LAY, Chief Judge.

Matthew Sylvester Two Bulls was charged with aggravated sexual abuse, 18 U.S.C. § 1153, 2241(a)(1) (1989), and sexual abuse of a minor, 18 U.S.C. § 1153, 2243 (1989), arising out of the rape of a fourteen-year-old girl on the Pine Ridge Indian Reservation in South Dakota. The police seized the underwear the girl was wearing before and after the incident. The Federal Bureau of Investigation (FBI) laboratory isolated the semen stain on her underwear by using a scientific technique called DNA (Deoxyribonucleic Acid) profiling. After testing Two Bulls' blood, the government concluded that there was a very high probability that the semen on the underwear came from Two Bulls. Before trial, Two Bulls made a motion for a suppression hearing challenging the admissibility of that evidence. At the pre-trial hearing the district judge ruled, after hearing the testimony of the government's first witness, that it had been sufficiently established that DNA evidence was generally accepted by the scientific community so that the evidence could be presented to the jury.

After the hearing, Two Bulls entered a conditional guilty plea, pursuant to a plea agreement, to a superseding Information charging sexual abuse in violation of 18 U.S.C. § 1153, 2242(1) (1989). He was sentenced to 108 months in prison followed by two years of supervised release. The sentence was delayed and Two Bulls was discharged on bond pending this appeal.

On appeal, Two Bulls argues that the trial court erred because it applied Federal Rule of Evidence 702 in determining the admissibility of the DNA evidence instead of using the test in Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), or a more rigid standard. He argues that the district court violated his due process because the pre-trial suppression hearing was incomplete.

This is a case of first impression in the federal circuit courts. See United States v. Jakobetz, 747 F. Supp. 250 (D.Vt. 1990). It is generally conceded that DNA profiling is relatively new and has been the subject of controversy in both the legal and scientific fields. Several state courts, however, have admitted DNA evidence. Andrews v. State, 533 So.2d 841 (Fla.Dist.Ct.App. 1988) (finding DNA evidence admissible using a relevancy test); Caldwell v. State, 260 Ga. 278, 393 S.E.2d 436 (1990) (admitting DNA evidence but not population statistics); Cobey v. State, 80 Md. App. 31, 559 A.2d 391 (1989) (admitting DNA evidence using Frye test); State v. Schwartz, 447 N.W.2d 422 (Minn. 1989) (admitting DNA evidence if tests performed properly); People v. Castro, 144 Misc.2d 956, 545 N YS.2d 985 (Sup.Ct. 1989) (using three step test to determine whether to admit DNA evidence); People v. Wesley, 140 Misc.2d 306, 533 N.Y.S.2d 643 (Albany County Ct. 1988) (admitting DNA evidence under the Frye test); State v. Pennington, 327 N.C. 89, 393 S.E.2d 847 (1990) (finding DNA evidence admissible); State v. Ford, ___ S.C. ___, 392 S.E.2d 781 (1990) (admitting DNA evidence and, because technique generally accepted, future Frye hearings unnecessary); Glover v. State, 787 S.W.2d 544 (Tex.Ct.App.) (admitting DNA evidence using the Frye test), review granted (1990); Spencer v. Commonwealth, 238 Va. 275, 384 S.E.2d 775 (1989) (using reliability test to find DNA evidence admissible) (case one), cert. denied, ___ U.S. ___,

110 S.Ct. 759, 107 L.Ed.2d 775 (1990); Spencer v. Commonwealth, 238 Va. 295, 384 S.E.2d 785 (1989) (admitting DNA evidence) (case two), cert. denied, ___ U.S. ___, 110 S.Ct. 1171, 107 L.Ed.2d 1073 (1990). Although DNA analysis has been used in forensics only recently, it has been used for several years in diagnostics. Andrews, 533 So.2d at 848-49; Caldwell, 260 Ga. at 286, 393 S.E.2d at 441; Ford, ___ S.C. at ___, 392 S.E.2d at 783. It has also been used in determining parentage. In re Baby Girl S., 140 Misc.2d 299, 532 N.Y.S.2d 634 (Sur. Ct. 1988) (finding it unnecessary to have hearing on DNA evidence admissibility when state statute provided for admission of blood genetic marker tests).

The Congressional Office of Technology Assessment has found that DNA tests are valid and reliable in forensics when performed and analyzed properly by skilled personnel. U.S. Congress, Office of Technology Assessment, Genetic Witness: Forensic Uses of DNA Tests 7-8, OTA-BA-438 (Washington, D.C.: U.S. Government Printing Office, July 1990). Commentators have also stated that "[t]here is nothing controversial about the theory underlying DNA typing. Indeed, this theory is so well accepted that its accuracy is unlikely even to be raised as an issue in hearings on the admissibility of the new tests." Thompson Ford, DNA Typing: Acceptance and Weight of the New Genetic Identification Tests, 75 Va.L.Rev. 45, 60 (1989) (emphasis added). These same commentators have stated that:

It would ease the burden on trial lawyers and triers of fact to make proper implementation a threshold issue for the admissibility of DNA typing tests. Before the test offered by a particular laboratory is admitted, there should be a showing, during an evidentiary hearing, that the specific protocol employed by the laboratory is accepted as reliable by disinterested scientists familiar with the procedure. In routine cases, then, the attorneys could focus their attention on the tractable question of whether an accepted protocol was accurately followed instead of the enormously more difficult question of whether the protocol itself is good or bad.

Id. at 58 (emphasis added).

Two Bulls asserts that a three step test should be used to determine the admissibility of DNA evidence similar to the test used in Castro. In Castro, the court stated that the three step analysis would aid in evaluating and resolving the admissibility issue. Castro, 144 Misc.2d at 959, 545 N.Y.S.2d at 987. The three steps are (1) whether the scientific community generally accepts the theory that DNA tests produce reliable results; (2) whether techniques currently exist in DNA testing that are generally accepted by the scientific community and which are capable of producing reliable results; and (3) whether the laboratory that performed the tests used these techniques in analyzing the samples in this case. Id. at 959, 545 N.Y.S.2d at 987. In Castro, the court focused on resolving the third question. The court observed that "'[p]erhaps the most important flaw in the Frye test is that by focusing attention on the general acceptance issue, the test obscures critical problems in the use of a particular technique.'" Id. at 960, 545 N.Y.S.2d at 987 (quoting Giannelli, The Admissibility of Novel Scientific Evidence: Frye v. United States, A Half-Century Later, 80 Colum. L.Rev. 1197, 1226 (1980)). The court stated that:

It is the view of this court that given the complexity of the DNA multi-system identification tests and the powerful impact that they may have on a jury, passing muster under Frye alone is insufficient to place this type of evidence before a jury without a preliminary, critical examination of the actual testing procedures performed in a particular case.

Id. at 960, 545 N.Y.S.2d at 987.

The government argues that Castro stands alone and provides too stringent a standard, necessitating long drawn out testimonial procedures before trial. The government urges that Rule 702 creates a liberal rule of admissibility which now supersedes Frye and which is contrary to the Castro standards. We read it differently. The Frye Court stated that scientific evidence should be distinguished according to whether it was in an experimental or demonstrative stage. Frye, 293 F. at 1014.

In discussing the admissibility of DNA evidence, we find that Frye and Rule 702 both require that a proper foundation be laid for any scientific testing or laboratory procedure. See United States v. Distler, 671 F.2d 954, 961-62 (6th Cir.) (admitting gas chromatography analysis), cert. denied, 454 U.S. 827, 102 S.Ct. 118, 70 L.Ed.2d 102 (1981); United States v. Cyphers, 553 F.2d 1064, 1072 (7th Cir.) (finding microscopic comparison of hair samples admissible), cert. denied, 434 U.S. 843, 98 S.Ct. 142, 54 L.Ed.2d 107 (1977); United States v. Franks, 511 F.2d 25, 33 (6th Cir.) (admitting voice print analysis), cert. denied, 422 U.S. 1042, 95 S.Ct. 2654, 45 L.Ed.2d 693 (1975). Regardless of which rule may be followed, we feel Rule 702 and Frye both require the same general approach to the admissibility of new scientific evidence. Neither rule should permit speculative and conjectural testing which fails normal foundational requirements necessary for the admissibility of scientific testimony or opinion.

For example, as we discuss later, hypnosis evidence is inadmissible unless certain procedures are carried out, see Sprynczynatyk v. General Motors Corp., 771 F.2d 1112, 1122-24 (8th Cir. 1985), cert. denied, 475 U.S. 1046, 106 S.Ct. 1263, 89 L.Ed.2d 572 (1986), and polygraph testing is still inadmissible until there exists greater scientific reliability. See United States v. Alexander, 526 F.2d 161, 163-64 (8th Cir. 1975). Blood tests are inadmissible unless foundational evidence establishes that the testing procedures are reliable. Gardner v. Meyers, 491 F.2d 1184, 1189 (8th Cir. 1974). Any experimental evidence requires certain proofs of similarity or identity to be deemed admissible.

The Castro court held that DNA evidence was admissible under Frye because (1) DNA identification was generally accepted by the scientific community and (2) identification techniques were generally accepted and were capable of producing reliable results. People v. Castro, 144 Misc.2d at 979, 545 N.Y.S.2d at 999. The court also found that a pre-trial hearing should be held to determine whether the laboratory used accepted techniques to yield results reliable enough to be admitted as a question of fact for the jury. Id. We fail to see that this analysis is sui generis or that it in any way allows intrusion of the court into deciding problems of weight vis a vis admissibility. When the evidence is so prejudicial and the admissibility seriously challenged, trial courts routinely hear motions in limine preliminary to the offer at trial.

Because DNA evidence is so new and the resulting prejudice to the defendant is sufficiently great, it is imperative that the court satisfy itself that there exists a sufficient foundational basis as to the overall admissibility of the evidence. This must be done before the government exposes the jury to the lab results. If the court has explored only scientific acceptability and the reliability of acceptable testing procedures in camera, and then, at trial the government fails to show that the lab tests did conform to reliable procedures, the court would have to exclude the evidence for lack of foundation. In doing so, the resulting prejudice to the defendant would be obvious. Notwithstanding the fact that an objection is sustained and the evidence excluded, aside from valuable trial time wasted, the jury would be exposed to prejudicial proofs and left to speculate as to why the defendant opposed the ultimate result.

These general principles apply to the admissibility of any questionable opinion evidence, in which proper foundation as to acceptability as well as to reliability is a focal issue. This approach, whether it be under Rule 702 or Frye, should require the court to satisfy itself that the evidence meets all three tests laid out in Castro. Depending on the circumstances, the court may exercise its discretion in allowing preliminary cross examination by the defendant or to allow on voir dire the defendant's counter evidence. We agree with the government that generally such counter evidence will go only to the weight and not to the admissibility. However, this is not always true. Generally, in determining the scientific acceptability and reliability of testing procedures, the court should hear both sides of the issue. Simply because one expert may verify such scientific acceptability does not make it true. For example, the court would not accept and feel bound by the testimony from one polygrapher that polygraph tests were scientifically acceptable.

Although several courts have found DNA evidence to be admissible because it is reliable and generally accepted, many cases state that DNA evidence remains subject to attack based on prejudice, relevancy, and laboratory procedures. See Pennington, 327 N.C. at ___, 393 S.E.2d at 854; Ford, ___ S.C. at ___, 392 S.E.2d at 784. It seems only logical that if scientific evidence is questionable in terms of acceptability and reliability that the district court should hear both sides of the question before ruling. Certainly one government witness testifying that new evidence is acceptable does not necessarily make it so.

This circuit has applied this reasoning to expert evidence dealing with polygraph tests, Alexander, 526 F.2d at 163-64, and to opinion evidence discussing addictive gambling, United States v. Lewellyn, 723 F.2d 615, 619 (8th Cir. 1983). In Lewellyn, the court stated that reliability was one of the most important factors in considering whether the scientific principle met the Frye test. Id. In Sprynczynatyk, the court found that hypnotically enhanced testimony was admissible if its reliability was established and if it was not overly prejudicial. Sprynczynatyk, 771 F.2d at 1123. The court stated that the trial court should consider (1) whether procedural safeguards had been used, (2) whether hypnosis was appropriate for the type of memory loss involved, and (3) whether there was any corroborating evidence. Id. The court also stated that once the trial court decided the evidence was reliable then it had to determine whether the probative value of the evidence outweighed its prejudicial effect. Id.

We hold that it was error for the trial court to determine the admissibility of the DNA evidence without determining whether the testing procedures used by the FBI lab in this case were conducted properly. In weighing the overall admissibility of such evidence, the court should hear testimony from experts on both sides as to the scientific acceptability and reliability of any novel scientific tests. The trial judge should rule as a matter of law (1) whether the DNA evidence is scientifically acceptable, (2) whether there are certain standard procedures that should be followed in conducting these tests, and (3) whether these standards were followed in this case. Cf. Caldwell, 260 Ga. at 286, 393 S.E.2d at 441. If the trial court is preliminarily satisfied that these requirements have been met the evidence should be admitted and the jury should be allowed to determine the weight that should be allocated to it.

We order Two Bulls' conviction vacated and the conditional plea set aside. We remand the case to the trial court with instructions to hold an expanded pre-trial hearing on the admissibility of the DNA evidence. The trial court is to decide (1) whether DNA evidence is generally accepted by the scientific community, (2) whether the testing procedures used in this case are generally accepted as reliable if performed properly, (3) whether the test was performed properly in this case, (4) whether the evidence is more prejudicial than probative in this case, and (5) whether the statistics used to determine the probability of someone else having the same genetic characteristics is more probative than prejudicial under Rule 403. Cf. Schwartz, 447 N.W.2d at 428 (limiting use of population frequency statistics because "juries in criminal cases may give undue weight and deference to presented statistical evidence and we are reluctant to take that risk"); State v. Joon Kyu Kim, 398 N.W.2d 544, 548 (Minn. 1987) (suppressing statistical population frequency evidence because it may have "`potentially exaggerated impact on the trier of fact'" (quoting State v. Boyd, 331 N.W.2d 480, 482 (Minn. 1983))).

It is so ordered.