A REVIEW OF EVALUATIONS ON POLYGRAPH RESULTS AND INTERPRETATION TECHNIQUES SHOWS THAT LABORATORY TESTING OF NONCRIMINAL SUBJECTS IS NOT AS ACCURATE AS THE TESTING OF SUSPECTS IN REAL LIFE SITUATIONS. THE REASON FOR THIS DEVIATION ACCORDING TO COOK (1968) IS THAT FEELINGS OF GUILT AND FEAR OF DETECTION PRESENT IN FIELD INVESTIGATION SUBJECTS (CRIMINAL SUSPECTS) ARE INVARIABLY MISSING IN LABORATORY SUBJECTS. ON APPLYING POLYGRAPH TESTS TO ACTUAL CRIME INVESTIGATIONS, VARIOUS EXPERTS HAVE REPORTED ACCURACY RESULTS BETWEEN 95 AND 100 PERCENT. A POLYGRAPH STUDY IN INDIA USED A SAMPLE OF 80 SUBJECTS, OF WHICH 40 WERE SUSPECTS IN CRIME CASES AND 40 WERE GOVERNMENT EMPLOYEES. THE SUBJECTS WERE PRESENTED A PACK OF CARDS WITH PHOTOGRAPHS, ASKED TO PICK ONE AND WRITE DOWN ITS NUMBER AND THE NAME OF THE SUBJECT ON THE PHOTOGRAPH, AND THEN RETURN THE CARD TO THE PACK WITHOUT LETTING THE EXAMINER KNOW WHICH CARD THEY HAD SELECTED. THEN THE EXAMINERS PRESENTED THE ENTIRE PACK OF CARDS TO THE SUBJECTS, ASKING THEM TO REVEAL THAT CARD WHICH THEY HAD SELECTED. THE SUBJECTS, HOWEVER, HAD BEEN PREVIOUSLY INSTRUCTED TO GIVE NEGATIVE RESPONSES ON ALL CARDS. POLYGRAMS FOR THE TOTAL SAMPLE WERE ANALYZED AND EVALUATED TO REVEAL 80 PERCENT ACCURACY IN EXAMINERS' CHOOSING OF THE CORRECT CARD. ACCURACY OF DIAGNOSIS IN RESPECT TO VOLUNTEERS AND CRIME SUSPECTS WAS 70 PERCENT AND 90 PERCENT, RESPECTIVELY. TABULAR DATA AND REFERENCES ARE PROVIDED. (RCB)