The opinion
HUNTER, Justice,
dissenting.
I must respectfully dissent to the opinion of the majority of this Court. The results of the neutron activation analysis lacked the reasonable scientific certainty which is requisite to the admission of expert scientific testimony. It may be that neutron activation analysis is or will become a commonly accepted and consistently reliable scientific procedure. The state has not produced sufficient evidence here, however, to justify the evidentiary use of its methods and results in the instant case.
That is not to say that FBI Agent Riley was not a properly qualified expert witness on the subject. His qualifications were amply demonstrated, including his role in the initial neutron activation research, his lectures at George Washington University, his •work at the Oakridge Associates and the National Bureau of Standards, and his thirteen years in the laboratories of the Federal Bureau of Investigation. His status as an expert does not legitimize the methods and results of neutron activation analysis, however. The scientific community of course abounds with experts whose particular theses await the future acceptance or rejection of peers.
So it is that an expert’s testimony that a certain conclusion is possible is without probative value. Palace Bar, Inc. v. Fearnot, (1978) 269 Ind. 406 , 381 N.E.2d 858 . There, Pivarnik, J., stated for this Court:
“A doctor’s testimony can only be considered evidence when he states that the conclusion he gives is based on reasonable medical certainty that a fact is true or untrue. A doctor’s testimony that a certain thing is possible is no evidence at all. *135 His opinion as to what is possible is no more valid than the jury’s own speculation as to what is or is not possible. Almost anything is possible, and it is thus improper to allow a jury to consider and base a verdict upon a ‘possible’ cause of death.” 269 Ind. at 415 , 381 N.E.2d at 864 [emphasis in original].
Here, following his lengthy explanation of the manner and means by which his assessment had been made, Agent Riley was able to state only that the bullet taken from the victim “could have come” from the same box of ammunition as two of the three lead bullets seized from defendant. It contravenes Palace Bar, Inc. v. Fearnot, supra, to hold that the very tentative results of Agent Riley’s neutron activation analysis are admissible; his conclusion simply lacks the reasonable scientific certainty to imbue it with probative value. Id.
Furthermore, defendant has cited authority wherein the results of neutron activation analysis have been rejected. He states in his brief:
“In the case of State v. Holt, (1969) 17 Ohio St.[2d] 81, 246 N.E.2d 365 , the State secured a conviction based in large part on evidence that hair samples analyzed by neutron activation analysis matched. The Ohio Court reversed the conviction stating that the evidence concerning the neutron activation analysis and its supposed results should not have been admitted over the defendant’s objection.”
In State v. Holt, (1969) 17 Ohio St.2d 81 , 246 N.E.2d 365 , the Ohio Supreme Court reversed the defendant’s conviction on the lack of certainty of the expert’s ultimate conclusion. The Court explained:
“The following question was asked the witness by the state:
“ ‘Based on the Neutron Activation Analysis * * * do you have an opinion based on reasonable scientific certainty as to the similarity or dissimilarity of these specimens?’ (Emphasis supplied.)
“His response was:
“ ‘The samples * * * are similar and are likely to be from the same source.’ (Emphasis supplied.)
“The term ‘likely’ is a weaker one than ‘reasonably certain’ or ‘probability’ and carries appreciably less weight. Thus, in the case of Ottgen v. Garey, 41 Ohio App. 499, 505 , 181 N.E. 485, 487 , Lloyd, J., correctly stated:
“ ‘The word “likelihood” imports something less than reasonable certainty.’
******
“Because of the witness’s educational background and his apparent prestige, his testimony undoubtedly made an impression on the jury and was accorded greater weight than it was entitled to. Defendant was charged with a heinous and dastardly crime, and the jury apparently had difficulty in reaching a guilty verdict. “Our conclusion is that the testimony of the witness, respecting the comparative hair analysis, the procedures used in such analysis and his ultimate conclusion based thereon, did not reach that degree of certainty which the law demands, and his testimony should have been rejected; the failure to do so constituted error prejudicial to defendant.” 17 Ohio St.2d at 85-86 , 246 N.E.2d at 367-68 [emphasis in original].
Here, the statement that the bullets “could have come” from the same box of ammunition does not even approach the “likelihood” at issue in State v. Holt, supra. The lack of certainty in Agent Riley’s ultimate conclusion is further emphasized by his testimony that as many as 100,000 bullets are produced from the same batch of lead, that 200 boxes of bullets of similar composition would in turn result, and that all retailers in a particular geographic area might consequently market bullets of similar composition. Furthermore, Agent Riley stated that “there are other leads that are going to be manufactured some place along the way that can be very close to this composition, if not the same.” In stating that the bullets “could have come” from the same source, he offered little more than an educated guess which, under Palace Bar, Inc. v. Fearnot, supra, had no probative value. Consequents ly, it was error for the trial court to admit the expert testimony over defendant’s con *136 tinuing objection, duly acknowledged by the trial court, that the methods and results lacked the requisite scientific reliability and acceptance. Id. See also, Zupp v. State, (1972) 258 Ind. 625 , 283 N.E.2d 540 .
It is noted that a compelling case for the general admissibility of neutron activation analysis testimony was made in United States v. Stifel, (6th Cir. 1970) 433 F.2d 431 . It is the particular circumstances of this case, which culminated in the very tentative conclusion of Agent Riley which renders the expert testimony inappropriate here. Those peculiar factors are not wholly unlike those considerations which prompt our refusal to admit the results of polygraph examinations. Zupp v. State, supra. Consequently, it is not implied that the general admissibility of expert testimony regarding neutron activation analysis is proscribed.
Here, however, the testimony should not have been admitted. And, for much the same reasons that prompted the reversal of judgments in Palace Bar, Inc. v. Fearnot, supra, and State v. Holt, supra, the trial court’s error cannot be regarded as necessarily harmless.
The evidence regarding defendant’s complicity in the robbery was wholly circumstantial in nature. Defendant, in a signed statement introduced into evidence, indicated that he drove Armen Sylvester to the Nehi-Royal Crown building without knowledge of his passenger’s intent to commit robbery. He contended that he left the scene unaware that a robbery had occurred. He attributed the fact that his tires squealed as he drove away to the fact that the car was unfamiliar to him and the presence of snow, a condition which was corroborated by other evidence. Witnesses at the scene estimated the top speed of the auto as it was driven away was thirty-five miles an hour. Policemen who viewed the auto a short time later estimated it was traveling at approximately thirty miles per hour and indicated that it was not being driven in an erratic fashion. The stolen currency was found in Sylvester's possession. The only other evidence regarding defendant’s participation in the robbery was the four bullets found in his pocket. Jones stated that the jacket belonged to his brother and that he had no idea that the bullets were present.
In determining whether the erroneous admission of evidence warrants reversal, the question is not whether there is sufficient evidence to support the conviction absent the improper evidence. Rather, the focal point of our inquiry is whether the evidence was likely to have had a prejudicial impact on the jury’s conclusion that, beyond a reasonable doubt, the defendant was guilty. Sulie v. State, (1978) 269 Ind. 204 , 379 N.E.2d 455 ; Mitchell v. State, (1972) 259 Ind. 418 , 287 N.E.2d 860 ; Otto v. State, (1980) Ind.App., 398 N.E.2d 716 .
Here, the erroneous admission of the neutron activation analysis results was not cumulative evidence. Nor are we confronted with other evidence which points so unerringly and overwhelmingly toward defendant’s guilt that the impact of the erroneously admitted evidence was necessarily minimal. See Whitten v. State, (1975) 263 Ind. 407 , 333 N.E.2d 86 . Rather, as in White v. State, (1971) 257 Ind. 64 , 272 N.E.2d 312 , where this Court found such prejudicial impact, the evidence is wholly conflicting in nature. And while in White v. State, supra, there were three eyewitness identifications of defendant as the perpetrator, we have only the circumstantial evidence of the state vis-a-vis the credibility of the defendant to measure here. Inasmuch as the jury’s decision rested on circumstantial and conflicting evidence, and in turn on an evaluation of the credibility of witnesses, we cannot conclude that the erroneous admission of the expert testimony had no prejudicial effect on the jury’s conclusion that defendant was guilty beyond a reasonable doubt. Compare, Mitchell v. State, supra; White v. State, supra; Otto v. State, supra; Morris v. State, (1979) Ind.App., 397 N.E.2d 1056 ; Johnson v. State, (1979) Ind.App., 384 N.E.2d 1035 . It cannot be doubted that, as the Supreme Court of Ohio reasoned in State v. Holt, supra, extensive and impressive credentials such as those revealed here, together with testimony of a *137 sophisticated technological nature far beyond the average layman’s understanding, may impress a jury and influence inordinately the weight which is attached to the expert’s testimony.
The erroneous admission of Agent Riley’s testimony constitutes reversible error. The judgment of the trial court should be reversed and the cause remanded for a new trial. I dissent.