dissenting:
I dissent from the vacation of appellant’s robbery sentence. The double jeopardy rule of Commonwealth v. Tarver, 493 Pa. 320, 426 A.2d 569 (1981), should not be extended to this case.
In Tarver the Pennsylvania Supreme Court held that the Fifth Amendment guarantee against multiple punishments for the same offense protects a defendant from receiving consecutive sentences for felony-murder and the underlying felony. The Court found that the nature of felony-murder as developed under common law and by statute compelled the conclusion that felony-murder and its underlying felony were the “same offense” for double jeopardy purposes.
Under the common law, murder was distinguished from other types of homicide by the element of malice. In the case of felony-murder the necessary malice was not express, but implied. The common law felony-murder rule permitted the factfinder, in the case of a killing occurring during a felony, to “constructively infer” the essential element of malice from the actual malice accompanying perpetration of the underlying felony. Commonwealth ex rel. Smith v. Myers, 438 Pa. 218, 261 A.2d 550 (1970).
The statute under which appellant and Tarver were convicted did not alter the common law definition of murder, but divided the crime into two degrees. Murder of the first degree encompassed two subcategories: “willful, deliberate and premeditated” killing, and felony-murder. Any murder not falling into either of these categories was murder of the second degree. Act of June 24, 1939, P.L. 872, § 701.
Tarver was convicted after a degree-of-guilt hearing at which a three-judge court specifically found him guilty of *590felony-murder. Later he was convicted and sentenced consecutively for the underlying robbery. On appeal from the sentence for robbery the Supreme Court reasoned that the robbery had supplied not only the malice essential to make the accompanying killing murder, but also the aggravating factor necessary to raise the murder to murder of the first degree. Thus unable to separate first-degree felony-murder from its underlying felony, the Court found a constitutional prohibition against consecutive punishments for each crime.
Appellant’s case diverges quite markedly from Tarver, both on the facts and in analysis.
Appellant robbed a gas station at gunpoint, stole the cash register containing $130 in cash, abducted twenty-year-old Billy Gilmour, the gas station attendant, and drove him twelve miles to an isolated area. There, despite repeated pleas by young Gilmour for his life, appellant fired six bullets into Gilmour’s head and body. When Gilmour continued to writhe, appellant beat him over the head with a tire iron, then kicked Gilmour’s body over an embankment.
The only direct testimony as to the brutal and merciless manner in which appellant killed Gilmour came from Edward Zigga, a trusty at the county prison where appellant was incarcerated. According to Zigga appellant not only related to him the sordid details of the crimes, but also solicited Zigga’s help in having appellant’s wife killed because of her knowledge about the crimes. Zigga’s story of how the killing occurred was fully borne out by the medical, physical, and photographic evidence introduced at trial to prove the cause of Gilmour’s death and the condition and location of his body when found.
Clearly the evidence at appellant’s trial established beyond any conscionable doubt that whoever killed Gilmour specifically intended to kill him. In other words, that the killing was willful, deliberate, and premeditated. Commonwealth v. Davis, 308 Pa.Super. 204, 454 A.2d 92 (1982). The majority now allows that, nevertheless, the jury may have ignored the plain evidence of intent and pronounced appel*591lant guilty of first-degree murder solely on a felony-murder theory. Consistently with the Tarver analysis, then, what the majority is saying is that we must consider Gilmour’s murder as the “same offense” as robbery because the jury might have chosen to “constructively infer” that the killing was malicious from the evidence that appellant robbed a gas station twelve miles down the road from the murder scene. I see no need to give appellant the benefit of such an unreasonable and phantom doubt as to what the jury found.
The majority conjures its doubt from the trial court’s alternative first-degree murder instructions, which told the jury they could convict appellant if they found either that he fired the killing bullets “while in the perpetration” of a robbery, or that he killed Gilmour willfully, deliberately, and with premeditation. These “alternative” legal theories, however, in this case had a single factual theory to which they could apply. Gilmour was shot six times from behind in the head and body and also clubbed over the head with a tire iron. If the jury in finding appellant guilty placed any reliance on the felony-murder theory, under the charge they had to attribute at least the shooting to him. As the Supreme Court said in Commonwealth v. Green, 294 Pa. 573, 584, 144 A. 743, 747 (1929), “We can imagine no act more strongly indicative of an intent to kill than the deliberate shooting of a bullet into another’s head. From such circumstances without qualifying facts the presumption of fact that the slayer intended to take life must arise in a reasoning mind. . . .” The factual presumption of intent logically increases sixfold in this case, where the killer used six bullets instead of one. See also Commonwealth v. Glass, 486 Pa. 334, 405 A.2d 1236 (1979); Commonwealth v. Prenni, 357 Pa. 572, 55 A.2d 532 (1947); Commonwealth v. Drum, 58 Pa. 9 (1868); but see Commonwealth v. O’Searo, 466 Pa. 224, 352 A.2d 30 (1976); cf. Commonwealth v. Sullivan, 472 Pa. 129, 371 A.2d 468 (1977) (evidence of brutality of killing probative of specific intent to kill).
*592Even assuming, then, that appellant’s jury found guilt on the theory that he killed “in the perpetration of a robbery,” is it conceivable that they scrupulously avoided the question whether appellant, while pumping six bullets into Gilmour, intended the boy’s death? Such a conceit is, I believe, a chimera, and a flimsy ground for dissolving a serious felony sentence. An intentional homicide committed in the course of a felony is nonetheless first-degree murder of the willful, deliberate, and premeditated variety, see Commonwealth v. Davis, supra; Commonwealth v. Olds, 322 Pa.Super. 442, 469 A.2d 1072 (1983), and that is the variety of murder we are dealing with here.
The majority believes that Commonwealth v. Sparrow, 471 Pa. 490, 370 A.2d 712 (1977), compels us to doubt that appellant’s jury found intent to kill. I disagree. In Sparrow the trial court gave an alternative first-degree murder charge similar to the one given in this case. But Sparrow killed his victim with one shot while holding him up on the street. The Supreme Court pointed out that while there was ample evidence in the case to support a finding of willful, deliberate, and premeditated slaying, the Court could not rule out that the jury had based its verdict solely on the felony-murder doctrine. Of course, the evidence in Sparrow could rationally and legitimately support a finding that the defendant did not specifically intend the death of his victim. That is emphatically not the case here, as even appellant’s counsel in his closing statement to the jury conceded:
I’m sure there’s not one of you in the jury box ... that will disagree with the opening statement that Mr. Brooks made to you people when this trial began a week ago yesterday, when he talked about what a horrible crime it was, what an execution it was, and we completely and totally agree, to that extent, with the district attorney. There can be no question in anyone’s mind that this was a horrible, horrible, ruthless, cold-blooded murder, the killing of young Billy Gilmour.
*593(Emphasis mine). It was not felony-murder but willful, deliberate, premeditated murder for which appellant was convicted, and I cannot subscribe to any legal argument that would characterize it as the “same offense” as robbery.1
In addition, the majority ignores a sound jurisprudential reason why the Tarver double jeopardy rule should not be extended to this case. By giving the benefit of the rule to appellant, the majority has decided that the rule will be available to any petitioner regardless whether his conviction was final at the time Tarver was announced. Although this Court has held that relief on Tarver grounds must be *594afforded to prisoners whose cases were on direct appeal when Tarver came down, Commonwealth v. Starks, 304 Pa.Super. 527, 450 A.2d 1363 (1982), such relief heretofore has not been applied retroactively beyond the Tarver case itself. My reading of recent Supreme Court pronouncements on the retroactive effect of decisions changing prior law convinces me that the Tarver rule should not apply to judgments of sentence final at the time of the Tarver decision (as Gillespie’s was). See Commonwealth v. Cabeza, 503 Pa. 228, 469 A.2d 146 (1983) (opinion by Larsen, J.; dissenting opinion by Nix, J.); Commonwealth v. Geschwendt, 500 Pa. 120, 454 A.2d 991 (1982) (plurality opinion by Nix, J.); Commonwealth v. Brown, 494 Pa. 380, 431 A.2d 905 (1981). See also United States v. Johnson, 457 U.S. 537, 102 S.Ct. 2579, 73 L.Ed.2d 202 (1982); Desist v. United States, 394 U.S. 244, 89 S.Ct. 1030, 22 L.Ed.2d 248 (1969); Stovall v. Denno, 388 U.S. 293, 87 S.Ct. 1967, 18 L.Ed.2d 1199 (1967); cf. Robinson v. Neil, 409 U.S. 505, 93 S.Ct. 876, 35 L.Ed.2d 29 (1973) (double jeopardy rule given full retrospective application where prior constitutional decision did not depart from prevailing law, so that convictions obtained in violation of decision were void ab initio).
Chief Justice Bell, dissenting in Commonwealth ex rel. Smith v. Myers, supra, quoted Justice Cardozo from The Nature of the Judicial Process, 66-67: “When they [judges] are called upon to say how far existing rules are to be extended or restricted, they must let the welfare of society fix the path, its direction and its distance ... The final cause of law is the welfare of society ...” 438 Pa. at 248, 261 A.2d at 565. I can conceive of no way in which the welfare of society might be served by today’s extension of the Tarver rule. Under the 1939 Crimes Code the degrees and penalties for intentional murder and felony-murder were identical, so there rarely was any reason for a jury deciding to convict a defendant on combined murder-felony counts to specify whether the defendant acted with intent to kill. Now all defendants convicted by means of such unspecific verdicts are ipso facto felony-murderers, regardless *595whether the jury would have labelled them intentional killers had they foreseen the legal loophole now available. Scores of premeditated murderers can now expect to walk through the loophole to have their consecutive felony sentences expunged in Post Conviction Hearing Act proceedings.
Of course, all these murderers are left, along with appellant, serving mandatory life sentences. But their newfound blanket categorization as felony-murderers is not a distinction without a difference. It is well known among sentencing judges that one reason for adding a consecutive felony sentence to a life sentence is to increase the time served by a defendant before he comes up for parole. See Commonwealth v. McClendon, 495 Pa. 467 n. 3, 434 A.2d 1185 n. 3 (1981) (dissenting opinion by Roberts, J., and cases there cited). I take it on eminent, albeit second-hand, authority that “ ‘As the Court well knows a life sentence in Pennsylvania at the present time carries an average sentence in the neighborhood of fifteen years.’ ” Spaeth, J., dissenting in Commonwealth v. Burno, 310 Pa.Super. 564, 568, 456 A.2d 1080, 1081-82 (1983) (quoting an unnamed assistant district attorney). I am troubled by the prospect that the enormity of acts like Mr. Gillespie’s might be forgotten in the passage of time, and that today’s decision could help such a cold-hearted young man and others similarly situated to reach the street after many hard-bitten years in the state penitentiary, but sooner than their sentences intended.2
. The majority misses the point when it says in footnote 12 of its opinion that a jury finding of felony-murder "would always be legitimate.” Nobody is denying that a jury is empowered to return any verdict it sees fit if circumstances warrant. For example, a jury may exercise uncontrolled leniency to lessen the degree of the crime proven, or even acquit the defendant, although the verdict be manifestly inconsistent with the evidence. See Commonwealth v. Jones, 457 Pa. 563, 319 A.2d 142, cert. denied, 419 U.S. 1000, 95 S.Ct. 316, 42 L.Ed.2d 274 (1974); Commonwealth v. Staples, 324 Pa.Super. 296, 471 A.2d 847 (1984); but see Commonwealth v. Carter, 502 Pa. 433, 466 A.2d 1328 (1983) (McDermott, J., concurring). However, ordinarily we require that a jury’s verdict of guilty be rationally based in the evidence. See Commonwealth v. Moore, 463 Pa. 317, 325-29 & n. 7, 344 A.2d 850, 855-56 & n. 7 (1975) (Roberts, J., concurring). Here there simply is no reason to suppose that the jury might have found appellant guilty of felony-murder. Besides the fact that the evidence unequivocally proved premeditated murder, the inherent, arbitrary powers of a jury to find a lesser crime or exercise mercy simply were not in play, since premeditated murder and felony-murder at the time were both first-degree and carried identical penalties. Compare, e.g., Commonwealth v. Hoffman, 439 Pa. 348, 266 A.2d 726 (1970) (power of jury to return voluntary manslaughter verdict on murder indictment).
I am not attempting in the least to arrogate the jury’s prerogative to “weigh the evidence." I am stating the plain and commonsensical fact that there is not the slightest possibility the jury relied exclusively on a felony-murder theory in reaching its verdict. The majority, by saying the jury might have relied on the theory anyway, is importing unwarranted irrationality to a facially sound verdict. I disagree with the majority because its application of a constitutional rule of law turns on an irrational possibility it has manufactured. I do not, as the majority suggests, shrink from my judicial responsibility to apply the rule in Tarver, as my opinions applying the rule attest. Commonwealth v. Maddox, 307 Pa.Super. 524, 453 A.2d 1010 (1982); Commonwealth v. Fortune, 305 Pa.Super. 441, 451 A.2d 729 (1982).
. In belittling our reliance on the wisdom of an unnamed assistant district attorney, the majority also slights an important point: that nobody really knows how much time a defendant sentenced to life imprisonment will have to serve. It is probably common knowledge that a sentence of life imprisonment does not necessarily spell life in prison for the defendant; certainly it is common for juries in murder cases to inquire about the actual length of a life term. See Commonwealth v. Johnson, 368 Pa. 139, 81 A.2d 569 (1951); see also Commonwealth v. Rogozinski, 387 Pa. 399, 128 A.2d 28 (1956); Commonwealth v. Ballem, 386 Pa. 20, 123 A.2d 728 (1956); Commonwealth v. Bibalo, 375 Pa. 257, 100 A.2d 45 (1953); cf. Commonwealth v. Chapasco, 436 Pa. 143, 258 A.2d 638 (1969) (jury recommended parole not be considered); Commonwealth v. Celijewski, 324 Pa.Super. 185, 471 A.2d 525 (1984) (semble). Despite the concern jurors often display about what will happen if they sentence someone to life, it is improp*596er for trial judges or district attorneys to discuss with them the possibilities of commutation and parole. Commonwealth v. Aljoe, 420 Pa. 198, 216 A.2d 50 (1966); Commonwealth v. Mills, 350 Pa. 478, 39 A.2d 572 (1944). See generally Annot., 16 A.L.R.3d 1137 (1967).
Given the general confusion about the meaning of a life term, some Pennsylvanians might be surprised to learn that there are no statutory guidelines governing decisions to parole life prisoners. The answer to the question "How much time will a lifer serve?” falls somewhere in the gray area among the bailiwicks of various policy-making officials in the administrative branch of government. In the first instance, before a life prisoner may be paroled the Governor must commute his minimum sentence to a term of years. This the Governor may do only on the recommendation of the Board of Pardons. Pa. Const, art. 4, § 9; The Administrative Code of 1929, § 909, as amended by Act of July 31, 1968, P.L. 769, No. 240, art. VI, § 609(8), 71 P.S. § 299. Next, after expiration of the term of years as commuted, the Board of Parole must exercise its discretion whether to parole the prisoner, subject, however, to the authority of the Board of Pardons to order parole if the Board of Parole does not. Act of Aug. 6, 1941, P.L. 861, § 21; as amended, 61 P.S. § 331.21. Of course, in deciding whether to grant parole the Parole Board must consider any recommendations made by the judge who heard the case and who may even have imposed the original sentence. But the final decision on how much time a life prisoner will serve remains purely an administrative matter.
Although my statistics are not completely up to date, they indicate that parole is a looming possibility for a life prisoner. For the years 1967-1980, the Board of Pardons heard 1251 applications by life prisoners for commutations of sentence, and in 379 (30%) of these cases recommended to the Governor that sentence be commuted. Analyses of the Action of the Board of Pardons Sessions for the Calendar Years 1967-1980. And according to tables prepared for the Parole Board, over the years 1962-1971 inclusive 180 life prisoners were released after having served an average of 19 years, 7 months in prison; for the years 1971-1980 inclusive 219 were released after an average of 17.7 years served. (Figures for the year 1971 are different in the two tables I have access to, "Lifers Released on Parole During the Past Ten Years [1962-1971],” and "Time Served by Commuted Lifers 1971 to 1980").
Of course, averages are only averages, and who is to say what they will mean when the Governor and Boards of Pardon and Parole come to assess Mr. Gillespie’s case? Maybe our unnamed assistant district attorney’s estimate “in the neighborhood of fifteen years” will prove inaccurate, and Gillespie will stay in jail much longer than that. On the other hand, there have been quite a few life prisoners who have won their freedom well before serving fifteen years. See, e.g., Pennsylvania Board of Parole, A Report of Special Commutation Cases, Table 3 (of 29 first-degree murderers released during period 12/7/52-6/21/54, 23 had served less than 15 years; of these, 7 had served less *597than 10 years). Whatever eventually happens in Mr. Gillespie’s case, he moves one step closer to the prospect of release with his victory in Court today.