In this appeal the defendant, Calvin 0. Atkins, challenges his second degree murder conviction on two principal grounds. First, he challenges the validity of the practice relating to the employment of a private prosecutor to assist the public prosecutor. Second, he alleges the trial court erred in permitting the State to impeach the defendant by questioning him on cross-examination as to his two prior criminal convictions, a practice condemned by this Court in State v. McAboy, _W. Va. _, 236 S.E.2d 431 (1977).
I
PRIVATE PROSECUTOR
Defendant urges us to declare that the practice of employing a private prosecutor to assist the public prosecutor is an anachronism and that it should be abolished on the ground that it essentially leads to overzealous and unfair criminal prosecution. He also contends that in the present case, the private prosecutor exceeded normal bounds by virtually trying the entire case. Finally, he asserts error was committed when the trial court *504refused to require the private prosecutor to disclose the amount of fee he was receiving to act as private prosecutor.
The right to hire a private prosecutor appears to evolve from the common law, where under old English practice the Crown did not supply a public prosecutor to handle routine felonies. The victim or his family was therefore required to hire counsel to bring the guilty party into the criminal justice system. 1 J. Bishop, New Criminal Procedure (2d ed. 1913), p. 245; 63 Am. Jur. 2d Prosecuting Attorneys § 9; Note, Private Prosecution— the Entrenched Anomaly, 50 N.C.L. Rev. 1171 (1972).
W. Va. Code, 7-7-8, explicitly recognizes the practice.1 This statute makes a clear distinction between a special prosecutor, who is appointed by the court because of a disqualification or inability to act on the part of a public prosecutor, and the private prosecutor. We discussed at some length the court appointment of a special prosecutor in State ex rel. Goodwin v. Cook, _W. Va. -, 248 S.E.2d 602 (1978).
The role of a private prosecutor has not received extensive treatment by this Court, although in several cases the specific conduct of the private prosecutor has been discussed. See, e.g., State ex rel. Moran v. Ziegler, _W. Va._, 244 S.E.2d 550 (1978); State v. Lohm, 97 W. Va. 652, 125 S.E. 758 (1924); State v. Stafford, 89 W. Va. 301, 109 S.E. 326 (1921). In none of these cases, however, was there a total challenge to the employment of a private prosecutor.
*505In most jurisdictions the courts have generally upheld the private prosecutor system against some specific claim of error without having to deal with the merits of the system as a whole. See, e.g., Powers v. Hauck, 399 F.2d 322 (5th Cir. Tex. 1968); Brooks v. State, 45 Ala. App. 196, 228 So.2d 24 (1969); Thomas v. State, 59 So.2d 517 (Fla. 1952); Territory v. Chong Chak Lai, 19 Hawaii 437 (1909); State v. Bartlett, 105 Me. 212, 74 A. 18 (1909); Commonwealth v. Knapp, 27 Mass. (10 Pick.) 477, 20 Am. Dec. 534 (1830); Goldsby v. State, 240 Miss. 647, 123 So.2d 429 (1960); Baca v. Padilla, 26 N.M. 223, 190 P. 730 (1920); State v. Best, 280 N.C. 413, 186 S.E.2d 1 (1972); State v. Kent, 4 N.D. 577, 62 N.W. 631 (1895); Lopez v. State, 437 S.W.2d 268 (Tex. Crim. 1968); contra, State v. Harrington, 534 S.W.2d 44 (Mo. 1976).
We observe initially that W. Va. Code, 7-7-8, does not affirmatively authorize the system of private prosecutors. The specific statutory language is nothing more than a proviso or exception designed to demonstrate that the Legislature did not intend to abolish the common law rule permitting the employment of a private prosecutor.2
We do not agree with the defendant’s contention that the right to retain a private prosecutor should be abolished. The right to obtain a private prosecutor in this State was never absolute and was always subject to judicial control and review. In State ex rel. Moran v. Ziegler, supra, we clearly established that a private *506prosecutor is subject to the same high standards of conduct in the trial of the case as is a public prosecutor. This rule blunts, if it does not entirely dissipate, the major criticism of private prosecutors, that they will be overzealous to convict and consequently ignore the public prosecutor’s fundamental obligation to do justice.8
There are several positive reasons for retaining the right to employ a private prosecutor. First, we recognize that there may be occasions when the public prosecutor is in need of assistance in order to carry out his duties effectively.4 Second, there may be those occasions when the employment of a private prosecutor would satisfy the public’s concern that a given case is not being given perfunctory treatment. Finally, it is not inappropriate to consider that in certain cases, the victim’s family may wish to satisfy itself that the case is being vigorously prosecuted.
While we recognize that there are legitimate reasons for retaining private prosecutors, we emphasize again that the right is not absolute. The ultimate responsibility for permitting the participation of a private prosecutor rests with the trial judge. A competent public prosecutor should ordinarily not be forced to accept a private prosecutor as an associate. It should also be stressed that our recognition of the role of a private prosecutor does not imply that he should be favored for selection as *507special prosecutor where the regular prosecutor is disqualified under W. Va. Code, 7-7-8.
In regard to the specific errors raised by the defendant, we do not find error in the fact that the private prosecutor did not disclose the amount of his fee. As required by State v. Lohm, 97 W. Va. 652, 125 S.E. 758 (1924), he disclosed who had employed him. He also stated that the amount of his fee would be based upon the time spent in connection with the case, an indication that it was to be calculated on a quantum meruit basis, and thus could not be determined until the trial had been completed. We are not aware, nor have we been cited any case, where the failure to disclose the amount of the private prosecutor’s fee was held to be reversible error.
The defendant also asserts that the private prosecutor dominated the prosecution of the trial and that where the public prosecutor plays only a limited role at trial, error should arise from this fact alone. We decline to adopt such a rule. In State v. Stafford, 89 W. Va. 301, 109 S.E. 326 (1921), a similar argument was advanced and rejected. There, as here, the record revealed that the public prosecutor was present at the trial and assisted in the case, although in Stafford the public prosecutor left the trial when his wife became seriously ill. Here, the public prosecutor was present throughout the trial. It does appear, however, that he did not participate in the examination of witnesses to the same extent as the private prosecutor.
We do not believe it is possible or desirable to establish an arbitrary line as to the degree of participation of the public prosecutor in the trial. The fundamental question is not the extent of the private prosecutor’s participation, but whether his actions result in prejudicial error to the defendant. As previously stated, the private prosecutor is bound to follow the same high standards of conduct in the trial of the case as the public prosecutor, and this should afford the defendant adequate protection.
*508II
McABOY AND THE RULE OF HARMLESS ERROR
The defendant contends and the State concedes that it was error under State v. McAboy,_W. Va.-, 236 S.E.2d 431 (1977), for the trial court to permit the State, over the objection of the defendant, to impeach him on cross-examination by questioning him in connection with his two prior criminal convictions. The State, however, urges that under the totality of the evidence, the error should be deemed harmless. It is important to note initially that McAboy is an evidentiary rule and not a constitutional rule.
Few, if any, trials can be conducted completely free of any error. See, e.g., Lutwak v. United States, 344 U.S. 604, 73 S.Ct. 481, 97 L. Ed. 593 (1953). A perfect trial is the ideal, but it can seldom be attained. The central question is what error can be deemed harmless.
The development of the doctrine of harmless error may be a by-product of the increasing complexity of the rules governing the trial of criminal cases. At least one distinguished jurist views the doctrine as a salutary development to free the common law from its rigid rule that any error, no matter how technical, would cause automatic reversal on appeal, irrespective of the error’s impact on the jury verdict.5
Another commentator has concluded that while cases involving the doctrine of harmless error appear chaotic in their formulation of uniform standards, a unifying principle can be detected: The standard of harmless error must bear a close relationship to the burden of evi-dentiary proof required in the trial. Thus, a distinction can be made between standards for harmless error in a civil case, where the degree of proof is by a preponder-*509anee of the evidence, and in a criminal case, where proof beyond a reasonable doubt is mandated.6
Moreover, in a criminal case a further distinction can be and has been made between constitutional and non-constitutional error. It is generally thought that a constitutional error, because it involves a more fundamental right, carries greater potential for harm than does a nonconstitutional error.
The United States Supreme Court recognized this distinction in Chapman v. California, 386 U.S. 18, 17 L. Ed. 2d 705, 87 S.Ct. 824 (1967), and Cooper v. California, 386 U.S. 58, 17 L. Ed. 2d 730, 87 S.Ct. 788 (1967). In Chapman, the Court set the harmless error standard for a federal constitutional error in holding that “the court must be able to declare a belief that it was harmless beyond a reasonable doubt.” 386 U.S. at 24, 17 L. Ed. 2d at 711, 87 S.Ct. at 828. In Cooper, it stated that where no federal constitutional error was found, “the State is free, without review by us, to apply its own state harmless-error rule to such errors of state law.” 386 U.S. at 62, 17 L. Ed. 2d at 734, 87 S.Ct. at 791.
The harmlessness of nonconstitutional error in the federal courts initially arose from former 28 U.S.C. § 391, which required the federal appellate courts to “give judgment after an examination of the entire record before the court, without regard to technical errors, defects, or exceptions which do not affect the substantial rights of the parties.” In Kotteakos v. United States, 328 U.S. 750, 90 L. Ed. 1557, 66 S.Ct. 1239 (1946), the Court elaborated on the meaning of this statute by establishing for criminal cases a nonconstitutional harmless error test:
“If, when all is said and done, the conviction is sure that the error did not influence the jury, or had but very slight effect, the verdict and the judgment should stand, except perhaps where *510the departure is from a constitutional norm or a specific command of Congress. [Citations omitted] But if one cannot say, with fair assurance, after pondering all that happened without stripping the erroneous action from the whole, that the judgment was not substantially swayed by the error, it is impossible to conclude that substantial rights were not affected. The inquiry cannot be merely whether there was enough to support the result, apart from the phase affected by the error. It is rather, even so, whether the error itself had substantial influence. If so, or if one is left in grave doubt, the conviction cannot stand.” [328 U.S. at 764-65, 90 L. Ed. at 1566-67, 66 S.Ct. at 1248]
Our own doctrine of harmless error in criminal cases has not been static. Our earliest decisions followed the rigid common law rule that prejudice is presumed from any error. State v. Cremeans, 62 W. Va. 134, 57 S.E. 405 (1907); State v. Hull, 45 W. Va. 767, 32 S.E. 240 (1899); Crookham v. State, 5 W. Va. 510 (1871).
Exemplifying the rigidity of this rule is State v. Moore, 57 W. Va. 146, 49 S.E. 1015 (1905), where the Court reversed jury convictions of first and second degree murder for two jointly indicted defendants because the record revealed that at the beginning of the trial, when they were formally arraigned to answer the indictment, their pleas of not guilty were made by their attorneys and not by the defendants themselves.7
In our more recent cases, where a nonconstitutional error has been asserted, we have adopted the rather general rule that the case will not be reversed unless the error is prejudicial to the defendant. State v. Riley, 151 W. Va. 364, 151 S.E.2d 308 (1966); State v. Flint, 142 *511W. Va. 509, 96 S.E.2d 677 (1957); State v. Simon, 132 W. Va. 322, 52 S.E.2d 725 (1949).8
Notwithstanding the obvious incompatability between the common law test and our more recent test, there has been no formal recognition of this conflict.
From an overall standpoint, there are two broad categories of trial error. The first category consists of errors relating to the receipt or rejection of evidence, including the ultimate sufficiency of the evidence, and the second category is composed of all nonevidentiary errors.
The realm of evidentiary error in a criminal case consists of two general subcategories. The first is that evidence introduced by the State which is claimed inadmissible, and the second, that evidence offered by the defendant but rejected at trial.9
With regard to the first subcategory of evidentiary error, two questions must be addressed in order to determine whether the error is harmless. First, the inadmissible evidence must be removed from the State’s case and a determination made as to whether the remaining evidence is sufficient to convince impartial minds of the defendant’s guilt beyond a reasonable doubt.10 If the *512remaining evidence is found to be insufficient, the error is not harmless. Should it be determined that the remaining evidence is sufficient to support the conviction, an analysis must then be made to determine whether the error had any prejudicial effect on the jury.
For all other kinds of error, the test for harmless error involves only the single step of weighing the prejudicial impact of the error on the jury verdict. The reason for this is that it is only when the State has placed inadmissible evidence in the case that there is an initial need to determine the evidentiary sufficiency of the State’s case without the inadmissible evidence. The other class of evidentiary error, defendant’s evidence rejected by the trial court, has no direct bearing on the sufficiency of the State’s evidence.
The same reasoning applies to the second category— nonevidentiary errors. There is ordinarily no necessity to weigh the evidence to determine if it is sufficient to meet the Starkey standard, since the error, viz., improper jury selection or prosecutorial misconduct, does not relate to an evidentiary point. The focus is simply the impact of the error on the jury verdict.
We turn now to the application of these harmless error rules to the present case. The admission into evidence of the defendant’s two prior criminal convictions occurred during the State’s cross-examination of the defendant. The initial inquiry is whether the State’s case, absent this testimony, proved the crime such that impartial minds would be convinced of the defendant’s guilt beyond a reasonable doubt.
As is the case with most trials, there was not unanimity among the witnesses as to what occurred. It is undis*513puted, however, that the pathologist found that the victim’s death was caused by five bullet wounds in the back.
A number of the State’s witnesses testified that the victim, Mr. Joseph Shaffer, and the defendant together left a Clarksburg bar at approximately 9:00 p.m. on the evening of November 3, 1976. Shaffer had gone to the bar with a companion, a Mr. Wolfe, for the express purpose of finding the defendant. There was a history of prior business and gambling relations between Shaffer and Atkins.
According to the State’s witnesses, Shaffer and Atkins argued outside the bar but there was no physical altercation. They did not see Shaffer with a gun, but did see Atkins with a handgun. They observed Shaffer turn his back to the defendant and start to cross the street, at which time the defendant fired the first shot, and Shaffer staggered forward and fell in the street. According to the State’s witnesses, the defendant then shot Shaffer, who was lying motionless and face down, several times in the back, left the gun lying next to the victim, and fled the scene.
The defendant, along with several witnesses, testified that Shaffer was carrying a gun in his belt at the time he entered the bar. Both the defendant and his companion, a Mr. Moran, stated that Shaffer and Wolfe seized each of them by the arm and pulled them outside the bar. Shaffer engaged the defendant in an argument over money and, according to the defendant, Shaffer pulled out his gun, the defendant grabbed for it, and in the ensuing scuffle the gun went off.11 The defendant was vague about the actual events of the shooting and unable to remember where Shaffer was struck by the bullets. Moran, other than seeing the scuffle and hearing the shots, was also unable to identify where Shaffer was *514struck. The defendant also had several witnesses who testified as to Shaffer’s violent nature.
The State’s case was not circumstantial, but consisted of eyewitness testimony. None of the State’s eyewitnesses saw Shaffer draw a gun on the defendant nor engage in any physical violence with him prior to the shooting. All of the State’s witnesses saw the defendant shoot several times at close range at Shaffer’s back. The pathologist testified that Shaffer had been shot five times in the back, and that no entry wounds were found in the front of his body. Excluding the erroneous prior conviction evidence, we are of the view that impartial minds would be convinced that the State did prove its case beyond a reasonable doubt under State v. Starkey, supra.
The second step of the harmless error test is to analyze the impact of the error on the jury verdict. The erroneously admitted prior convictions were a conviction for contributing to the delinquency of a minor, in which there was no elaboration of the nature of the underlying charge, and a manslaughter conviction occurring some eighteen years prior to the trial. The trial court instructed the jury that the prior criminal convictions were admitted “for the sole purpose of affecting the credibility of the defendant” and were not “proof that he is guilty of the offense here charged.”
While our cases state the test as whether the error had a prejudicial impact on the jury, the difficulty is in applying the test. The more tangential the error to the ultimate issue of guilt, the less likely its prejudicial impact.
In any inquiry into the prejudicial impact of the error, we will be guided by whether the record reveals that the error was repeated or singled out for special emphasis in the State’s argument. We will scrutinize the record to determine if the error became the subject of a special instruction to the jury, or produced question from the jury. Also of importance is the overall quality of the *515State’s proof. While on appellate review under Starkey, the State is entitled to have its case viewed in the most favorable light, this is because the jury’s verdict of guilty is taken to have resolved factual conflicts in favor of the State in recognition of the jury’s role in evaluating the credibility of the witnesses. If the case contains a number of substantial key factual conflicts or is basically a circumstantial evidence case, or is one that is largely dependent on the testimony of a co-participant for conviction, there is an increased probability that the error will be deemed prejudicial.
Further, if the error is related to critical testimony of the defendant, the more likely that it will be deemed prejudicial. Cf. People v. Mayrant, 43 N.Y.2d 236, 401 N.Y.S.2d 165, 372 N.E.2d 1 (1977); People v. Dickman, 42 N.Y.2d 294, 397 N.Y.S.2d 754, 366 N.E.2d 843 (1977) (holding prior conviction evidence not harmless in context of New York balancing test for admissibility of prior convictions). A further consideration is the cumulative effect of the error in the context of the entire trial. We have recognized that there may be error which, standing alone, is not sufficient to require a reversal, but that when cumulated with other marginal error, the combined effect may be sufficient to warrant a reversal. Cf. e.g., State v. Wilson, _ W. Va. _, 202 S.E.2d 828 (1974) (constitutional and nonconstitutional errors combined).
In the present case, we do not find any of the foregoing factors. The prior convictions were not only stale, but were admitted under a limiting instruction by the court that they could not be considered with respect to the issue of guilt. There was no attempt by the prosecutor to place special or repeated emphasis on the issue. No mention of it was made in the opening statements, and while the closing arguments were not included in the record, defendant does not assert that the State used this issue in its closing argument.
We have discussed the fact that the State’s case was built on direct eyewitness account, and not on circum*516stantial evidence or the sole testimony of a co-participant in the crime. The key factual elements were not substantially disputed by the defendant, and the error did not relate directly to his defense of self-defense. While the prior convictions were directed at impeaching the defendant, their effect must be gauged in the light of the defendant having had other witnesses to corroborate his testimony. Other courts have held that the harmfulness of a prior conviction must be evaluated in light of how evenly the evidence is balanced between the prosecution and the defendant. Compare, e.g., United States ex rel. Lowry v. Myers, 364 F.2d 297 (3d Cir. 1966), and Commonwealth v. King, 316 A.2d 878 (Pa. 1974) (error held not harmless), with United States v. Coades, 549 F.2d 1303 (9th Cir. 1977) and People v. Morrison, 67 Cal. App. 3d 425, 136 Cal. Rptr. 650 (1977) (error held harmless).
In viewing the verdict of second degree murder against the backdrop of all the evidence, we cannot conclude that this error had any prejudicial impact on the verdict. We therefore conclude the error was harmless. It should be clearly understood that while the McAboy violation in the present case has been found to be harmless error, this should not be construed to mean that we are retreating from the application of the McAboy principle in other cases. A prosecutor acts at his peril in violating McAboy, and the case will be reversed unless a clear showing can be made that it meets the stringent harmless error standards set out in this opinion.
Finally, the defendant asserts that the trial court erred in permitting one of his witnesses, Roger Gains, to be subjected to cross-examination by the prosecutor through the use of “comparative testimony.” Over the defendant’s objection, Gains was asked several questions concerning statements by two State witnesses which were in direct conflict with the facts that Gains had stated on direct examination. In each instance, after summarizing the particular fact statement made by the State witness, the prosecutor asked Mr. Gains if “you *517say he is not telling the truth about that?” After an objection by the defense counsel Afras overruled, Mr. Gains then ansAvered, stating: “That is correct.” Defendant urges that this should constitute reversible errror since the same practice Avas condemned in State v. Ramey, _W. Va._, 212 S.E.2d 737 (1975), overruled on other grounds, State v. McAboy, _ W. Va. _, 236 S.E.2d 431 (1977):
“This practice is to be condemned because, Avith his other burdens, it forces a defendant to characterize the testimony of other Avitnesses. In responding, the defendant must either admit the truth of the testimony of such Avitnesses given against him or, in effect, accuse such Avitnesses of perjury. Besides placing an unAvarranted burden upon the defendant, the tactic invade the province of the jury. It is for that body to resolve the ultimate facts and to give credence and Aveight to the testimony of Avitnesses, on proper instructions of the court.” [_W. Va. at-, 212 S.E.2d at 745]
The State contends that Ramey is not applicable because it dealt Avith cross-examination of the defendant, not one of his Avitnesses, and furthermore, Avhile the practice Avas condemned, it Avas not held to be reversible error. Ramey relied on the Virginia case of Plymale v. Commonwealth, 195 Va. 582, 79 S.E.2d 610 (1954), overruled on other grounds, Wooden v. Commonwealth, 208 Va. 629, 159 S.E.2d 623 (1968), Avhich involved a purported oral confession made by the defendant to the sheriff and fire warden, which the defendant denied having given. On cross-examination, the state solicitor asked the defendant if he were accusing the witnesses of falsifying their testimony and committing perjury. The court theorized that a great deal of the harm was the attempt to elicit an accusation of perjury against the sheriff, who was charged with the custody of the jury trying the defendant and who was in the courtroom throughout the trial.
*518We do not conceive that Ramey intended to prohibit any cross-examination of a witness in regard to inconsistencies in his testimony, and specifically, to prohibit cross-examination which requires the witness, in light of the contrary testimony of another, to reconsider his own testimony to determine if he himself were mistaken. Rather, Ramey is directed against compelling the witness to give an opinion as to the truth or falsity of another witness’ testimony, which is an ultimate fact to be determined by the jury. But, even if what Ramey condemns does occur, the question remains as to whether the occurrence is automatically reversible error. We think not.
A number of cases have considered this question, but none has reversed on this point alone. In State v. Hariott, 210 S.C. 290, 42 S.E.2d 385 (1947), the court indicated that it was error, but was more concerned with the sufficiency of the evidence proving the defendant’s guilt. Later South Carolina cases, while condemning the practice, have not deemed it reversible error. State v. Avery, 255 S.C. 570, 180 S.E.2d 190 (1971); State v. Outen, 237 S.C. 514, 118 S.E.2d 175 (1961). In People v. King, 240 Cal. App. 2d 389, 49 Cal. Rptr. 562 (1966), cert. denied, 385 U.S. 923, 17 L. Ed. 2d 146, 87 S.Ct. 235, the court stated this type of questioning to be improper, but concluded that it was harmless since the court had sustained an objection to the question.
The Louisiana Supreme Court is somewhat ambiguous on the point. In State v. Bradford, 298 So.2d 781 (La. 1974), it sanctioned the technique of cross-examining a witness as a contradiction between his testimony and that of a prior witness. The court acknowledged in State v. Shaffer, 260 La. 605, 257 So.2d 121 (1971), on the other hand, that it was improper to cross-examine with the question, “Well, if they testified to the contrary, who would be telling the truth?”, but the case was not reversed on this point.
In Illinois, the intermediate appellate courts appear to recognize that cross-examination on contradiction in tes*519timony can be undertaken, and while the witness can be asked if the other witness was “wrong,” “mistaken” or “incorrect,” it is improper to ask if the other witness’ testimony is “true” or “false.” Compare People v. Morse, 33 Ill. App. 3d 384, 342 N.E.2d 307 (1975), cert. denied, 426 U.S. 953, 49 L. Ed. 2d 1191, 96 S.Ct. 3179 (1976), with Ryan v. Monson, 33 Ill. App. 2d 406, 179 N.E.2d 449 (1961). At least one leading authority appears to sanction the general practice of exposing inconsistencies by asking the second witness to characterize the accuracy of the first witness’ account. C. McCormick, Law of Evidence § 47 (2d ed. 1972) at 100.12
We conclude that it is not improper on cross-examination to direct a witness to specific previous testimony of another witness and ask the witness whether he agrees or disagrees with such testimony. We also conclude that it is objectionable on cross-examination to require a witness to state whether another witness’ testimony is true or false, since this is the ultimate question that a jury must decide. We do, however, state that the failure to sustain an objection to such improper questioning will not necessarily result in error unless the technique has been used so pervasively and abusively in the cross-examination as to substantially distort the witness’ testimony on critical trial issues. Here, we do not find that the comparative testimony technique reached this level, and therefore decline to hold that it was error.
For the reasons stated above, we affirm the judgment of the Circuit Court of Harrison County.
Affirmed.
W. Va. Code, 7-7-8, in pertinent part reads:
“If, in any case, the prosecuting attorney and his assistants are unable to act, or if in the opinion of the court it would be improper for him or his assistants to act, the court shall appoint some competent practicing attorney to act in that case. ... No provision of this section shall be construed to prohibit the employment by any person of a practicing attorney to assist in the prosecution of any person or corporation charged with a crime.” [Emphasis supplied]
The specific provision of W. Va. Code, 7-7-8, relating to private prosecutors, is:
“No provision of this section shall be construed to prohibit the employment by any person of a practicing attorney to assist in the prosecution of any person or corporation charged with a crime.”
We recognize that there is a distinction between a “proviso” and an “exception.” See, e.g., Eaton v. County Court, 140 W. Va. 498, 85 S.E.2d 648 (1955); State v. Cunningham, 90 W. Va. 806, 111 S.E. 835 (1922). For the present purpose, it is clear that no matter which term is applied to the language of W. Va. Code, 7-7-8, that section does not operate to authorize the assistance of a private prosecutor.
While not at issue in the present case, most courts have held that a private prosecutor has no right to initiate indictments before the grand jury or to appear before it. Nichols v. State, 17 Ga. App. 593, 87 S.E. 817 (1916); Wilson v. State, 70 Miss. 595, 13 So. 225 (1893); Flege v. State, 93 Neb. 610, 142 N.W. 276 (1913); Hartgraves v. State, 5 Okla. Crim. 266, 114 P. 343 (1911). By analogy our case of State v. Frazier, _ W. Va. _, 253 S.E.2d 538 (1979), would suggest this result.
In Syllabus Point 2 of State v. Pratt,_W. Va._, 244 S.E.2d 227 (1978), we counseled against “an oppressive overmatch between prosecutors and defense counsel.” ... While Pratt involved the defense standpoint, it is clear that there may be occasions where the situation is reversed with a newly elected, inexperienced prosecutor.
R. Traynor, The Riddle of Harmless Error (1970), pp. 3-17.
See Saltzburg, The Harm of Harmless Error, 59 Va. L. Rev. 988-89 (1973).
We recently overruled this holding in State v. Grimmer, _ W. Va._, 251 S.E.2d 780, 786 (1979), where we concluded that the error was harmless where the entire case had been vigorously tried on the basis of a not guilty plea, even though the record failed to show any plea of not guilty by either the defendant or his counsel.
As to a constitutional level error, we have followed the rule of Chapman v. California, 386 U.S. 18, 17 L. Ed. 2d 706, 87 S.Ct. 824 (1967), that the error must be shown to be harmless beyond a reasonable doubt. State v. Boyd, _W. Va. _, 233 S.E.2d 710 (1977).
Theoretically there is a third and a fourth subcategory — evidence offered by the State but rejected at trial, and evidence offered by the defendant and admitted over the State’s objection. These two categories are ordinarily not encountered on appeal, since the State has no right to appeal an adverse verdict.
This standard is, of course, our normal standard for appellate review of the sufficiency of the evidence in a criminal trial, as more fully set out in Syllabus Point 1 of State v. Starkey, _W. Va. _, 244 S.E.2d 219 (1978). In State v. Frazier, _W. Va. _, 253 S.E.2d 538 (1979), we recognized that Starkey’s test for the sufficiency of the evidence on appellate review in a criminal case was similar to the federal rule stated in Burks v. United States, 437 U.S. 1, 57 L. Ed. 2d 1, 98 S.Ct. 2141 (1978). The same test *512has recently been made applicable to federal habeas corpus review of the sufficiency of the state’s case. Jackson v. Virginia,_U.S. _,_L. Ed. 2d_,_S.Ct_, 47 U.S.L.W. 4883 (June 28, 1979). Our formulation of this first step of the harmless error analysis is similar to that found in Note 18 of Kotteakos v. United States, 328 U.S. 750, 90 L. Ed. 1557, 66 S.Ct. 1239 (1946).
The State’s ballistic testimony was that all of the bullets in the deceased were from a 45-caliber gun, which was the type of gun carried by Atkins.
The pertinent statement from McCormick is:
“Of course, the contradicting witness may simply state the facts as he asserts them, without references to the prior testimony which is being contradicted. It seems, however, that where appropriate the contradiction may be more direct. Thus it would seem acceptable to recite in the question the pertinent part of the prior testimony of the first witness, and inquire, ‘What do you say as to the correctness of this statement?’ ” [Footnote omitted]