OPINION OF THE COURT ON RECONSIDERATION
GONZALES, Judge:
Contrary to his pleas, the appellant was found guilty, by a general court-martial composed of officer members, of attempted unpremeditated murder, two specifications of presenting a false official document, housebreaking, solicitation to commit graft, and obstruction of justice in violation of Articles 80, 107, 130, and 134, Uniform Code of Military Justice, 10 U.S.C. §§ 880, 907, 930, and 934 (1988) [hereinafter UCMJ]. The convening authority approved the adjudged sentence of a bad-conduct discharge, confinement for one year, and reduction to Private El.
Before this court, the appellant asserts, inter alia, that the military judge erred by admitting the opinion testimony of the government’s expert witness on the results of a luminol test for hidden blood without a showing that the test was sufficiently rehable. We hold that the military judge erred in admitting the scientific opinion testimony because the government failed to establish the evidentiary reliability of luminol tests for the purpose for which the government sought to use it. However, we further find that the appellant was not prejudiced by this error. We initially issued an opinion in this case on 29 July 1994. We vacated that decision, United States v. Hill, ACMR 9300891 (A.C.M.R. 23 Aug. 1994) (order) (unpub.), on our own motion so that we might more clearly articulate the reasons for our decision.1
I. The Facts
Specialist (SPC) Robinson, the victim of the attempted unpremeditated murder, lived in an apartment in Lawton, Oklahoma, adjacent to Fort Sill. He was scheduled to testify against the appellant in a special court-martial concerning the two specifications of presenting a false official document of which the appellant was found guilty in this case. On 1 December 1992, a few weeks before the appellant’s trial, SPC Robinson was attacked in his off-post apartment when he returned from his unit’s morning physical training (PT). His attacker was a man with a stocky build wearing an Army winter PT uniform, a black wool cap with two small eye slits over his face, and a pair of black leather work gloves. During the incident, the assailant removed both of his gloves to tie SPC Robinson’s hands behind his back with a telephone cord and to tape SPC Robinson’s black wool cap over his mouth.
After SPC Robinson’s hands were tied, the assailant stabbed him once on the right side of his back with such force that the knife broke. He obtained another knife from the kitchen and sliced the left side of SPC Robinson’s neck. The assailant then ran out of the apartment, leaving his gloves behind. Specialist Robinson observed his assailant drive away in a blue Chevrolet Camaro IROC with black louvers and “ground effects” and “IROC” written on the side of the car. This matched the description of the appellant’s car. An examination of the gloves left behind by the assailant revealed the initials “N.H.” inside the right glove and “N. Hill” or “V. Hill” inside the left glove. Based on his observations, SPC Robinson told the Lawton police that he suspected that his attacker was the appellant, Sergeant Norman Hill.
On the evening of the attack, pursuant to a search warrant, the Lawton police seized the appellant’s complete PT uniform from the dryer in the laundry room of his residence. The appellant’s wife indicated that her husband had worn the uniform that morning and that it had been washed. No trace of blood [598]*598or blood stains were visible on the uniform. The uniform items were sent to a criminal laboratory for hidden blood testing.
At the appellant’s court-martial, the defense moved to exclude all references to alleged hidden blood or blood stains on the appellant’s PT uniform-on the basis that such evidence was irrelevant and highly prejudicial under Military Rules of Evidence 402 and 403 [hereinafter Mil.R.Evid.]. Specifically, the defense moved to exclude the testimony of Mr. Darwin Hormann, a senior criminologist with the Oklahoma State Bureau of Investigation. Mr. Hormann performed a luminol test on the appellant’s PT uniform and the result was a “presumptive positive” for blood. However, he was unable to perform a follow-up test confirming the presence of human blood because of the absence of blood crystals or crystalline residue.
The trial counsel, responding to the motion to exclude the results of the luminol test, argued that under Mil.R.Evid. 401, the luminol testimony was relevant. The victim’s wounds were on the left side of his neck and right side of his back. The left cuff of the appellant’s sweatshirt and the right knee of his sweatpants tested presumptive positive for blood. Thus, the luminol test would support an important part of the prosecution: that the blood stains on the appellant’s PT uniform were consistent with the location of blood stains on clothing worn by SPC Robinson’s attacker.
The military judge, in an Article 39(a), UCMJ, session, previewed Mr. Hormann’s testimony to determine whether his testimony and photographs would “contribute to the panel’s understanding of this case” and whether it would be “more prejudicial than probative in this issue.”2 Although no confirmatory test was performed, he ruled that Mr. Hormann’s testimony on the presumptive positive luminol test was admissible to show that there was blood on the appellant’s PT uniform. This included his determination that the testimony’s probative value was not. substantially outweighed by the danger of unfair prejudice.
-II. The Law
Several state courts have mentioned luminol testing, beginning with State v. White, 293 N.C. 91, 235 S.E.2d 55 (1977). There, the North Carolina Supreme Court reversed a trial court’s denial of a defendant’s motion for dismissal of his case, in part because the blood found on the carpet by luminol testing had not been further tested for blood type or otherwise. Id. 235 S.E.2d at 57. Since then, reported cases show that trial judges have admitted testimony on luminol tests with and without evidence of a confirmatory test. In the majority of these eases, including one federal circuit court of appeals case, such testimony has been admitted without any evidence that a confirmatory test had also been performed. See United States ex rel. Savory v. Lane, 832 F.2d 1011, 1013 (7th Cir.1987); Thompson v. State, 768 P.2d 127, 128 (Alaska App.1989); State v. Bible, 175 Ariz. 549, 562, 858 P.2d 1152, 1165 (1993); Zinger v. State, 313 Ark. 70, 852 S.W.2d 320, 322 (1993); Thomas v. State, 312 Ark. 158, 847 S.W.2d 695, 698 (1993); Larimore v. State, 309 Ark. 414, 833 S.W.2d 358, 361 (1992); Johnston v. State, 497 So.2d 863, 870 (Fla.1986); People v. Henne, 165 Ill.App.3d 315, 116 Ill.Dec. 296, 299, 518 N.E.2d 1276, 1279 (1988); State v. Gonzales, 245 Kan. 691, 783 P.2d 1239, 1242 (1989); State v. Greene, 324 N.C. 1, 376 S.E.2d 430, 437 (1989); State v. Brown, 306 N.C. 151, 293 S.E.2d 569, 575 (1982); State v. VanNatta, 506 N.W.2d 63, 71 (N.D.1993); Fisher v. State, 827 S.W.2d 597, 605 (Tex.App.1992).
On the other hand, testimony on luminol testing has also been admitted where there was evidence of a confirmatory test. See Waterhouse v. State, 429 So.2d 301, 304 (Fla. 1983); People v. Hendricks, 145 Ill.App.3d 71, 99 Ill.Dec. 20, 32, 495 N.E.2d 85, 97 (1986); Jimenez v. State, 105 Nev. 337, 775 P.2d 694, 695 (1989); State v. McElrath, 322 N.C. 1, 366 S.E.2d 442, 448 (1988); State v. Wages, 87 Ohio App.3d 780, 623 N.E.2d 193, 196 (1993); Robedeaux v. State, 866 P.2d 417, 425 (Okla.Crim.App.1993).
[599]*599In none of these cases, however, was the issue of the admissibility of luminol testing directly raised on appeal. That did not occur until the Supreme Court of Arkansas decided Brenk v. State, 311 Ark. 579, 847 S.W.2d 1 (1998). In a very clear voice, that court held that luminol testing, without any confirmatory testing, is unreliable. Id. 847 S.W.2d at 9. Thus, that court held that it was error for the trial court to admit expert testimony on the presumptive positive luminol test that was performed, when there was no follow-up testing to establish that the substance causing the luminol reaction was, in fact, human blood related to the alleged murder. Id.; see also Young v. State, 316 Ark. 225, 871 S.W.2d 373, 377 (1994); Palmer v. State, 315 Ark. 696, 870 S.W.2d 385, 386 (1994).
We can find no general consensus among the state appellate courts on the proper standard to apply for the admission of expert testimony on luminol testing. However, the Supreme Court ruled “in light of sharp divisions among the courts regarding the proper standard for the admission of expert testimony,” the Frye “general acceptance” test has been superseded by enactment of the Federal Rules of Evidence [hereinafter Fed. R.Evid.]. Daubert v. Merrell Dow Pharmaceuticals, Inc., — U.S.-,-, 113 S.Ct. 2786, 2792, 125 L.Ed.2d 469 (1993).3 The Court held that the Federal Rules of Evidence provide the standard for admitting expert scientific testimony in a federal trial. Id. at-, 113 S.Ct. at 2794. Specifically, Fed.R.Evid. 702 requires trial judges to ensure that any and all scientific testimony or evidence admitted is not only relevant, but reliable. Id. at-, 113 S.Ct. at 2795. Our Military Rules of Evidence adopted the Federal Rules of Evidence to military use, and Mil.R.Evid. 702 is exactly the same as Fed. R. Evid. 702.4
The Court’s discussion of Fed.R.Evid. 702 in Daubert requires that the subject of the expert’s testimony must be “scientific knowledge,” which establishes the standard for evidentiary reliability (trustworthiness). Id. at- n. 9, 113 S.Ct. at 2795 n. 9. The standard for relevance of the testimony is that it must also “assist the trier of fact to understand the evidence or to determine a fact in issue” (helpfulness). Id. at-- -, 113 S.Ct. at 2795-2796.
When faced with a proffer of expert scientific testimony, the trial judge must exercise some “gatekeeping responsibilities” by making a preliminary assessment of whether the testimony’s underlying reasoning or methodology is scientifically valid and whether it properly can be applied to the facts at issue. Id. at-, -, 113 S.Ct. at 2796, 2800. The trial judge must ensure that an expert’s testimony rests on a reliable foundation and is relevant to the task at hand. Id. at -, 113 S.Ct. at 2799. While we are cognizant that Daubert did not presume to set out a checklist, many factors will bear on this inquiry, including (1) whether the theory or technique in question can be (and has been) tested, (2) whether it has been subjected to peer review and publication, (3) its known or potential error rate, and (4) whether it has attracted widespread acceptance within a relevant scientific community. Id. at---, 113 S.Ct. at 2796-2798.5 The inquiry is a flexible one. Id. at -, 113 S. Ct. at 2797. However, Daubert did not open the courtroom door to all scientific evidence. Bible, 175 Ariz. at 579, 858 P.2d at 1182.
[600]*600Despite the Supreme Court of Arkansas’ definitive ruling on the unreliability of a luminol test without a confirmatory test, we will apply these four factors to our analysis of the reliability of the expert’s testimony under Mil.R.Evid. 702 in a court-martial. See United States v. Rodriquez, 37 M.J. 448, 449 (C.M.A.1993); United States v. Williams, 39 M.J. 555, 557 (A.C.M.R.1994).6
III. Discussion
Applying the first factor of evidential reliability, we must determine whether the luminol technique can be (and has been) tested. Luminol is a mixture of chemical agents that, when sprayed on clothing or other surface, can detect hidden blood. Luminol reacts with the central iron group in blood by giving a glow that can be photographed in the dark. If any portion of the area sprayed emits a sharp and intense, light blue-green fluorescent glow that lasts from ninety seconds to a couple of minutes before a respray, there is a presumptive positive presence for blood. However, this test is unable to determine whether any possible blood present is human or animal. Further testing is necessary to determine whether what caused the reaction is actually blood and, if blood, whether it is human or animal. The actual presence of human blood can only be confirmed if blood crystals or crystalline residue containing hemoglobin remain on the clothing. Washing can remove all blood crystals from the clothing.7
Mr. Hormann testified that he performed a luminol test on the appellant’s PT uniform
and the results indicated a presumptive positive presence of blood on the left cuff of the sweatshirt and right knee of the sweatpants. These two areas emitted a sharp and intense, light blue-green fluorescent glow. However, he could not confirm the presence of human blood on these two areas because no blood crystals remained in the PT uniform.
Mr. Hormann also testified that he performed a second test on the appellant’s PT uniform called orthotolidine, a procedure he has done three to four hundred times. The orthotolidine test is not a confirmatory test for blood, nor is it designed to help reduce the risk of misidentification by luminol. In fact, it is less definitive than luminol. Thus, when an area sprayed with orthotolidine turns blue, there is only a presumptive presence for blood.
The chemical reaction between the orthotolidine and the appellant’s PT sweat suit caused two areas to turn blue — the left cuff of the sweatshirt and the right knee of the sweatpants. Mr. Hormann compared the light blue-green glow areas of the luminol test with the blue stained areas from the orthotolidine test and found that both areas matched. Because of the intensity of not one, but of both tests, Mr. Hormann opined that had he been able to perform a confirmatory test, he would expect it to establish that the reaction to the luminol test was caused by blood.
We note that one court member asked the single most important question concerning the luminol test, “How reliable are these tests?” This question should have been [601]*601posed by the military judge when determining admissibility under Mil.R.Evid. 702, rather than being asked only later by a court member. Mr. Hormann answered, “Luminol is an investigative tool, and any positive test is just the inducement to go on for a confirmatory [test] or further investigation, so we rely on it heavily because the next series of tests will either confirm or not____” His answer limited luminol’s reliability to that of just an investigative tool.
We find that luminol is a technique that can detect hidden blood and that this technique “can be and has been” tested. However, such testing has revealed that luminol is limited to disclosing only a presumptive positive presence for blood, not a confirmatory presence for blood.
The second reliability factor is whether luminol has been the subject of peer review and publication. Mr. Hormann’s testimony included reference to one three-page article that was admitted as Appellate Exhibit XLY. The article was written by Mr. J. Douglas Perkins, who was Mr. Hormann’s luminol test instructor. The article was used by Mr. Hormann to explain that a false positive for blood would result when luminol is sprayed on other substances containing iron, such as dishwasher detergent, Comet cleanser, a penny, and bleach. He testified that while the fluorescent glows from these substances were similar to that for blood, the glows of these substances were different in either intensity, duration, and/or shade of blue. The article had two significant footnotes citing other publications for the historical facts that luminol was first synthesized in 1902 and initially used for medico-legal blood identification in 1937. Therefore, we find that this limited evidence was sufficient to establish that luminol has been the subject of peer review and publication and that the articles concluded that luminol can only show a presumptive positive presence for blood.8
A further indicator of the reliability of the luminol test is its known or potential error rate. The frequency of erroneous results produced by the luminol test is an important component of reliability. Mr. Hormann testified that although he has performed presumptive positive luminol tests twenty to thirty times, he has only performed a followup confirmation test seven or eight times. In these few instances, he confirmed blood five or six times. The intensity of the fluorescence of the glow, its duration, and its color are the three characteristics of the luminol test that are personally observed to determine a presumptive presence for blood. Blood will emit a sharp and intense, light blue-green glow lasting from ninety seconds to three minutes. We find that these three characteristics are the standard that must be satisfied before a presumptive positive for blood can be made. According to Mr. Hormann, this standard was met in this case, and we so find.
Mr. Hormann also indicated that the error rate was 10 percent, but using his own numbers, we calculate the range of error is from 14 percent (one out of seven) to 37.5 percent (three out of eight). We note that these percentages may not be a true reflection of luminol’s known or potential error rate because of the few confirmation tests that Mr. Hormann performed during his four years as a forensic serologist. Nevertheless, in the absence of evidence to the contrary, we find that luminol’s known or potential error rate in this ease was no better than that indicated by Mr. Hormann’s experiences. We further find that a 14-to-37.5 percent rate of error indicates that luminol can only disclose a presumptive positive presence for blood; not a confirmatory presence.
The final factor in determining reliability is the particular degree of acceptance of luminol within a relevant scientific community. Mr. Hormann testified that the luminol test is a widely used crime scene and investigative tool. It is accepted nationwide. The Oklahoma State Bureau of Investigations started using luminol in 1986, and has continued to use it to the present. Thus, we find that luminol testing has gained acceptance in the criminology community, but only as a [602]*602preliminary test for the presumptive positive presence of hidden blood.
IV. Decision
The record clearly demonstrates that all four of the safeguards discussed by the Supreme Court in Daubert to evaluate whether an expert’s scientific testimony rests on a reliable foundation were met in this case. However, the evidence indicated that luminol is reliable only to show a presumptive positive presence for blood and not to confirm the presence of blood. We find that the government, as the party offering the expert testimony, failed to meet the Daubert reliability standard for the admission of expert testimony because it showed no more than that a luminol test reveals only the presumptive positive presence of blood. Accordingly, we hold that the military judge erred in admitting the expert testimony on luminol testing of the appellant’s PT uniform to show that there was blood on it.
Because our determination on the reliability issue is favorable to the appellant, we need not address the issue of relevancy under MiLR.Evid. 402, 403, and 702.
Having found error, we must now test for prejudice to the appellant on the offenses of attempted unpremeditated murder and housebreaking. UCMJ art. 59(a). Error not of a constitutional dimension may be found harmless only upon the determination either that the court members were not influenced by it, or that the error had but a slight effect on the resolution of the issues in the case. United States v. Mann, 21 M.J. 706, 710 (A.F.C.M.R.1985), aff'd, 26 M.J. 1, cert. denied, 488 U.S. 824, 109 S.Ct. 72, 102 L.Ed.2d 49 (1988).
Because of the question by a court member on the reliability of luminol, as noted above, we find that the court members did not give it more weight than it deserves as an investigative tool. Our review of the court members’ other questions convinces us that they had a complete and correct understanding of the meaning and effect of presumptive positive results, false positive readings, and luminol’s inability to distinguish between human and animal blood. Unlike a similar situation in State v. Moody, 214 Conn. 616, 573 A.2d 716, 723 (1990), we find that the court members were not misled.
We also find that the government presented sufficient, credible evidence in the form of testimony by SPC Robinson, Sergeant Gibson, Staff Sergeant Butteris, Sergeant Nicholson, Mr. Kidneigh, Mr. Naranjo, Officer Whitaker, Detective Reid, Detective Whittington, and Doctor Penrod, as well as the physical 'evidence of the black gloves and a car matching the description of the appellant’s car at the scene of the attack, against the appellant’s unbelievable alibi defense, to convince us of the appellant’s guilt beyond a reasonable doubt. Accordingly, we hold that the admission of Mr. Hormann’s testimony was harmless error and that the appellant was not prejudiced by it.
We pause at this point to emphasize that our decision should not be interpreted to exclude luminol tests altogether from court-martial proceedings. The Court of Military Appeals has ruled that testimony on luminol testing is admissible under Mil.R.Evid. 403 when it is introduced for the limited purpose to explain why an investigation focused on an accused, and such testimony is accompanied by an appropriate limiting instruction by the military judge. United States v. Burks, 36 M. J. 447, 452 (C.M.A.1993).
Luminol testimony may also be admissible to corroborate an accused’s confession or the credibility of a witness. United States v. Garries, 19 M.J. 845, 858 (A.F.C.M.R.1985), aff'd, 22 M.J. 288 (C.M.A. 1986), cert. denied, 479 U.S. 985, 107 S.Ct. 575, 93 L.Ed.2d 578 (1986). It may also be introduced to explain how a footprint was found or how a footprint was enhanced for photography. State v. Lewis, 543 So.2d 760, 766 (Fla.App.1989); State v. Bullard, 312 N. C. 129, 322 S.E.2d 370, 373 (1984).
Our decision is based on the government’s failure to prove the reliability of luminol testing by showing that it can confirm the presence of blood. However, if the government can satisfy the reliability element of Mil. R.Evid. 702, luminol testimony may be admissible to establish a major premise of the government’s case.
[603]*603We have carefully considered the remaining assignments of error, including those personally raised by the appellant pursuant to United States v. Grostefon, 12 M.J. 431 (C.M.A.1982), and hold that none warrant relief.
The findings of guilty and the sentence are affirmed.
Senior Judge CREAN and Judge EDWARDS concur.