dissenting.
I disagree with the majority’s conclusion that the testimony about defendant’s behavior towards the two waitresses at work was not so prejudicial as to require a new trial. The majority assumes that it was error to admit the testimony of defendant’s employer because the race of the two waitresses was irrelevant to rebut otherwise proper testimony and then concludes that the admission of this testimony was harmless error. I agree with the Court of Appeals’ conclusion that “the ages and the race of the waitresses was not relevant to rebutting the defendant’s evidence that he was a good employee.” State v. Cotton, 99 N.C. App. 615, 394 S.E.2d 456 (1990) (emphases added). However, I agree with Judge Johnson’s dissent in the Court of Appeals that this error was not harmless because the evidence of defendant’s guilt in this case is not so overwhelming as to remove any reasonable possibility that if this testimony had not come in, a different result would have been reached at trial.
A review of the evidence in this case is crucial to a resolution of this issue since we must determine if the error was prejudicial. The majority opinion sets out the relevant testimony given by defendant’s employer concerning defendant’s behavior as exhibited towards two specific waitresses while he was working at the restaurant. That testimony reveals the employer testified that while defendant “messed” with all the waitresses, “it was two more than anybody else.” “Messing” with the waitresses consisted of defendant’s touching these two waitresses on various parts of their bodies, telling them dirty jokes, and talking about sex. The two waitresses who were the targets of defendant’s actions were eighteen and forty-seven years old, and both were white. The victims in the present case were ages twenty-two and forty-one, and both were white.
*770The evidence concerning defendant’s identity is less than overwhelming. As Judge Johnson points out in his dissent:
First, as to victim one, the evidence also tended to show that she was nearsighted and was not wearing her glasses during the attack upon her, and the only illumination in the room was from a street lamp filtering through her blinds; that during the time her assailant was in her presence he made efforts to keep her from seeing his face; that upon viewing a photographic lineup on 31 July containing six photos, one of which was of defendant, she initially chose two pictures from the array, one of which depicted defendant. After examining those two pictures for a number of minutes, she told the investigating officer that defendant’s photo “looks most like him.” On 8 August, she viewed a physical lineup consisting of seven men. Defendant was the only participant whose picture had been among those in the photographic array. Again, the victim was instructed to choose the one that looked the most like her assailant. After viewing the participants for a while, she told the officer that it was between participants numbers four and five. She then stated that number five, defendant, “looks the most like him.”
State v. Cotton, 99 N.C. App. at 624-25, 394 S.E.2d at 461 (Johnson, J., dissenting). Furthermore, the physical evidence gathered at this victim’s home, semen stains on her bed sheets, was inconsistent with defendant’s blood type.
As to the second victim, the evidence also tended to show that on the two occasions that the assailant entered her house he directed the beam of a flashlight in her face; that other than the flashlight beam the only source of light in the house was from a television set which was not on when her attacker entered the second time. On 31 July, the second victim viewed the same photographic lineup of six photos, including defendant’s photo, that the first victim had viewed. Likewise, she was told to pick out the photo of the individual who most resembled her assailant. She failed to pick out anyone from this array. When she viewed the physical lineup on 8 August 1984, and picked out a Kenneth Watkins as her attacker, she thereafter asked the officer conducting the lineup if she had picked out the right person. On cross-examination she stated that she tried to pick out the right man, but had made a mistake.
*771State v. Cotton, 99 N.C. App. at 624, 394 S.E.2d at 461 (Johnson, J., dissenting). In addition, the second victim was present at the first trial but did not indicate until shortly before the second trial, some three and one-half years after the incident, that she could identify defendant as her assailant.
As with the first victim, the physical evidence gathered at the scene of the second incident was inconsistent with defendant’s blood type. At the home of the second victim, the officers recovered a pair of the victim’s panties which had semen stains, and these stains were inconsistent with defendant’s blood type. The evidence at the second victim’s home indicated that the assailant had broken an outside light on the porch when breaking into the home, and the police gathered a sample of fresh blood found on the storm door at the victim’s home. This blood sample was likewise inconsistent with defendant’s blood type. Thus, there was no evidence gained from the laboratory testing of the samples collected at the scene of either incident linking defendant with these two attacks.
Under the circumstances, I conclude that there is a reasonable possibility that, had the error not occurred, a different result would have been reached at trial. N.C.G.S. § 15A-1443(a) (1988). Thus, I find the error prejudicial.
Chief Justice EXUM joins in this dissenting opinion.