY ACCEPTED IN SCIENTIFIC AND LEGAL COMMUNITIES. MOST STATE STATUTES PRESUME THAT A DRIVER WITH A BAC OF LESS THAN 0.05 PERCENT WEIGHT TO VOLUME IS NOT INTOXICATED AND THAT A DRIVER WITH A BAC GREATER THAN OR EQUAL TO 0.10 PERCENT WEIGHT TO VOLUME IS INTOXICATED. BIOLOGICAL VARIABLES INTRODUCE THE POSSIBILITY OF SUBSTANTIAL ERROR IN ANY ESTIMATION OF BAC FROM BREATH ALCOHOL ANALYSIS. CONSEQUENTLY, BREATH ANALYSIS INSTRUMENTS DO NOT ALWAYS DETERMINE BAC WITH THE CERTAINTY REQUIRED BEFORE PLACING A DEFENDANT IN THE POSITION OF HAVING TO REBUT A PRIMA FACIE CASE. COMPOUNDING FALLACIES IN BASIC SCIENTIFIC ASSUMPTIONS ABOUT BREATH ALCOHOL ANALYSIS IS THE ADDITIONAL CONCERN THAT ALL EXPERIMENTAL DATA HAVE BEEN GENERATED UNDER HIGHLY CONTROLLED LABORATORY CONDITIONS WITH PRESUMABLY COOPERATIVE SUBJECTS. NO EXPERIMENT HAS BEEN PERFORMED UNDER OPERATIONAL CONDITIONS WHERE BREATH SAMPLES OF SUSPECTED DRUNK DRIVERS ARE COMPARED TO ACTUAL BAC'S OF THOSE DRIVERS. UNTIL THIS TYPE OF WORK IS DONE, THE PREDICTION OF BAC FROM BREATH ALCOHOL ANALYSIS CANNOT BE REGARDED AS SCIENTIFICALLY RELIABLE. EXPERIMENTAL STUDIES OF BAC LEVELS ARE NOTED. FOOTNOTES AND CASE LAW CITATIONS ARE INCLUDED. (DEP)