The opinion of the Court was delivered by
Weintraub, C. J.George Merritt, Jr., and Gail Madden were convicted of murder in the first degree and sentenced to life imprisonment. The Appellate Division reversed the convictions and ordered a new trial. We granted the State’s petition for certification. 59 N. J. 287 (1971).
The victim was a police officer, John Y. Gleason. He was stationed at the perimeter of an area in Plainfield, N. J., which had just experienced serious racial rioting. The officer sought to arrest one Bobby Lee Williams, who allegedly had attempted to strike some boys with a hammer. As the officer sought to make the arrest, spectators threw sundry missiles at him. In response to some movement by Williams, who still had the hammer, the officer fired his .weapon, wounding Williams. Thereupon a mob came at the officer. The officer tried to escape, but after a pursuit involving a considerable distance, he was brought down. Members of the mob inflicted a savage beating which continued until he died.
Twelve persons were indicted for the murder. Eleven were tried together. One was acquitted by order of the trial court; seven were found not guilty by the jury; the two respondents were convicted; and the jury was deadlocked with respect to the remaining defendant.
As will hereinafter be spelled out, the evidence was sufficient to warrant a finding of first degree murder against *383both defendants. As to Merritt, there was testimony that he struck the deceased with a meat cleaver, and as to Madden, a woman weighing about 300 pounds, there was testimony from which it could be found she jumped upon and stomped the victim as he lay prostrate on the ground. A finding of such participation in the brutal attack on the deceased would warrant a conviction of murder in the first degree without regard to whether the blows either one inflicted in fact caused death.
But the case went to the jury, not only on the claim of such physical participation in the attack, but also on the premise that each defendant would be equally guilty if the participation consisted of words or conduct encouraging others to inflict the beating. Thus in defining aiding or abetting, the trial court included “all assistance rendered by words, acts, encouragement, support, or presence, actual or constructive, to render assistance, if necessary.” It is the charge on aiding or abetting which raises difficult questions in the ease, and since we cannot know whether the jury accepted as true the testimony of physical participation referred to above, we cannot say the verdict did not rest upon a finding of such other activity as would come within the aiding or abetting instruction. The charge upon aiding or abetting included instruction upon the liability of conspirators. It was that aspect of the charge which led the Appellate Division to reverse the judgments and order a new trial.
An inquiry into the validity of the instruction upon aiding or abetting the murder of a police officer acting in the execution of his duty necessarily involves a consideration of the constituent elements of that category of murder. We will deal first with that subject, and then with the question whether the aiding or abetting charge had the infirmity the Appellate Division found or some other or further infirmity.
*384I
At common law, murder was a killing with "malice aforethought.” It is difficult today to know precisely what the term meant throughout the development of the common law. There is authority that malice aforethought, as late as the sixteenth century, meant “a deliberate, premeditated intent to kill formed some time beforehand and that no killing 'on a sudden’ even without provocation or on slight provocation was considered murder,” a killing "on a sudden” being manslaughter. State v. King, 37 N. J. 285, 299-300 (1962). As King pointed out, the concept of murder changed. Apparently an intent to kill, without a prior design to kill, was later sufficient for "malice aforethought.” 2 Burdick, Law of Crime (1946) §§ 448c-450, pp. 164-169.
There were other situations at common law in which such actual malice was not required to support a charge of murder. One category was a killing in the course of the commission of a felony, as to which the intent to commit the felony sufficed even though there was no intent to kill. Another category, and the one with which we are here concerned, was described by Coke in his Institutes as the killing of a magistrate, sheriff, constable or other officer executing a lawful warrant or in the performance of any other official duty. 2 Burdick, Law of Crime (1946) § 451, p. 170, and § 453, pp. 171-73. A killing in those circumstances was, without more, deemed murder, and this because of the need for further protection of such officers when acting in obedience to their official duty and because an attack upon officers thus engaged is an attack upon the institution of justice itself and therefore an attack upon the security of all citizens. Brown v. State, 62 N. J. L. 666, 697-698 (E. & A. 1899); Bullock v. State, 65 N. J. L. 557, 570 (E. & A. 1900).
A like provision as to the killing of officers has appeared in our homicide statutes at least since 1829 (Act of Feb. 17, 1829, § 66, p. 128; Rev. of 1846, § 3, p. 258; L. 1898, c. *385235, § 106, p. 824). Our present statute, N. J. S. A. 2A:113-1, defines murder, as follows:
If any person, in committing or attempting to commit arson, burglary, kidnapping, rape, robbery, sodomy or any unlawful act against the peace of this state, of which the probable consequences may be bloodshed, kills another, or if the death of anyone ensues from the committing or attempting to commit any such crime or act; or if any person kills a judge, magistrate, sheriff, constable or other officer of justice, either civil or criminal, of this state, or a marshal or other officer of justice, either civil or criminal, of the United States, in the execution of his office or duty, or kills any of his assistants, whether specially called to his aid or not, endeavoring to preserve the peace or apprehend a criminal, knowing the authority of such assistant, or kills a private person endeavoring to suppress an affray, or to apprehend a criminal, knowing the intention with which such private person interposes, then such person so killing is guilty of murder.
It will be noted that the statute reflects the three categories of common law murder to which we have already referred. As to killings which are not within the category of a killing in the commission of the specified common law felonies and which are not in the category of the killing of an officer, our statute requires proof of an intent to kill or intent to do grievous bodily harm. State v. Gardner, 51 N. J. 444, 456-459 (1968). If the intent is to inflict injury less than serious bodily harm, the offense is involuntary manslaughter.
As to killings in the commission of the specified common law felonies, it is enough for present purposes to say that there need not be an intent to inflict any injury at all. The question is whether an intent to injure is required in the third category — the killing of an officer in the execution of his office or duty. Here the law is quite obscure. It is clear that there need not be an intent to kill. But the authorities do not call for an intent to do harm of a grievous character, although usually that intent is evident from the nature of the weapon or force applied. That surely was the situation in our cases involving the killing of officers. Brown v. State, supra, 62 N. J. L. 666; Bullock v. State, *386supra, 65 N. J. L. 557; State v. Metalski, 116 N. J. L. 543 (E. & A. 1936); State v. Genese, 102 N. J. L. 134 (E. & A. 1925); State v. Sage, 99 N. J. L. 229 (E. & A. 1923); State v. Schilling, 95 N. J. L. 145 (E. & A. 1920). We are satisfied that the total context requires a holding that intent to inflict any injury upon such an officer is all that is required, so that if any force applied with intent to injure does in fact result in death, the offense is murder.
But we cannot say the Legislature intended to make the offense murder if there was no intent to do some physical injury. If the unlawful act, unaccompanied by an intent to do some injury to the officer, does result in his death, the crime would be involuntary manslaughter. See State v. Weisengoff, 85 W. Va. 271, 101 S. E. 450 (Sup. Ct. 1919); Dickby, “Homicides in Resisting Arrest,” 18 Corn. L. Q. 373, 377-78 (1933).
We come now to the truly ambiguous event in the statutory history in our State. Until 1965, the killing of an officer in the execution of his duty was murder in the second degree and that had been true since degrees of murder were first introduced by L. 1839, pp. 147-48. Until 1965 murder was in the first degree only if “perpetrated by means of poison, or by lying in wait, or by any other kind of willful, deliberate and premeditated killing” or if committed in perpetration or attempt to perpetrate the specified common law felonies already mentioned. All other murder was second degree. In 1965, by c. 212, § 1, there was added to murder in the first degree, a murder:
* * * which is perpetrated in the course or for the purpose of resisting, avoiding or preventing a lawful arrest, or of effecting or assisting an escape or rescue from legal custody, or murder of a police or other law enforcement officer acting in the execution of his duty or of a person assisting any such officer so acting.
Thus for the first time since the year 1839, a killing of a police officer, if in the execution of his duty, became another category of capital offenses, that is, punishable by death. *387(The killing of the other public officials included in the definition of murder in N. J. S. A. 3A:113-1 remains murder in the second degree.)
The question is whether the 1965 statute should be read to lift into the category of a capital offense every killing of a police officer in the execution of his duty which theretofore was second degree murder punishable by a term of years. More precisely, the question is whether the legislators intended to make punishable by death all of these killings of an officer: (1) where there was an intent to kill; (3) where there was an intent to do grievous bodily harm; (3) where there was an intent to do harm less than grievous bodily harm.
Upon this critical question, there is no history at all within the legislative records. The bill had no statement attached. Nor is there a record of any expression by any member of the Legislature. It is apparent from the discussion hereinabove that the full reach of the murder statute would not have been understood without tracing its antecedents in the common law. The only available legislative history is found in the records of the then Governor who approved the bill. That history persuades us that the amendment must be read to make the murder of a police officer punishable by death only if the killing was intentional.
Prior to the 1965 amendment, as is shown by the foregoing discussion, the fact that the victim was a police officer killed in the execution of his duty had no legal significance in a charge of murder in the first degree. To make such a killing murder in that degree, the same additional facts were required as in the killing of a non-officer. Thus (passing a killing by poison or lying in wait — categories rarely invoked), the State had to prove the murder of a policeman was a “willful, deliberate and premeditated killing,” or was committed in perpetrating or attempting to perpetrate the specified common law felonies. With this introduction, the following facts in the Governor’s file are readily understood.
*388The Prosecutor of Essex County said in a letter to the Governor dated October 14, 1965:
I have had occasion in recent days to examine very carefully our homicide statute as it pertains to the killing of a police officer. That statute, 2A:113-1, provides that the killing of an officer of justice is murder. Murder under our law is presumed to be murder in the second degree. 2A:113-2 provides that felony murders and wilful, deliberate and premeditated murders constitute murder in the first degree.
It is my strong feeling that the intentional killing of a police officer whether in the course of a felony or not; whether premeditated or not; should constitute first degree murder, the punishment for which is either death or life imprisonment.
A bill (S. 366) was introduced and passed, and the Governor in his press release dated December 37, 1965 said: “This amendment was first suggested” by the Prosecutor of Essex County.
On December 30, 1965 a citizen wrote to the Governor stating her concern that the amendment then awaiting his signature might go too far and be unrealistic. In part the citizen said:
It seems possible that Sen. Sarcone’s new bill would cover situations in which a police officer is accidentally killed. Does it not put a grand jury, a prosecutor’s office or a courtroom jury in the position of finding for first degree murder, with no chance or choice to consider any extenuating circumstances there might be? Is it properly worded so that no jury in the future finds itself caught in the throes of an arbitrary law that it cannot in conscience support? Is it just, and will it serve as a deterrent?
The Governor answered, saying:
* * * I assure you that extremely careful consideration was given this measure, and I am satisfied that the law does not cover the purely accidental slaying of a police officer.
It is clear that the county prosecutor proposed no more than that the offense be capital if there was an intent to kill. He wanted to relieve the State of the necessity of proving *389anything in addition to an intent to kill to raise the murder from second to first degree. The correspondence between the Governor and the citizen is somewhat equivocal. The citizen referred to an “accidental” killing. “Accidental” signifies without design, and applied to a homicide, the words can refer either to the means or to the result the means bring about. If an act that is intended in fact leads to a death that is not intended, the act is not accidental but the result is. Thus ambiguity inheres in “accidentally killed” and “accidental slaying.” But since the subject is capital crime, it is surely arguable that the citizen was speaking of an unintended death rather than of an inadvertent act, and that the Governor responded on the same predicate.
In any event, it may be doubted the several legislators were familiar with the gloss upon the homicide statute which one might find upon extended study, and that, being thus informed, they intended to resurrect a definition of a capital crime abandoned in the year 1839. The purpose the county prosecutor expressed in seeking the amendment, i. e., to make the intentional killing of a police officer a capital crime, was compatible with contemporary thought. There was nothing before the legislators to alert them to a larger objective.
We must enforce the legislative will if it is within constitutional limits whether we approve of the legislative intent or not. But it is our initial task to seek that intent, and to that end we must consider any history which may be of aid. New Jersey Pharmaceutical Association v. Furman, 33 N. J. 121, 130 (1960). Surely in so grave a matter as a capital offense, the Court ought be certain that it does not read a statute to exceed the true legislative objective. Here we can be confident only that the Legislature intended to authorize a death sentence if there was an intent to kill.
Hence we read the 1965 amendment to elevate to first degree the murder of a police officer in the execution of his duty if there is an intent to kill. If the intent is to do bodily *390harm, the offense remains murder in the second degree, and this, as we have said, whether the intended harm is grievous or less than grievous.
As the trial court submitted the case to the jury, the degree of murder depended solely upon whether the jury found the officer was killed while acting in the execution of his duty. A first degree conviction was sought only on the hypothesis that the officer was so acting.1
The trial court charged (1) that murder in the second degree required an intent to kill or an intent to do grievous bodily harm, and (2) that the murder is in the first degree if it is also found the officer was in the execution of his duty. That instruction was erroneous as to the defendants in one respect and as to the State in another. It was incorrect to say that the murder was in the first degree if there was an intent to do grievous bodily harm, for an intent to kill was necessary. On the other hand, an intent to do any bodily harm (even though less than grievous) would make the homicide murder in the second degree if the officer was killed in the execution of his duty. On the thesis of the prosecution, the instruction should have conveyed these propositions with respect to murder in the second and first degrees: (1) an unlawful homicide is murder in the second degree (a) if there was an intent to kill or to do grievous bodily harm, or (b) if the victim is an officer in the execution of his duty and there was an intent to inflict any physical injury upon him; and (2) the murder is elevated to first degree if (a) the officer was in the execution of his duty and (b) there was an intent to kill.
II
We turn then to the instructions as to aiding or abetting.
Preliminarily we note the general proposition that parties to the commission of an unlawful homicide may be *391guilty of offenses of different degrees depending upon their respective participation and intent. State v. Fair, 45 N. J. 77, 94—96 (1965); 40 Am. Jur. 2d, Homicide, § 44, p. 334 (1968); annotation, 12 A. L. R. 275, 276 (1921). Accordingly if several participate in an attack and death results, the degree of the offense as to each may turn upon his own acts and purpose. People v. Monaco, 14 N. Y. 2d 43, 248 N. Y. S. 2d 41, 197 N. E. 2d 532 (Ct. App. 1964); People v. May, 9 A. D. 2d 508, 195 N. Y. S. 2d 792 (1st Dept. 1960); Mickey v. Commonwealth, 72 Ky. 593 (1873). Thus if two should attack and one of them intends only a simple assault and battery and is unaware of the intent of the other to use deadly force, he would be culpable only according to his own intent and wrong. But if he participates in an attack or continues in it with an awareness of the purpose of others to kill or to do grievous bodily harm, he is chargeable with that further intent and result.
Thus in the case at hand, a defendant who physically participated in the attack upon the officer, although not intending to kill or to do grievous bodily harm, would be chargeable with such an intent if he or she entered upon the attack or continued the attack with an awareness that others harbored such intent. The jury could have found in this case that each defendant did join in the physical attack with an intent to kill by reason either of the severity of the blows he or she struck or by reason of an awareness that other participants were seeking the death of the officer.
But as we said earlier, the case was given to the jury on the further premise that a defendant could be guilty of murder, second degree or first, on the basis of other conduct or on the basis of words of encouragement to the actual killers. It was one aspect of that instruction which led the Appellate Division to order a new trial.
N. J. S. A. 2A:85-14, which we will refer to hereinafter as the aiding or abetting statute, reads in part:
*392Any person who aids, abets, counsels, commands, induces or procures another to commit a crime is punishable as a principal.
The pervasive character of this statute is evident. We note that our statute embraces conduct which at common law fell within the concept of an accessory before the fact. State v. Wilson, 79 N. J. L. 241 (Sup. Ct. 1910), affirmed, 80 N. J. L. 467 (E. & A. 1910). All who aid or abet or counsel or command or induce or procure another to commit a crime are equally guilty with the perpetrator as principals.
A
In instructing the jury upon the aiding or abetting statute, the trial court said in part:
In legal effect, the State charges that the defendants acted together, that they conspired to kill John Y. Gleason, and that each one aided and abetted the others. The State contends, therefore, that it makes no difference which one inflicted the fatal blow or blows that killed John V. Gleason, and that, if it has been proved beyond a reasonable doubt that the defendants together, in words or by conduct, agreed together to do an unlawful act, in the course of which there was a killing, all are equally responsible.
For the purposes of this case, it is sufficient to say that a criminal conspiracy is an agreement to do an unlawful act. Our statute, as I have told you, provides that: Any person who aids, abets, counsels, commands, induces or procures another to commit a crime is punishable as a principal. Under our law, all those who conspire to commit a crime and participate in some way in its commission are joint principals and each is as guilty as the person who actually commits the crime and is liable to the same punishment.
A conspiracy may be proved by direct evidence, that is, proof that the defendants together had a conversation and agreed in words to do an unlawful act. But it is not necessary that it be proved by such direct evidence. It is not essential to show an actual verbal agreement to commit a crime. Concerted action does not have to be proved by direct evidence of a formal plan to commit a crime, verbally concurred in by all who are charged, nor is it essential that all enter into the conspiratorial agreement at one and the same time. The proof may be circumstantial. Participation and acquiescence can be established or inferred by you from conduct as well as from spoken words. One who joins a conspiracy after its formation is equally guilty with the original conspirators.
*393The Appellate Division held it was error thus to introduce the subject of conspiracy in the instructions to the jury. The Appellate Division pointed out that a conspiracy to commit a crime is an offense distinct from the substantive offense committed pursuant to the agreement. It found “no evidence as to any plan or agreement such as is necessary to sustain a conspiracy prosecution,” and added, apparently, that “an agreement to act in the future” is not itself enough to constitute the participation contemplated by our aiding or abetting statute.
We do not doubt that a conspirator, as such, may be guilty of the ensuing substantive crime under the aiding or abetting statute. Whether entering into a conspiracy will itself lead to such guilt depends upon whether the making of the conspiratorial agreement did in fact aid or abet or counsel or induce or procure another to commit the substantive offense. When the commission of the substantive crime is specifically agreed upon as the object of the conspiracy, it can readily be found that everyone who agreed upon the commission of the crime did thereby in fact aid or abet or counsel or induce or procure the actual perpetrator to commit that crime. See State v. Dolbow, 117 N. J. L. 560, 562-563 (E. & A. 1937); Pinkerton v. United States, 328 U. S. 640, 645-647, 66 S. Ct. 1180, 90 L. Ed. 1489, 1495-1497 (1946); Nye & Nissen v. United States, 336 U. S. 613, 618-620, 69 S. Ct. 766, 93 L. Ed. 919, 925-26 (1949); Model Penal Code (Proposed Official Draft, 1962) § 2.06(3) (a); Tentative Draft No. 1 (1953) § 2.04(3) and comments, pp. 20-24; Proposed New Jersey Penal Code (1971) Vol. I, § 2C:2-6(c) and comment, Vol. II, p. 58.
A conspirator who is thus chargeable with the substantive crime under our aiding or abetting statute is a principal, and on the trial of that charge, the fact of the conspiracy may be proved even though the indictment does not also charge the crime of conspiracy or allege the fact of a conspiracy. State v. Manney, 26 N. J. 362, 370 (1958).
*394Of course, one may be liable as a principal under our aiding or abetting statute even though no conspiracy existed between him and the immediate perpetrator of the substantive crime. And the immediate perpetrator does not become a conspirator with another merely because the other aided or abetted him. The crime of conspiracy requires an actual agreement for the commission of the substantive crime, and unless that agreement in fact existed, each defendant is liable with respect to the substantive offense on the basis of his own conduct with respect to that substantive offense, the perpetrator on the basis of the actual commission of the crime and the other on the basis of any participation in the perpetrator’s purpose and deed as comes within the aiding or abetting statute.
In the case at hand the instruction as to conspiracy was inappropriate. There was no evidence of an actual agreement made in advance of the criminal event, as for example when a number of men agree to embark upon a holdup. In the latter situation, it is appropriate to tell the jury of the common liability of all with respect to the ensuing substantive offense. That was the situation in State v. Cooper, 10 N. J. 532, 568 (1952), and also in State v. Smith, 32 N. J. 501, 521 (1960), cert. denied, 364 U. S. 936, 81 S. Ct. 383, 5 L. Ed. 2d 367 (1961), upon which the trial court apparently relied. But when, as here, there is no proof of an actual prior agreement and each defendant’s liability inevitably depends upon what that defendant did at the time and place of the homicide, the case is unnecessarily encumbered by the introduction of an instruction upon conspiracy.
The question is whether the defendants were harmed by the charge. If the jury understood that a defendant could not be found to be a conspirator except on the basis of the very acts which would otherwise have sufficed to inculpate him or her under the aiding or abetting statute, he or she would not have been harmed by the instruction. Indeed, upon that premise the State’s burden might have been enlarged if the jury gained the impression that a defendant *395could not be convicted unless the defendant were also found to have entered into a conspiracy for the murder of the victim. But a defendant could have been harmed if the jury, not finding that the defendant actually participated in the inflicting of the fatal blows, understood that the evidence would nonetheless warrant a finding of a conspiracy to kill which would then charge the defendant with all the acts of the actual killers. We agree with the Appellate Division that the evidence would not have justified a finding of an actual agreement. The formation of the mob was a spontaneous affair; it would be speculative to say a conspiracy to kill was agreed upon by all who joined in or encouraged the pursuit.
Equally troublesome is the statement in the first paragraph of the quoted instruction, “that, if it has been proved beyond a reasonable doubt that the defendants together, in words or by conduct, agreed together to do an unlawful act, in the course of which there was a killing, all are equally responsible.” That language could be understood to depart from the initial reference to “conspired to kill” and to attribute vicarious responsibility for murder if it were found that defendants agreed to “an unlawful act” and that the killing occurred in the course of that act. Thus, if a defendant were found to have conspired for the pursuit of the victim — an unlawful act — his guilt of murder would be established even if there were no agreement by the defendant that the officer be beaten or killed.
Finally, it appears that the State did not try the case on the theory of conspiracy. The subject was introduced in the trial court’s instructions after the proofs were closed. Defendants were thereby denied an opportunity to testify that they did not intend to enter into an agreement and did not do so.
For these reasons we agree with the Appellate Division that the instruction with respect to conspiracy was erroneous.
*396B
The instruction with respect to aiding or abetting is troublesome in other respects.
In defining murder, the trial court said that the killing had to be malicious and that malice consisted of an intent to kill or to do grievous bodily harm. The trial court added a further mental element which it called “criminal intent” and defined as an “evil meaning mind.” In defining aiding or abetting, the trial court required that there be participation in the “intent” of the actual perpetrators, but the charge was obscure in that regard because the definition of murder included two intents, one in the definition of “malice” and the other in the definition of “criminal intent” as an “evil meaning mind.”
Thus the jury was told that “our courts have held that one is an aider and abettor in the commission of a crime where he or she was an active partner in the intent, which was the crime’s basic element.” Elsewhere the court charged that a defendant’s acts must have been done pursuant to “a criminal purpose held in common between them.” We are dealing here with a mob pursuit of a policeman. Conceivably the intents could be many and different. There could be an intent to kill, or an intent to injure or an intent to prevent the arrest of Williams by no more than a threat of injury, or an intent to harass or humiliate the officer. Each of those intents would be a “criminal intent,” an “evil meaning mind.”
When, as here, a mob is involved in circumstances in which their understanding, intent and purpose may vary, it is difficult, but no less necessary, that the basis of the liability of the members of the mob be defined in the light of those sundry understandings, interests or purposes. If a defendant is to be held, not as a participant in the fatal attack, but as one who by words or acts encouraged others thus to attack, the jury should be plainly told that one cannot be held as an aider or abettor unless it is found that he shared *397the same intent required to be proved against the actual perpetrator.
Part of the difficulty in this case stems from the failure to deal in specifics. If the State sought the conviction of each of the defendants before us solely on the claim that he or she did physically attack the officer, the instruction should have zeroed in upon that claim. But if the State sought a conviction for murder, second or first degree, on the basis of other acts or upon the basis of words, that position should have been stated plainly, to the end that the trial might be conducted on a concrete basis and the matter submitted to the jury with the guidance that such specificity would permit.
In this connection, defendants complain of the following instruction:
Mere presence at the scene or near the scene of a crime does not make one a participant in the crime, nor does the failure of a spectator to interfere make him a participant in the crime. It is, however, a circumstance to be considered with the other evidence in determining whether he or she was present as an aider and abettor, but it is not in itself conclusive evidence of the fact.
The criticism made is that neither presence nor a failure to intervene suggests guilt, and neither fact should be deemed to augment the case against a defendant. See State v. Smith, supra, 32 N. J. at 521; State v. Sullivan, 43 N. J. 209, 236-237 (1964); State v. Fair, supra, 45 N. J. at 95; State v. Mayberry, 52 N. J. 413, 438-439 (1968).
Whether presence has any inculpatory thrust depends upon circumstances. Cf. State v. De Simone, 60 N. J. 319 (1972). Of course, when presence is relevant to the State’s case, it may be proved. In that sense, presence may, with other circumstances, establish guilt. But it does not follow that presence in a public street which is wholly consistent with innocence can itself add inculpatory strength to the State’s case. The instruction on this topic must be phrased in the light of the facts. In the case at hand, neither mere presence at the scene nor a failure to intervene could itself bespeak *398criminal involvement. In these circumstances the instruction should have stopped at the end of the first sentence. We nonetheless are satisfied that in the context of the whole charge, the jury could not have been misled.
III
The final questions before us relate to the subject of manslaughter. The trial court ruled out involuntary manslaughter but left voluntary manslaughter to the jury as a possible verdict.
As already noted, the trial court charged the State had to prove the officer was killed in the execution of his duty as an element of murder. In this connection the jury was told the deceased was not in the execution of his duty if he did not have probable cause to arrest Williams or if he employed excessive force to accomplish an arrest. But the charge barred a verdict of voluntary manslaughter if the jury found the officer was in the execution of his duty when he was killed. Otherwise stated, the trial court ruled that neither a lawful arrest nor the use of necessary force to accomplish the arrest can constitute provocation justifying a finding of manslaughter. This was undoubtedly correct; the protection intended an officer who is thus in the execution of his duty would be nullified if the lawful execution of his duty were held to be an affront competent to reduce the crime to manslaughter. See Brown v. State, supra, 62 N. J. L. at 713-714; Bullock v. State, supra, 65 N. J. L. at 570-571.
The jury was permitted to consider voluntary manslaughter if it found the police officer was not in the execution of his duty. But the instruction was given in abstract terms, without identifying the facts which could legally constitute provocation of these defendants. We do not know what conduct of the officer the trial court had in mind, or whether it intended that circumstances not of the officer’s making could serve legally as provocation. In this connection we note again that it was Bobby Lee Williams who was the *399object of the arrest and who was wounded, and not the defendants. And the officer was not killed while he was making the arrest or using force to that end; the mob had succeeded in preventing consummation of the arrest by driving the officer from the scene, and the killing occurred after a pursuit of about a city block.
The trial court also charged, again without referring to any facts in the case, that “it is not necessary that any specific event be the provocative one”; that “a course of ill treatment or oppression” and “the prospect of its continuance” could legally constitute provocation. That part of the charge undoubtedly was taken from State v. Guido, 40 N. J. 191 (1963). There we dealt with a course of brutality to which the defendant was allegedly subjected by the deceased, her husband. It was in that context that we held “that a course of ill treatment which can induce a homicidal response in a person of ordinary firmness and which the accused reasonably believes is likely to continue, should permit a finding of provocation,” adding that “we merely acknowledge the undoubted capacity of events to accumulate a detonating force, no different from that of a single blow or injury.” 40 N. J. at 211. We are not aware of any facts in the present case which could have warranted that instruction.
Defendants now urge that the jury should have been told (1) that racial or like antagonism or grievances may constitute provocation, and may so serve whether or not the officer was acting in the execution of his duty; and (2) that the illegal arrest or the shooting of Williams could constitute a provocation of bystanders.
The first proposition is frivolous. It is no more than this: that racial, ethnic, religious, economic, political, or like grievances within our social order should be deemed sufficient to provoke a member of one group to kill a member of a group he deems opposed to him, or to kill every police officer as a symbol of the social order within which such grievances exist. The proposition is foreign to the subject of voluntary manslaughter. Voluntary manslaughter is a slaying com*400mitted in a transport of passion or heat of blood induced by a provocation the law deems adequate to arouse such passion in a man of ordinary firmness, and then only if the killing occurs before the passage of time sufficient for an ordinary person in like circumstances to cool off. State v. Guido, supra, 40 N. J. at 209-210; State v. King, supra, 37 N. J. at 300-301. The conventional picture is that of a defendant who exploded in response to some injury or affront to which he was subjected by the deceased. None of this can be found in the proposition defendants now advance. No wrong at all would be attribatable to the deceased; his mere status would be the infuriating fact. It would hardly do to accept the proposition that anyone who assumes public office thereby provokes one group or another to slay him. What defendants really ask is that the Court declare that the intentional killing of a police officer shall constitute only manslaughter notwithstanding the legislative judgment that it is murder.
Nor do we see substance in the second proposition, that an unlawful arrest or the use of excessive force in making an arrest should serve as provocation to bystanders. We are not talking about whether a defendant could defeat a criminal charge on the plea that he acted in defense of another who was being subjected to excessive force. Here defendants do not say the officer was killed to protect Williams from further physical injury. The officer had already been driven away by the mob. Rather defendants say the encounter between the deceased and Williams should permit a finding of provocation with respect to a murder committed after a chase.
Defendants are not related to Williams, and hence we do not have the question whether an illegal arrest or the use by a policeman of excessive force may constitute provocation of a member of the victim’s family. The most defendants can assert is a friendship with Williams. The cases are scant, and in disagreement, upon the question whether provocation can be found in an attack upon a friend. See 1 Wharton, Criminal Law and Procedure (Anderson, 1957) § 282, p. 597; 2 *401Burdick, Law of Crime (1946) § 462h, pp. 195-97. We need not consider that question, because here another dimension is present, for the deceased was a police officer, and although we now proceed hypothetically on the premise that the attempt to arrest Williams and the wounding of Williams were unlawful, nonetheless we cannot approach the issue as if the deceased was a private citizen embroiled in a private encounter.
In State v. Williams, 29 N. J. 27, 36 (1959), we said:
Police officers are not volunteers. They are armed and required to act to enforce the law. They may err in their judgment and exceed their authority in the sense that they misjudge the need for extreme measures or their right to resort to them. Yet, where the purpose is to comply with duty, it would be unreasonable to impose the measure of criminal responsibility applicable to the citizen whose involvement does not originate in a legal compulsion to act and who is free to turn away.
We were there speaking of the criminal liability of a mistaken officer, but the difficulties inherent in his role are relevant here.
It was in that same light that we held that a citizen may not resist an illegal arrest by one he knows to be a police officer. State v. Mulvihill, 57 N. J. 151, 155-156 (1970). The safe and sufficient course is to turn to the courts rather than to engage in a combat in which one or the other will surely be wrong. Nor may a third person interfere with an illegal arrest by one known to be a policeman. And the same policy reasons dictate that a third person may not say an illegal arrest provoked him to kill one he knew to be a policeman.
And as to the wounding of Williams, which for the purpose of the present discussion is assumed to have involved excessive force, we see no basis for a claim of provocation in the undisputed circumstances of this ease. Surely citizens should not be encouraged to sit in judgment of a police officer’s role in a physical combat with another and attack the officer with deadly force upon their estimate of the rightness of the officer’s behavior. It is true that the issue is not whether *402the killing of the officer can be justified or excused, but whether the killing may be found to be manslaughter if a policeman’s use of excessive force did in fact induce in a third person the required transport of passion. We think provocation should not be recognized in those circumstances, for two reasons. The first is that we believe a man of ordinary firmness would not be provoked to a deadly attack. The second reason, which buttresses the first, is that the protection a police officer should have in the public interest might be diluted if his mistaken use of force were accepted as an affront to a bystander.
A claim of provocation could be understandable in a close relative who is on the scene, but if so, we can see no social utility in accepting that claim by other bystanders. Especially is that true in the present case, in which the deadly force was not invoked to protect Williams from further physical injury, but to punish the officer who was trying to escape a menacing mob and was brought down after a substantial pursuit. It may be added that, if a bystander is found guilty of murder in the second degree (i. e., on the hypothesis that the officer was not in the execution of his duty and that the killing was not premeditated and deliberated upon), the sentencing judge has ample authority to make the sentence fit the bystander’s individual circumstances. State v. Guido, supra, 40 N. J. at 210.
As stated earlier, the trial court instructed the jury that involuntary manslaughter was not in the ease. On the hypothesis that the police officer was not acting in the execution of his duty, involuntary manslaughter would be a possible verdict if the State seeks a conviction notwithstanding that less than grievous bodily harm was intended. Involuntary manslaughter might arguably be involved upon other factual theses the State or the defense might press on the retrial. The subject was not explored in the argument, and since it may be academic, we will not pursue it.
The judgment of the Appellate Division is affirmed and the matter remanded to the trial court for retrial.
As we have already noted, the State did not seek a first degree conviction on the thesis of a “willful, deliberate and premeditated killing.”