On January 27, 1989, the nude body of Pamela Smith was found floating in a barrow pit in Pope County. Mrs. Smith had been missing for four days. The body had been in the water for about twenty-four hours. Death was attributed to a loss of blood from multiple stab wounds. There were numerous abrasions and contusions on the body and evidence of sexual abuse.
In July appellant Alfred Palmer consented to the testing of a bedroom in his house with luminol. The results were positive. Two and a half years later, on January 22, 1992, luminol tests were again conducted in the bedroom, this time pursuant to a search warrant, again with positive results. Palmer was arrested the same day. On January 24 he was charged with the capital murder of Pamela Smith.
*697Palmer filed motions in limine to exclude the luminol test results at trial. After a hearing those motions were denied. Palmer was not found guilty of either capital or first degree murder, but was convicted of second degree murder and sentenced to twelve (12) years in the Department of Correction. Palmer appeals on the single point that the trial court erred in admitting the results of luminol testing without additional evidence that the test results were attributable to human blood related to the crime. Finding error in the admission of the luminol test results, we reverse the judgment of conviction and remand for a néw trial.
Two witnesses for the state from the State Crime Laboratory, Donald Lee Smith, criminologist, and Scott Sherrill, forensic serologist, testified regarding the results of the luminol tests. Smith testified that the chemical luminol provides a presumptive test for blood; it can detect minute amounts — one drop of blood in fifteen gallons of water — and is not affected by the passage of time, even centuries. Luminol cannot differentiate between animal and human blood, it registers positive for either and registers “false positives” to other agents: bleach, copper, nickel, cobalt and some plant enzymes.
Mr. Smith performed luminol tests in the northeast bedroom of Palmer’s house and tested several suspicious spots on one wall at eye level. Some were positive. At the northwest corner of the room there was a positive reaction to an area of the wall and floor where blood appeared to have been wiped up. Mr. Smith had no follow-up information relative to the tests nor any evidence that the substance detected by the luminol came from Pam Smith.
Scott Sherrill testified that he examined and tested four samples of wallpaper from the Palmer residence taken in July, 1989. On two he was able to identify human blood but not the blood type; on one he was able to detect blood but could not determine whether it was animal or human; in the remaining sample he could not identify blood of either kind. Tests on scrapings from Alfred Palmer’s vehicle were negative. Mr. Sherrill later examined additional wallpaper samples from the Palmer residence and found no blood. These samples were submitted to the FBI and the report was negative.
Citing Brenck v. State, 311 Ark. 579, 847 S.W.2d 1 *698(1993), the state agrees that it was error to admit the evidence of luminol testing without further corroboration. Brenck had not been decided at the time of Palmer’s trial, but Palmer’s motions in limine to exclude the luminol test results preserved the identical point decided in Brenck and unquestionably the rationale of Brenck would have governed the result. Here, there was conflicting testimony that Pam Smith had been seen in Palmer’s vehicle prior to her disappearance but no proof that she had ever been in Palmer’s house. The Brenck decision settles the issue that the mere presence of human blood by luminol testing without factors which relate that evidence to the crime is not admissible:
Since we have determined that luminol tests done without follow-up procedures are unreliable to prove the presence of human blood or that the substance causing the reaction was related to the alleged crime, we find it was error for the trial court to admit the evidence of luminol testing done by Mr. Smith where there was no follow-up testirig done to establish that the substance causing the luminol reaction was, in fact, human blood related to the alleged crime. See State v. Moody, 573 A.2d 716 (Conn. 1990); see also Lee v. State, 545 N.E.2d 1085 (Ind. 1989); cf. Commonwealth v. Yesilciman, 550 N.E.2d 378 (Mass. 1990):
Brenck v. State, supra, at 594.
Nor can we conclude the error is harmless. There is no contention by the state to that effect. When the state’s proof concerning luminol, including photographs of the bedroom, is weighed against the evidence generally, we cannot say that prejudice is lacking.
Reversed and remanded.