UDALL, Vice Chief Justice.

Defendant was tried for and convicted of possession of narcotics. Pursuant to a written stipulation entered into by defendant, his counsel and the county attorney before trial defendant submitted to a polygraph (lie-detector) examination. The stipulation also provided that the results of such examination would be admissible at the trial. Accordingly, the polygraph operator was permitted, over objection by defendant to testify to the results of the examination (unfavorable to defendant) at defendant's jury trial. After the jury returned a verdict of guilty and before sentence was entered, the trial court, pursuant to Rule 346 of the Arizona Rules of Criminal Procedure, 17 A.R.S., certified the following question to this court:

"In a criminal case, if prior to trial the defense attorney, on behalf of his client and with his client's consent, and the deputy county attorney agree in a written stipulation that the results of a polygraph test, to be taken by the defendant, will be admissible as evidence at the trial, on behalf of either the State of Arizona or the accused, may the trial court admit the results of the test over the objection of defense counsel?"

Because any case involving admissibility of lie-detector evidence would be one of first impression in Arizona and also because of the particular disposition herein of the question certified the following observations and review of the authorities are set forth for the guidance of the Bar.

The polygraph or lie-detector is a pneumatically operated device which simultaneously records changes in a subject's blood pressure, pulse, respiration rate and depth, psychogalvanic skin reflex (skin resistance to electrical current) and, in some cases, muscular activity. "The basis for the use of the so-called lie-detector * * * is the hypothesis that conscious deception can be deduced from certain involuntary physiological responses in the same manner as physicians diagnose various diseases. The thesis is that lying engenders emotional disturbances which are transmuted into tangible bodily manifestations." The machine itself reflects and records only the subject's physiological responses to the questions propounded by the operator. He then interprets the poly graph (meaning, literally, "many pictures") and determines whether the subject is lying

I Admissibility in General

The first reported American case involving admissibility of lie-detector evidence was Frye v. United States, 54 App. D.C. 46, 293 F. 1013, 34 A.L.R. 145 (1923). Frye, convicted of murder in the second degree, appealed alleging as his sole assignment of error the trial court's refusal to allow an expert to testify as to the results of a systolic blood pressure test to which Frye had submitted. In affirming the conviction and in upholding the trial court's refusal of the proffered testimony the Circuit Court observed:

"Just when a scientific principle or discovery crosses the line between the experimental and demonstrable stages is difficult to define. Somewhere in this twilight zone the evidential force of the principle must be recognized, and while courts will go a long way in admitting expert testimony deduced from a well-recognized scientific principle or discovery, the thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular filed in which it belongs.

"We think the systolic blood pressure deception test has not yet gained such standing and scientific recognition among physiological and psychological authorities as would justify the courts in admitting expert testimony deduced from the discovery, development, and experiments thus far made."

Ten years later the Supreme Court of Wisconsin reached the same result in State v. Bohner, 210 Wis. 651, 246 N.W. 314, 86 A.L.R. 611 (1933). Bohner's conviction for robbery was affirmed and it was held that the trial judge had correctly excluded defendant's offer of lie-detector results. The Wisconsin court added the following:

"We are not satisfied that this instrument, during the ten years that have elapsed since the decision in the Frye Case, the progressed from the experimental to the demonstrable stage." 210 Wis. at 658, 246 N.W. at 317.

And the judicial attitude toward lie-detector evidence expressed in Bohner has not changed markedly in the numerous cases decided since 1933. Thus, in 1961 a New Jersey appellate court was correct in pointing out:

"* * * that there is not a single reported decision where an appellate court has permitted the introduction of the results of a polygraph or lie-detector test as evidence in the absence of a sanctioning agreement or stipulation between the parties."

State v. Arnwine, 67 N.J.Super, 483, 495, 171 A.2d 124, 131 (1961). Consistent with this approach appellate courts have reversed convictions in cases where lie-detector results unfavorable to defendants were placed before the juries inferentially. E.g., State v. Arnwine, supra; People v. Wochnick, 98 Cal.App.2d 124, 219 P.2d 70 (1950). Further, it is uniformly held that a defendant is not permitted to introduce evidence of his willingness to take a lie-detector test. E.g., Commonwealth v. Saunders, 386 Pa. 149, 156-57, 125 A.2d 442, 445-46 (1956). Nor can a defendant's refusal to submit to polygraphic interrogation be shown by the state directly, State v. Kolander, 236 Minn. 209, 52 N.W.2d 458 (1952), or indirectly, People v. Carter, 48 Cal.2d 737, 752, 312 P.2d 665, 674 (1957).

But judicial reluctance to recognize generally the worth of lie-detector evidence in the court room has not been due to mere inertia. For, in affirming a first degree rape conviction, the Oklahoma Criminal Court of Appeals quoted from two leading authorities the following "`* * * factors which occasion the chief difficulties in the diagnosis of deception by the lie-detector technique * * *."

"`(1) Emotional tension —, "nervousness" — experienced by a subject who is innocent and telling the truth regarding the offense in question, but who is nevertheless affected by

"`(a) fear induced by the mere fact that suspicion or accusation has been directed against him, and particularly so in instances where the subject has been extensively interrogated or perhaps physically abused by investigators prior to the time of the interview and testing by the lie-detector examiner; and

"`(b) a guilt complex involving another offense of which he is guilty.

"`(2) physiological abnormalities, such as

"`(a) excessively high or excessively low blood pressure;

"`(b) disease of the heart;

"`(c) respiratory disorders, etc.

"`(3) Mental abnormalities, such as

"`(a) feeblemindedness, as in idiots, imbeciles, and morons;

"`(b) psychoses or insanities, as in manic depressives, paranoids, schizophrenics, paretics, etc.;

"`(c) psychoneuroses, and psychopathia, as among so-called "peculiar" or "emotionally unstable" persons — those who are neither psychotic nor normal, and who form the borderline between these two groups.

"`(4) Unresponsiveness in a lying or guilty subject, because of

"`(a) lack of fear of detection;

"`(b) apparent ability to consciously control responses by means of certain mental sets or attitudes;

"`(c) a condition of "sub-shock" or "adrenal exhaustion" at the time of the test;

"`(d) rationalization of the crime in advance of the test to such an extent that laying about the offense arouses little or no emotional disturbance;

"`(e) extensive interrogation prior to the test.

"`(5) Unobserved muscular movements which produce ambiguities or misleading indications in the blood pressure tracing.'"

Henderson v. State, 94 Okla. Cr. 45, 51-52, 230 P.2d 495, 501-2, 23 A.L.R.2d 1292, cert. denied, 342 U.S. 898, 72 S.Ct. 234, 96 L.Ed. 673 (1951). And in addition to the above enumerated scientific shortcomings of the polygraph technique the following objections to the unrestricted use of its results in the court room have been registered:

(1) The supposed tendency of judges and juries to treat lie-detector evidence as conclusive on the issue of defendants' guilt. See Highleyman, The Deceptive Certainty Of the "Lie Detector", 10 Hastings L.Rev. 47 (1958); Kleinfeld, The Detection of Deception — A Resume, 8 Fed.B.J. 153 (1947).

(2) Lack of standardization of test procedure, (Burack, A Critical Analysis Of The Theory, Method, And Limitations Of The "Lie Detector", 46 J.Crim.L., C. P.S., 414 (1955); Koffler, The Lie Detector — A Critical Appraisal Of The Technique As A Potential Undermining Factor In The Judicial Process, 3 N.Y.L.F. 123 (1957)), examiner qualifications and instrumentation.

(3) Difficulty for jury evaluation of examiners' opinions.

Finally, it appears "* * * that at the present time the technique is not an `accepted' one among the scientists whose approval is a prerequisite to judicial recognition." Inbau and Reid, Lie Detection and Criminal Interrogation, (3rd ed, 1953) at 130. See also Cureton, A consensus As To The Validity Of Polygraph Procedures, 22 Tenn.L.Rev. 728, 739-41 (1953). Of course absolute infallibility is not the standard for admissibility of scientific evidence. But at this time it seems wise to demand greater standardization of the instrument, technique and examiner qualifications and the endorsement by a larger segment of the psychology and physiology branches of science before permitting general use of lie-detector evidence in court. Accordingly, in the absence of a stipulation lie-detector evidence should not be received in an Arizona court for the present.

II Admissibility Upon Stipulation

The first reported decision involving stipulated admissibility of lie-detector results was LeFevre v. State, 242 Wis. 416, 7 N.W.2d 288 (1943). Before the trial defendant submitted to a lie-detector (Keeler polygraph) test the results of which were favorable to him. At trial, however, on objection by the district attorney to defendant's proffer of the test results, the trial judge excluded them. This ruling was upheld on appeal although the conviction was reversed on the grounds of insufficient evidence. Without mentioning the stipulation the Wisconsin Supreme Court simple cited the Bohner case, supra, and said "They were properly excluded." Oddly enough the court then referred to testimony of the district attorney to the effect that the test results indicated defendant was not lying, and remarked that:

"We have the word of the district attorney that those tests were favorable to the defendant. While the findings of these experts were properly excluded from the jury, the district attorney's testimony came in without objection and we regard it as very significant." 242 Wis. at 427, 8 N.W.2d at 293.

Further, a note critical of LeFevre in 1943 Wis.L.Rev. at 430 points out that the district attorney objected to the introduction of the reports by themselves and only on the ground that the examiners, not present in court, should have been called to testify to the results of the examination.

Five years later, however, a California District Court of Appeal held that the test results and the examiner's testimony relating thereto were properly admitted pursuant to a written stipulation. Determination of the value and weight of such evidence was left to the jury. People v. Houser, 85 Cal.App.2d 686, 193 P.2d 937 (1948). It does not appear from the decision whether an objection was made thereto at trial. In affirming Houser's conviction for a lewd and lascivious act the court remarked that:

"It would be difficult to hold that defendant should now be permitted on this appeal to take advantage of any claim that such operator was not an expert and that as to the results of the test such evidence was inadmissible, merely because it happened to indicate that he was not telling the truth. * * *." 85 Cal.App.2d at 695, 193 P.2d at 942.

The next case in this area was Stone v. Earp, 331 Mich. 606, 50 N.W.2d 172 (1951). In Stone, a civil case, plaintiff sought a declaration that he was the legal and equitable owner of a truck and trailer. During the course of the trial the court required the parties to take lie-detector examinations. They consented and the examiner testified that in his opinion the plaintiff was not telling the truth. A decree for defendant was affirmed by the Michigan Supreme Court. The reviewing court, however, pointed out that the trial court committed error (though not prejudicial error, because, aside from the test results, there was a preponderance of evidence in defendant's favor) in receiving the lie-detector results in evidence. And in this connection the court remarked that "* * * whether by voluntary agreement, court direction, or coercion, the results of such tests do not attain the stature of competent evidence." 331 Mich. at 611, 50 N.W.2d at 174.

In 1960 the Iowa Supreme Court held that there was sufficient evidence to affirm a woman's second-degree murder conviction and then disposed of her contention that lie-detector evidence was improperly admitted notwithstanding a stipulation by announcing that:

"We hold the lie-detector evidence was admissible by reason of her agreement."

State v. McNamara, 252 Iowa 19, 104 N.W.2d 568, 574 (1960). Significantly, "at the trial defendant [McNamara] strenuously objected to any evidence regarding the tests on the ground that they were unreliable and prejudicial." 252 Iowa at 28, 104 N.W.2d at 573.

An opposite result was reached by the Supreme Court of New Mexico in the recent case of State v. Trimble, 68 N.M. 406, 362 P.2d 788 (1961). Without extended discussion of the authorities, indeed with no mention at all of the Houser or McNamara decisions, the court reversed defendant's conviction for incest on the ground that "The signing of a waiver did not alter the rule with regard to the admissibility of * * * [lie-detector] evidence." 68 N.M. at 408, 362 P.2d at 789.

In addition to the express holdings in Houser and McNamara admissibility of lie-detector evidence upon stipulation has been indicated by implication in the following cases: State v. Arnwine, 67 N.J. Super. 483, 498, 171 A.2d 124, 132 (1961); Colbert v. Commonwealth, 306 S.W.2d 825, 71 A.L.R.2d 442 (Ky. 1957); Commonwealth v. McKinley, 181 Pa.Super, 610, 123 A.2d 735 (1956); State v. Lowry, 163 Kan. 622, 185 P.2d 147 (1947). For a discussion of unreported trial cases in which lie-detector results were admitted per stipulation see 1943 Wis.L.Rev. at 435; 26 J.Crim.L. C. 262 (1935-36).

Generally speaking, even those experts who warn against admissibility in the absence of a stipulation favor admission of lie-detector evidence upon a proper stipulation. And although polygraphic interrogation has not attained that degree of scientific acceptance in the fields to which it belongs to be admissible at the instance of either the state or defendant (supra, section I of this opinion), it has been considerably improved since Frye v. United States, supra, was decided in 1923. A conservative estimate of the accuracy of such tests is as follows:

(1) In 75-80 per cent of the cases the examination correctly indicates the guilt or innocence of the accused;

(2) in 15-20 per cent of the cases the results are too indefinite to warrant a conclusion by the examiner one way or the other; and

(3) 5 per cent or less is the margin of proven error.

With improvement in and standardization of instrumentation, technique and examiner qualifications the margin of proven error is certain to shrink. "Modern court procedure must embrace recognized modern conditions of mechanics, psychology, sociology, medicine, or other sciences, philosophy, and history. The failure to do so will only serve to question the ability of courts to efficiently administer justice." Chappell, J., concurring in Boeche v. State, 151 Neb. 368, 383, 37 N.W.2d 593, 596, 600 (1949). Although much remains to be done to perfect the lie-detector as a means of determining credibility we think it has been developed to a state in which its results are probative enough to warrant admissibility upon stipulation. Cf., People v. Zavaleta, 182 Cal.App.2d 422, 6 Cal.Rptr. 166, 171 (1960).

Accordingly, and subject to the qualifications announced herein, we hold that polygraphs and expert testimony relating thereto are admissible upon stipulation in Arizona criminal cases. And in such cases the lie-detector evidence is admissible to corroborate other evidence of a defendant's participation in the crime charged. If he takes the stand such evidence is admissible to corroborate or impeach his own testimony.

The "qualifications" are as follows:

(1) That the county attorney, defendant and his counsel all sign a written stipulation providing for defendant's submission to the test and for the subsequent admission at trial of the graphs and the examiner's opinion thereon on behalf of either defendant or the state.

(2) That notwithstanding the stipulation the admissibility of the test results is subject to the discretion of the trial judge, i.e. if the trial judge is not convinced that the examiner is qualified or that the test was conducted under proper conditions he may refuse to accept such evidence.

(3) That if the graphs and examiner's opinion are offered in evidence the opposing party shall have the right to cross-examine the examiner respecting:

a. the examiner's qualifications and training;

b. the conditions under which the test was administered;

c. the limitations of and possibilities for error in the technique of polygraphic interrogation; and

d. at the discretion of the trial judge, any other matter deemed pertinent to the inquiry.

(4) That if such evidence is admitted the trial judge should instruct the jury that the examiner's testimony does not tend to prove or disprove any element of the crime with which a defendant is charged but at most tends only to indicate that at the time of the examination defendant was not telling the truth. Further, the jury members should be instructed that it is for them to determine what corroborative weight and effect such testimony should be given.

The case as certified is remanded for action consistent with this opinion.

BERNSTEIN, C.J., and STRUCKMEYER, JENNINGS and LOCKWOOD, JJ., concur.