In this appeal from a district court denial of a petition for habeas corpus, brought under 28 U.S.C. § 2254,1 petitioner Jeffrey Sinnott seeks to set aside his 1987 conviction for first degree murder based on "extreme atrocity and cruelty." He was indicted in 1983, along with a co-defendant, Gary E. Mosso,2 for the murder of Anthony Tamburro, and found guilty by a Massachusetts Superior Court jury in 1984, the judgment being affirmed by the Massachusetts Supreme Judicial Court in Commonwealth v. Sinnott, 399 Mass. 863, 507 N.E.2d 699 (1987). After post-conviction proceedings were resolved against him, he initiated this action in the district court for the District of Massachusetts.

Two asserted trial errors are before us: (I) introduction into evidence of a written statement of the non-testifying co-defendant, Mosso, in violation of Bruton v. United States, 391 U.S. 123, 88 S. Ct. 1620, 20 L. Ed. 2d 476 (1968); and (II) certain jury instructions concerning malice, alleged to have improperly reduced the Commonwealth's burden of proof. As to each, the question we address is whether any error had a substantial and injurious effect or influence on the jury's determinations. See Brecht v. Abrahamson, 507 U.S. 619, 637, 113 S. Ct. 1710, 1721-22, 123 L. Ed. 2d 353 (1993). We have concluded that in each case the answer is "No."3

Our analysis will require detailed review of certain areas of evidence, but we here sketch the general scenario to give context. This criminal prosecution stemmed from a fight in the early morning of December 9, 1983 in a parking lot behind a nightclub in Oxford, Massachusetts, Kastle's Keep, where the attractions were drinking, dancing, and listening to a band, "The Fools." James Peloquin and his friend, Anthony Tamburro, arrived at Kastle's Keep at 8:30 in the evening of December 8, and at about 1:30 a.m. went to Tamburro's car in the parking lot to get a demonstration tape of another band. They encountered a group of three men, petitioner and co-defendants Mosso and Blanchard.

Conversation then took place with Tamburro concerning the purchase of some cocaine. Soon Mosso purchased a half gram for fifty dollars. Shortly thereafter, he complained about the quality and wanted his money back. A fight started between Tamburro and Mosso, and Tamburro fell or was thrown on the ground, where he was repeatedly struck by Mosso. At this point petitioner intervened, kicking Tamburro in the face. The extent of petitioner's continued participation is the focus of much of the testimony. Suffice it to say here that after only several minutes, Tamburro was rendered inert and unconscious, suffering from abrasions, bruises, lacerations, hemorrhaging and fractures, the injured areas ranging from cranium to testes. Mosso and petitioner, in a car driven by Blanchard, sped away from the scene as several persons summoned by Peloquin came out from the Keep. Tamburro soon stopped breathing, was administered oxygen, received CPR ministrations, and was taken to a hospital where he was pronounced dead at about 2:52 a.m.

The first error alleged by petitioner is the trial court's admission of a written statement given the police by Mosso. Petitioner submitted his own written statement to police in which he said that he had kicked Tamburro only once with the side of his foot, then had stepped aside, trying to keep others from interfering. Meanwhile, Tamburro was getting "beat bad" by Mosso. Mosso's statement represented petitioner as jumping in, early in the fray, and repeatedly kicking for several minutes while he, Mosso, was trying to persuade him to withdraw. Mosso ascribed all injuries inflicted on Tamburro to petitioner. Neither petitioner nor Mosso testified, and objection to admission of Mosso's statement was properly made.

While there was a misimpression of the admissibility of Mosso's statement at the trial stage of this litigation, based on no longer applicable caselaw, it is now undisputed that admission of the incriminating statement violated the Confrontation Clause of the Sixth Amendment because petitioner was deprived of the opportunity of cross examination. See Bruton, 391 U.S. at 137, 88 S.Ct. at 1628-29.

The Supreme Court has made it clear that on collateral review of habeas cases involving trial error, the test for harmless error is not so rigorous as that demanded by Chapman v. California, 386 U.S. 18, 24, 87 S. Ct. 824, 828, 17 L. Ed. 2d 705 (1967) (where the prosecution must establish harmlessness beyond a reasonable doubt), but follows the standard enunciated in Kotteakos v. United States, 328 U.S. 750, 776, 66 S. Ct. 1239, 1253, 90 L. Ed. 1557 (1946): did the error have "substantial and injurious effect or influence in determining the jury's verdict"? Brecht, 507 U.S. at 631, 637, 113 S. Ct. at 1718, 1722.

In arriving at this conclusion, the Court observed that a constitutional violation occurring during presentation of the case to the jury is amenable to harmless-error analysis because it " 'may ... be quantitatively assessed in the context of other evidence presented in order to determine [the effect it had on the trial].' " Id. at 629, 113 S.Ct. at 1717 (citing Arizona v. Fulminante, 499 U.S. 279, 307-308, 111 S. Ct. 1246, 1263-64, 113 L. Ed. 2d 302 (1991)). While the Court determined that habeas relief should not be granted based on merely a "reasonable possibility" that the error contributed to the verdict, id. at 637, 113 S.Ct. at 1721-22, it also recognized that relief should not be denied simply because a reviewing court felt that a petitioner "would have been convicted even if the constitutional error had not taken place." Id. at 642, 113 S.Ct. at 1724 (concurring opinion of Justice Stevens, who cast the deciding vote). As we said in Gilday v. Callahan, 59 F.3d 257, 269 (1st Cir.1995),

[W]e ... must consider--to restate the Brecht standard--whether the error was of such magnitude that it actually casts doubt on the integrity of the verdict. This is the difference between a possibility and a probability.

Moreover, our review is plenary and the burden of establishing harmlessness is on the prosecution. O'Neal v. McAninch, 513 U.S. 432, 436, 115 S. Ct. 992, 994-95, 130 L. Ed. 2d 947 (1995).

We derive some additional guidance for our review from Kotteakos itself. We are told that errors "conceivably may be altogether harmless in the face of other clear evidence, although the same error might turn scales otherwise level." 328 U.S. at 763, 66 S.Ct. at 1247. A court must have "the conviction ... that the error did not influence the jury, or had but very slight effect...." Id. at 764, 66 S.Ct. at 1248. The Court's careful comparative weighing to distinguish a case in which the same error resulted in a reversal, Berger v. United States, 295 U.S. 78, 55 S. Ct. 629, 79 L. Ed. 1314 (1935), see 328 U.S. at 766-74, 66 S.Ct. at 1248-52, informs our own analysis of the likely impact of the offending evidence.

The question we must consider, therefore, is whether admission of Mosso's statement implicating petitioner as the primary combatant had such a "substantial and injurious effect or influence" that it "casts doubt on the integrity of the verdict." Central to this issue is the evidence concerning the extent and nature of petitioner's involvement in the fracas and his state of mind. We shall examine in detail the relevant testimony of four non-party eyewitnesses, that of the pathologist who conducted the autopsy, written statements given the police by both petitioner and Mosso, and the closing argument of the prosecutor, who commented on the two defendants' statements. We shall then assess the likely impact of the admission and use of the Mosso statement.

(1) Eyewitness Testimony. We begin by identifying the crucial and relevant testimony of each of the four non-party witnesses as to the participation of petitioner in the beating of Tamburro. It is clear that the full import of the testimony cannot be adequately conveyed by merely listing the varying number of blows or kicks that each witness saw petitioner inflict. For not only did the witnesses observe for different periods of time, but the surrounding circumstances also add relevant detail. And since our task is to gauge the probability of Mosso's statement having had a substantial effect on the jury's deliberations, we strive to have as complete a picture in mind as did the jury.

James Peloquin was the first witness. A good friend of the victim, he was on the scene at the very beginning and remained until very nearly the end. From 8:30 p.m. until 1:30 a.m. he had only three alcoholic drinks. These are the major points he made:

-- After Mosso complained of the quality of the cocaine Tamburro had sold him, he demanded a sample of Tamburro's personal supply and was offered a "taste." He accused Tamburro of not trusting him, and then took off his jacket. Petitioner took off his own jacket, saying, "So it is going to be one of those."

-- Mosso then hit Tamburro in the face, Tamburro struck back, hitting him lightly, the two wrestled, and Tamburro tripped and fell with Mosso on top of him. Mosso pinned Tamburro's left arm and head, and elbowed him in the eye five or six times.

-- As Tamburro was wriggling, trying to free himself, Peloquin started over but was hit on the back of the head by petitioner and fell down as a result.

-- Petitioner then approached Tamburro and, as he was trying to get up, his hands behind his back in a lifting position, petitioner kicked Tamburro in the face "right square," the front of his ankle hitting the left side of the bridge of Tamburro's nose.

-- Tamburro fell on his back again, legs to one side. Mosso, standing at Tamburro's feet, started kicking him in the lower abdomen, and petitioner, standing above the upper part of his body, "was still kicking Tony" around the face. Peloquin could not say how many times either man struck Tamburro. During this time, Tamburro was apparently helpless, doing nothing.

-- Peloquin then grabbed Tamburro with a bear hug from the rear and dragged him several feet away. He was being hit by both Mosso and petitioner, and then petitioner kicked him in the ribs. Peloquin dropped Tamburro and headed toward the nightclub. Before entering the club to obtain help, he looked around and saw petitioner give one final blow to Tamburro's head. Peloquin acknowledged on cross-examination that he had given four separate statements to the police, all of which varied some in detail. When asked if he had given a statement in which he had said that he did not remember who administered the last kick, Peloquin replied that what he had said on the stand was his best memory.

-- Asked about petitioner's sobriety, Peloquin said he seemed very drunk, and was swaying and slurring his words. He did not recall his staggering or messy clothes, but said that counsel's description of petitioner as "falling down drunk" was a fair statement.

Stephen Amaral, who did not know Tamburro, had come to Kastle's Keep with his friend Ducharme. He had consumed many beers and White Russians (kahlua, vodka and milk) and described himself as very drunk by the time he arrived at the nightclub, at about 12:30 a.m. The two left at approximately 1:45 a.m. These were Amaral's observations about the fighting:

-- From the passenger seat of Ducharme's pickup truck, he saw two men going down on the ground, the smaller one on the bottom, the larger one on top.4 The larger man was facing the feet of the smaller one and punched him in the groin three times. Amaral looked away to talk to Ducharme. When he looked back, the smaller man was on the ground, his back leaning against a car. The larger man, whom he observed to have dark hair, kicked him twice in the head.

-- Then another person, a blonde man, was on the left of the reclining man; he and the larger man then each kicked the victim in the head twice. "They kicked him as if you were going to kick a football on a kickoff or something."

-- The victim then stood up, ran about ten feet, and fell down on his back. Then "[t]he kid with blonde hair went over to him and halfheartedly kicked the side of his face, turning his head from one side to the other." The dark-haired assailant approached the victim and appeared to be checking his pulse. As a group from the nightclub came running toward them, the two men got in a car and took off.

Gary Ducharme had drunk only two beers in two hours. When he and Amaral left Kastle's Keep, Ducharme was to drive. He testified that Amaral did not appear "to be bad to me at all," but could not be sure of his condition. Ducharme's relevant testimony follows:-- He was trying to fix the car radio when Amaral called his attention to a fist fight between two men, one about eight inches taller than the other. The two fell, the larger one on top and facing the other's feet; he punched his opponent in the groin three times. The smaller man was on the ground, tried to push himself up, but was kicked in the face two or three times.

-- At this point Ducharme observed "the other taller kid going in and kicking him also." This person, a blonde haired man, kicked the prone man in the face, alternating with the other man, kicking him "once or twice." The other man also kicked once or twice.

Douglas Cassidy had gone to the same high school as Tamburro, but the two were not friends. He had consumed approximately six to eight drinks from 9:30 p.m. to 1:30 a.m. Although "having a good time," he described himself as "quite aware." When he and his friends prepared to leave, he observed:

-- Two persons were standing beside another, who lay face up on the ground, and were kicking him. The blonde haired person was on the left and kicked the prone individual "a number of times" in "the upper body, head area." The other man also kicked a number of times in the same area.

-- Cassidy then ran over and said to petitioner, "You're a loser." He raised the victim to a sitting position, and the two men who had kicked him took off in a car.

(2) The Co-defendants' Statements. In addition to this testimony from non-party witnesses, a written statement from each defendant was admitted into evidence. Petitioner's statement was in the form of an interview with a state trooper and included the following points:

-- After Mosso purchased the cocaine from a man petitioner described as shorter than himself, Mosso tried it. Shortly thereafter, petitioner got out of the car and saw Mosso and the other man fighting.

-- Petitioner then got out of the car, ran over, and gave Tamburro a side kick in the ribs. Tamburro went down, got up again, and continued fighting.

-- Petitioner "kept other people from jumping into the fight."

-- The "kid" was "getting beat bad.... Gary (Mosso) kept kicking the kid after he was down. I think he went nuts."

-- That night, petitioner was wearing black combat boots.

-- The dispute was over the quality of the half gram of cocaine that Mosso had bought for $50. Mosso complained and wanted his money back. The "kid," about 5 feet eight or nine inches tall, refused, grabbed Mosso by the chest, and hit him in the forehead. The two wrestled to the ground, punching at each other. When Mosso tried to get up, the kid grabbed his right leg by the ankle; Mosso started kicking to get loose and managed to free his leg.

-- At this point, petitioner jumped out of the car and started kicking the kid, kicking him a few times in the face and shoulder area.

-- Mosso told petitioner, "Forget it, cut it out, let's get out of here." Although Mosso went back to their car, petitioner stayed and continued to kick the kid for three minutes.

-- Mosso did not think he "caught [the kid] with any punches at all," and believed that the kid was not hurt by him, but "was hurt bad after Jeff [petitioner] got through kicking him."

(3) The Medical Testimony. Dr. Sussman, Chief of Pathology of the Worcester City Hospital, conducted an autopsy, had photographs taken, and testified at length. In very summary fashion, these were his observations and conclusions:

-- External injuries included multiple abrasions, lacerations, and bruises on the head, face, and both ears, and abrasions on both cheeks; a crepitant nose, with a bubbly feeling; a swollen and hemorrhaging upper lip; abrasions on right and side portions of the back; multiple abrasions on right upper, left side, and front of chest; abrasions on top of the left shoulder; abrasions on knee; and a tear in one thumb.

-- Internal injuries included the brain, which was swollen and had moved from normal position to compress the pons medulla (responsible for controlling respiration and heart beat); the scalp, which experienced hemorrhaging on the left side, near the ear; the larynx and trachea, which contained blood and stomach contents; two fractured ribs, with underlying contusions to the lungs and associated bleeding; the abdominal wall and large intestine, which were hemorrhaging; a bruised spinal cord and hemorrhaging adjacent to the spine; hemorrhaging also seen in and around testicles.

-- All injuries were consistent with kicks except those to the testes, which were more consistent with blows from a fist or elbow.

-- Death was caused by "multiple blunt trauma," there being at least twelve injuries: at least two on the chest, five on the face, one on the back of the neck, one on the hand, one on the shoulder, one on the back, and one in the testes. The most likely cause of death lay in traumas to the head, although the injuries to the chest, neck, testes, and abdomen could have contributed.

(4) The Prosecutor's Closing Argument. Petitioner claims that the prosecutor's closing argument, although not evidence, "enhanced and fortified" the Bruton error by comparing and contrasting the two defendants' statements "with dramatic oral and visual fingerpointing." Here are the salient portions:

If you believe everything that [defendants' counsel] said on behalf of their respective clients, ... then Mr. Mosso, at least, and Mr. Sinnott, didn't do anything, and Mr. Tamburro is really not dead. Because if you listen exclusively to Mr. Sinnott and Mr. Mosso, neither one of them hurt [Tamburro]. If you listen to both of them, what have they done with respect to this trial ... all the way up to today in the final argument, when both of them in effect went like this (illustrating by crisscrossing one hand over the other and pointing index finger) "I'm not responsible."

...

Well, I would submit to you, ladies and gentlemen, that based on all the evidence that you have heard from all the other witnesses, that both Mr. Mosso and Mr. Sinnott were making self-serving statements. They both didn't tell the complete truth. They both told the police what they wanted to convince the police what happened. "Look, I was involved, but I am really not the one who is guilty here. It is him."

C. Our Assessment Absent the Mosso Statement

In Levasseur v. Pepe, 70 F.3d 187, 193 (1st Cir.1995), we identified three factors that are relevant to determining if the jury was "substantially swayed" by offending evidence: "(1) the extent to which the error permeated the proceeding, (2) the centrality of the issue affected by the error to the case as actually tried and (3) the relative strength of the properly admitted evidence of guilt." In this case, there is no question that the issues to which the Mosso statement is most relevant--petitioner's mental states and degree of atrocity--are central. We shall discuss the first factor, prevalence, when we consider the impact of the statement in the next section. In this section, we deal with perhaps the major test of harmlessness--the relative strengths of the cases for the prosecution and the defense absent the offending evidence.

We first sort out the extent of agreement or disagreement in, and the nature of, the non-party eyewitness testimony so that we can assess its general force and weight. We then evaluate the medical evidence. Finally, we consider the potency of the petitioner's case.

Although the evidence shows that each of the eyewitnesses had consumed some alcohol on the night in question, there is little basis for challenging the sobriety of two them--Peloquin and Ducharme, neither of whom had had much to drink. Cassidy was "feeling good," but his claim of being aware is consistent with the span of time over which he had consumed his drinks. Only Amaral admitted to being drunk, with some justification based on his record of consumption. As for opportunity to observe, Peloquin, significantly, was on the scene from the beginning to almost the end, and both Amaral and Ducharme were present for critical parts of the fight. Cassidy was on the scene for the final act.

Most of the evidence, and all of it concerning the core evidence of kicking, was remarkably consistent and corroborative. Although not evidence contributed by the eyewitnesses, petitioner's own statement that he was wearing black combat boots at the time establishes this notable fact. The consensus narrative began with the inception of the fight between Tamburro and Mosso, their wrestling, and falling to the ground.

Peloquin saw Mosso elbowing Tamburro in his eye five or six times. Amaral and Ducharme saw him punching Tamburro in the groin three times. We know from the medical testimony that the testes were injured. In either event, Tamburro at this juncture was supine and helpless. Then came petitioner's entry into the affray, kicking Tamburro. Peloquin is the only non-party witness to a specific first kick, but petitioner in his statement confirmed this fact, differing only in that he said it was a side kick in the ribs, not a side kick to the bridge of the nose.

Then came the period of joint kicking by both petitioner and Mosso, one on either side of Tamburro, the smallest of the three. The quantitative estimates do not dramatically vary: Peloquin refers to petitioner as "still kicking" around the face although he could not give a number of times either man kicked; Amaral said both men kicked Tamburro in the head twice, much in the manner of kicking a football at kickoff; Ducharme described both men as "alternating" in their kicking "once or twice"; and Cassidy described both men as kicking "a number of times" in "the upper body, head area."

Finally, there was substantial consensus, and no contradiction, as to a final kick by petitioner. Peloquin's best memory was that it was petitioner who administered the final blow. Amaral described it as a half-hearted kick, turning Tamburro's head from one side to the other. And Cassidy, while not identifying a single final blow, was sufficiently moved by what he saw to rush toward petitioner and say, "You're a loser."

Not corroborated by other witnesses, but in no way contradicted, was Peloquin's testimony that, at the very outset, petitioner not only removed his jacket along with Mosso, but made the statement, "So it is going to be one of those"; that, as Tamburro was wriggling to escape Mosso's elbowing in his eye, petitioner hit Peloquin on the back of his head as he started to intervene; and that, again, when Peloquin was trying to drag Tamburro to safety, petitioner kicked him in the ribs, causing him to drop Tamburro. Subsequent medical examination at St. Vincent's Hospital corroborated this testimony, revealing contusions in Peloquin's rib areas.

To this testimony must be added the findings of a pathologist reviewing his autopsy of Tamburro: at least twelve separate injuries, comprising abrasions, bruises, lacerations, swellings, discolorations, hemorrhaging, fractures to the head, chest, abdomen, neck, spine, and other vital parts.

All this presents a cohesive, consistent picture of an affray pitting two larger men against a smaller one, with Mosso initiating and petitioner intervening after the smaller man, on the ground and on his back, had suffered a series of hard blows to his eyes and/or testicles, then the two continuing to kick, with petitioner staying in to the end. The injuries were multiple and serious from testes to cranium, and inflicted largely through the medium of violent kicks with, at least in petitioner's case, heavy boots; so many discrete injuries occurred in so many parts of the victim's body that they could not have been produced by only a few kicks or blows. Moreover, the short time span of the affair argues for more than one participant. Perhaps even more persuasive of continued dual participation is the extent to which injuries were observed to both ears, both cheeks, and both sides of the victim's chest, indicating more than one assailant.

To borrow the language of the Court in Brecht, "the State's evidence of guilt was, if not overwhelming, certainly weighty." 507 U.S. at 639, 113 S.Ct. at 1722.

Against this evidence, petitioner's case is not impressive. His claim that all the eyewitnesses had been drinking overlooks the different amounts of alcohol consumed and the time span of consumption, indicating that only Amaral could be said to be drunk. As for the minor differences in the numbers of kicks observed, the witnesses' total testimony, as we have seen, was remarkably internally consistent for a violent, short-lived fracas, observed in the early morning hours by four people at slightly different times from different vantage points. Moreover, it is likely that petitioner's own statement hurt rather than helped him. That he would have rushed in after Mosso had Tamburro on the ground and kicked him in the ribs, then suddenly have withdrawn and merely tried to keep other people from jumping into the fight while Mosso was going "nuts" and beating Tamburro, seems counter intuitive.

After our minute review of all the evidence, quantitative and qualitative, from eyewitnesses and the pathologist, as well as from petitioner, we conclude that, absent Mosso's statement, the evidence was far from being in equipoise, but rather was impressively cogent on the part of the prosecution, any chipping away as the result of vigorous cross examination being canceled out by the unlikely persuasiveness of petitioner's version.

Petitioner argues that the extent of his participation in the beating was critical to two issues: whether he exhibited malice, which is a prerequisite to a finding of even second degree murder (requiring an intent to kill, an intent to do grievous bodily harm, or circumstances that would have led a reasonably prudent person to believe there was a plain and strong likelihood of death), and whether the killing was marked by "extreme atrocity and cruelty," one of the two bases for a first-degree murder finding (the jury having explicitly ruled out the alternative basis, premeditation). Petitioner argues that Mosso's statement "tilted the balance of the evidence," corroborating the testimony of Peloquin and Cassidy, who testified to more extensive involvement on his part than did Amaral and Ducharme.

As we have just concluded, we see no initial "balance" to be tilted. But assuming we are wrong in our assessment, what impact was likely from Mosso's statement? When we juxtapose Mosso's account against this critical body of evidence, we assess the likely effect as follows. Mosso's description of the beginning of the episode, ending with both him and Tamburro on the ground, adds nothing to other accounts except the highly unlikely detail that Tamburro, the smaller man who was content to let the transaction stand, started the fight.

Mosso's statement about petitioner's intervention and kicking Tamburro adds little to petitioner's own admission. It departs from petitioner's version by going beyond a single kick to a few kicks in the face and shoulder area. This observation concededly paralleled the testimony of Peloquin and Cassidy. If the number of kicks were the only important fact, and if Mosso's statement possessed earmarks of credibility, this might be a closer case.

But not only did the testimony of all four eyewitnesses point to a brutal boot-kicking of a defenseless, felled person in the head and upper body, inflicting multiple injuries, but Mosso's statement also had to have fared poorly in the jury's minds. In the first place, Mosso already had demonstrated his ability to fabricate, with a first statement in which he denied knowledge of Tamburro or any fight. Even more important, however, was the balance of the statement, in which Mosso portrayed himself as leaving the fray, asking petitioner to desist, and sitting in the car while petitioner continued kicking Tamburro for three minutes. This version not only contradicted all the eyewitnesses but also was inherently implausible. After all, it was Mosso who had paid the fifty dollars and was the initial fight participant. His further statement that he did not think he had inflicted any damage, and that all of the injuries had to be caused by his co-defendant, could not have affected the jury in any manner adverse to petitioner.

We therefore conclude that no reasonable jury could have been influenced in its determinations, not to mention substantially influenced, by Mosso's statement. We have been pointed to no place in the trial where it played a role except in the prosecutor's closing argument, a subject to which we now turn. Petitioner's claim is that, by coupling petitioner and his co-defendant with their reciprocal finger-pointing statements accusing the other, the prosecutor effectively canceled out petitioner's statement.

If the closing argument's use of the Mosso statement were both the dominant motif and the source of significant prejudice that otherwise would not have existed, the "prevalence" factor might be satisfied and perhaps the previous imbalance of evidence might be overcome.

But such is not the case here. In the first place, we note that petitioner is now asking us to ride two horses, going in opposite directions. On the one hand, petitioner asserts that the prosecutor's use of the Mosso statement "increased the likelihood the jury viewed Sinnott's participation as greater than it was." On the other hand, he also asserts that the evil lies in the prosecutor's deft claim that each was pointing the finger at the other, and thus Mosso's statement improperly canceled out that of petitioner. It is difficult to see how the Mosso statement could at the same time add to the weight of the state's case and also be dismissed as another transparent attempt to shift blame.

Nevertheless, even though the finger-pointing argument was not made at trial or before the state court, we shall, like the district court, consider it. In homage to a search for prevalence, we note that in the course of an eleven-day trial, occupying 1,389 pages of transcript, the prosecutor's challenged remarks occupied at most four pages, one at the very beginning of his 37-page argument and three two-thirds of the way through. The significant fact is that the predominant part of the prosecutor's closing rested on the evidence, not the statements of the defendants.

More to the point is that the basic message conveyed by the prosecutor could and undoubtedly would have been conveyed had there been no Mosso statement, namely, that petitioner was obviously trying to shift blame. When counsel argues that, by showing that another defendant succumbed to the instinct of self preservation, the prosecutor unfairly demeaned petitioner, we cannot quite see the logic. But perhaps, apart from logic, there was some impact from the prosecutor's dramatic ploy of crossing his hands with fingers pointing. All we can say as to this is that at the very most any additional effect on the jury's deliberations was "slight."

We are convinced from our meticulous review of this trial that the admission of the Mosso statement and the use made of it by the prosecutor in his closing argument did not in all probability have a substantial and injurious effect or influence on the determinations of the jury.

Petitioner has advanced four claims of error in the trial court's jury instructions. None was made at trial. The Massachusetts Supreme Judicial Court and the district court have considered them at considerable length and found no merit. At this juncture, we see no point in addressing issues of exhaustion, other procedural problems, or even the correctness vel non of the several instructions. Just as was the case with the Bruton error we have discussed, the substantive question is whether, in this collateral consideration of the record as a whole, any errors, singly or collectively, have been demonstrated by the Commonwealth as having had no substantial and injurious effect or influence in determining the jury's verdict. Libby v. Duval, 19 F.3d 733, 738-40 (1st Cir.1994) (applying Brecht harmless error review to error in giving mandatory presumption instruction).

Petitioner's underlying theme is that the instructions failed in various ways to satisfy Massachusetts case law on how to charge malice, a state of mind that is a prerequisite for finding either first or second degree murder. Massachusetts cases establish that malice can be proven in one of three ways: (1) showing that the defendant, without justification or excuse, possessed an intent to kill, (2) showing that the defendant, without justification or excuse, possessed an intent to cause grievous bodily harm, or (3) showing circumstances in which a reasonably prudent person would have known there was a plain and strong likelihood that death or grievous bodily harm would result. See, e.g., Commonwealth v. Grey, 399 Mass. 469, 470 n. 1, 505 N.E.2d 171, 173 n. 1 (1987); Commonwealth v. Huot, 380 Mass. 403, 408, 403 N.E.2d 411, 414 (1980).5 Although the trial court discussed malice at considerable length, it did not list the three alternatives as such.

As might well be expected in a criminal prosecution of three defendants charged with the most serious offenses, the court's instructions were lengthy, encompassing forty pages of transcript. Those relevant to the claims at issue occupied some fifteen pages. Petitioner's claims are based on passages appearing on three of these pages. Our task is to view these statements in the context of a complicated and lengthy set of instructions and to assess the extent to which they would likely have affected the jury's thought processes.

We discuss first an instruction that petitioner describes as "automatically equating use of a deadly weapon with malice." The court said:

A killing may be malice, and consequently murder, even though the slayer did not wish to cause death. If a man intentionally and without legal justification uses upon the body of another a deadly weapon, such as a gun or a knife, and death results, a jury may infer that malice aforethought is present.

Petitioner asserts that the court never explained that there must be an intent to kill, or to do grievous bodily harm, or that a prudent person would have known that he was creating a plain and strong likelihood of death.

This assertion overlooks a number of other efforts by the court to explain the kinds of actions that would be required to reach the malice threshold. We mention three. The first came several pages earlier than the above quoted passage and addressed the definition of a dangerous weapon. Although the court was discussing the charges against Blanchard, who was being prosecuted as an accessory after the fact to the crime of assault and battery by means of a dangerous weapon, it had to explain the offense of the principals. It identified two kinds of dangerous weapons, quoting from an unidentified case in describing the second type:

One type is by design a dangerous weapon, guns, daggers, they are designed to do serious bodily injury to people....

"[Another type] may also be defined as an instrument ... which, because of the manner in which it is used, or attempted to be used, endangers the life or inflicts great bodily harm, or is calculated as likely to produce death or serious bodily injury...."

It is for you to decide in this case whether as to a particular charge, an instrument such as a foot covering, was used in such a manner that in the particular circumstances of the case, it was a dangerous weapon.

With this discussion of a dangerous weapon and its use, the court, a few pages after the passage first quoted, fleshes out "malice aforethought" by stating that if the jury finds beyond a reasonable doubt that an assault and battery by means of a dangerous weapon has been committed, causing death, and

that the murder was the natural and probable consequence of the commission of the crime, and ... that that crime, assault and battery by means of a dangerous weapon, is a crime that involved inherent danger to human life, then the fact of the commission of that crime, with death resulting, allows you to find the presence of malice aforethought from those facts. So that under those circumstances, you would have at least a murder in the second-degree. (Emphasis added.)

From this overview of the instructions, it is clear, whether required or not, that all three alternative ways of meeting the malice requirement were discussed by the court. Petitioner himself concedes that "the court instructed, according to the first prong of malice, that malice may be found from an intent to kill." The third prong, knowledge of a prudent person that his actions created a plain and strong likelihood of death or grievous bodily harm, was substantially covered by the court's references to using an instrument in a way "calculated as likely to produce death or serious bodily injury," to murder being "the natural and probable consequence" of the use of the dangerous weapon, and to a crime "that involved inherent danger to human life." Any difference between these comments and the "plain and strong likelihood" language of the cases is so slight as not to weigh significantly in the scales of probable injurious effect.

As for the second prong, intent to do grievous bodily harm, the fact that the jury rendered a special finding of extreme atrocity and cruelty excluded any possibility that the jury was misled by the absence of a haec verba instruction. For the court, in addressing the requisites of first degree murder, after saying that more than ordinary atrocity or cruelty was needed, said:

Repeated violent blows have been held to evince such ferocity as would warrant a finding of extreme atrocity or cruelty.

"Evidence that a defendant struck the fatal blow 'for kicks,' " using the colloquial expression. "A crime that was committed with such savagery and brutality, where there were repeated violent blows, which were held to evince a particular ferocity that would constitute extreme atrocity and cruelty."

In our view, there is accordingly no likelihood that the jury would have acted on the assumption that mere use, simpliciter, of a deadly weapon, resulting in death, without the requisite intent or knowledge of consequences, was a sufficient basis for its conclusions. The same extracts from the instructions suffices to dispose of two other claimed errors.

Very similar to the first claim is the assertion that the court erred in creating a mandatory presumption, in violation of the teaching of Sandstrom v. Montana, 442 U.S. 510, 99 S. Ct. 2450, 61 L. Ed. 2d 39 (1979), by use of the following language:

If you have a death resulting under the circumstances I described by one who has committed a felony, an assault and battery by means of a dangerous weapon, then you have a murder in the second-degree.

This excerpt is found in the penultimate paragraph of the entire charge and is clearly part of a thumbnail summary. It is immediately preceded by a page listing the "circumstances" it references; the jurors were instructed that to find malice aforethought from the commission of the crime of assault and battery by means of a dangerous weapon, they must conclude that the crime is one in which "the natural and probable consequence" is death and one which involves "inherent danger to human life."

The language pinpointed by petitioner, suggesting a mandatory presumption, seems to us effectively qualified and explained by the earlier passage. See Hill v. Maloney, 927 F.2d 646, 649 (1st Cir.1990). In any event, even if any error remains, there is in our judgment no likelihood that it had any substantial influence upon the jury's deliberations.

The same analysis and conclusion apply to petitioner's third claim of error, directed to this extract:

There also, of course, can be actual malice. You can find there is ill will and there is hatred and harboring of a grudge, motives of revenge or punishment.

Petitioner claims that this instruction permitted the jury to find malice "upon proof of a mere intent to injure," considerably watering down the required intent or knowledge.

Once again, a statement is taken out of context and without regard to other qualifying instructions. In the first place, the statement follows a reference to situations permitting the inference of malice from use of a deadly weapon, and is part of a systematic effort to limn the contours of malice without at that point filling in details. In the second place, petitioner strains the text in his effort to find error, going beyond the listed examples which suggest mental states prior to the petitioner's involvement in a fight--ill will, hatred, a grudge, revenge, punishment. Ordinarily one would not characterize an instantaneous decision to step into a fight to help a friend as malice.

But, even according "ill will" the meaning of "intent to injure" urged by petitioner, the passages we have quoted above adequately flesh out the kind of action that must accompany the motive of ill will to meet the requirement of malice. Moreover, we cannot imagine how a jury could arrive at a finding of extreme atrocity and cruelty if it felt that petitioner harbored a mere intent to injure, rather than at least an intent to cause grievous bodily harm. This instruction, if not explained and cured by context and other instructions, most certainly possessed no likelihood of influencing the jury's deliberations.

The final claim of error is that, contrary to Massachusetts law, the jury was instructed that it might consider whether defendant had an impaired capacity because of his consumption of alcohol solely as to "first-degree murder, but not on any other degree of offense." Petitioner cites Commonwealth v. Henson, 394 Mass. 584, 476 N.E.2d 947 (1985), for the proposition that voluntary intoxication is relevant to all three species of malice, and thus to second-degree murder as well. The government responds that, even if there were an error of state law, such is not a basis for habeas corpus relief if there is no violation of the Constitution, laws, or treaties of the United States, citing Estelle v. McGuire, 502 U.S. 62, 67-68, 112 S. Ct. 475, 479-80, 116 L. Ed. 2d 385 (1991), and it points out that the defense of voluntary intoxication is not constitutionally required. See Montana v. Egelhoff, 518 U.S. 37, 56-58, 116 S. Ct. 2013, 2024, 135 L. Ed. 2d 361 (1996); Robinson v. Ponte, 933 F.2d 101, 104-105 (1st Cir.1991). We do not reject the government's argument, but prefer to rest on what we think is dictated by the jury's decision.

In the somewhat restrained review we conduct under the dictates of Brecht, we have difficulty in understanding how, in light of the court's lengthy instructions on alcohol-induced impairment of mental capacity to engage knowingly in acts of extreme atrocity or cruelty, one could argue that restricting such instruction to first degree murder was prejudicial. After all, there was no conviction as to any offense except first degree murder by extreme atrocity or cruelty. As to this, the court went beyond saying that the jury could consider defendant's mental and emotional state in light of his alcohol consumption. It added:

Let me be a little bit more precise regarding that. If the jurors determine from the evidence that a defendant had substantially reduced mental capacity at the time that the murder or the crime was committed, they can consider the effect, if any, that the defendant's impaired capacity had on his ability to comprehend the consequences of his choices. Thus, a defendant's mental impairment is to be weighed in evaluating the evidence of the manner and means of inflicting death....

In light of the special verdict finding of extreme atrocity or cruelty, pursuant to these instructions, we do not see how it was possible for the absence of a voluntary intoxication instruction on the issue of malice to have prejudiced petitioner. Although a negative finding as to extreme atrocity or cruelty would not, under the instructions, have barred a second degree murder conviction, the finding of extreme atrocity, notwithstanding the evidence of drunkenness, presupposed a determination of "ability to comprehend the consequences of his choices." That the jury found first degree murder in these circumstances eliminates the possibility that the absence of instruction on second degree murder was prejudicial. In sum, we discern no likelihood that this claimed error influenced the jury's thought processes.

Accordingly, our review of the instructions indicates no error cognizable under habeas corpus strictures.

AFFIRMED.