¶1Concurring Opinion by
¶2Without hesitation, I concur in the result reached by the majority today. And I wholeheartedly agree with the majority’s opinion to the extent that it holds unconstitutional
¶3*126My approach, is only slightly different. In this criminal case, the inference used by the trial court
¶4Here, as the majority concludes (and I agree), the inference that the possessor of recently stolen property is the thief does not satisfy the reasonable doubt standard.
¶5 This conclusion is compelled by the due process clause. U.S. Const, amend. XIV. The same result would be reached under article I, section 9 of the Pennsylvania Constitution.
¶6 I am constrained also to comment on the majority’s treatment of the differences between presumptions and inferences. As the majority knowingly explains, there are several reasons in logic and history for treating presumptions differently from inferences. However, in my view, there is no difference between them when the constitutionality of their use is tested.
¶7The notion that there is any distinction between presumptions and inferences for the purpose of determining whether their use affords an accused due process was repudiated by Barnes v. United States, 412 U.S. 837, 93 S. Ct. 2357 (1973). There, the United States Supreme Court held constitutional the inference of guilty possession from proof of possession of recently stolen checks, because the proven facts were “sufficient to enable the jury to find beyond a reasonable doubt that petitioner knew the checks were stolen.” Id. at 845, 93 S. Ct. at 2363. In reaching its result, the Court considered the teachings of its earlier cases without regard to whether they dealt with presumptions or inferences. Id. at 841-44, 93 S. Ct. at 2360-62.
¶8*125It is true that the result in Commonwealth v. Shaffer, 447 Pa. 91, 288 A.2d 727 (1972), was justified on an alternate ground that, in the view of the majority, inferences required different constitutional scrutiny than rebuttable presumptions. Thus, Commonwealth v. Owens, 441 Pa. 318, 271 A.2d 230 (1970), and the “more-likely-than-not” test were inapplicable. The view that presumptions and inferences are to be given different constitutional treatment was later rejected in Commonwealth v. MacFarland, 452 Pa. 435, 439, 308 A.2d 592, 594 (1973).
¶9In Leary v. United States, 395 U.S. 6, 89 S. Ct. 1532 (1969), the Supreme Court did not “reach the question whether a criminal presumption which passes muster when so judged [by the more-likely-than-not standard] must also satisfy the criminal ‘reasonable doubt’ standard if proof of the crime charged or an essential element thereof depends upon its use.” Id. at 36 n.64, 89 S. Ct. at 1548, n.64. See Barnes v. United States, 412 U.S. 837, 841-44, 93 S. Ct. 2357, 2361-62 (1973) ; Turner v. United States, 396 U.S. 398, 404-05, 90 S. Ct. 642, 646 (1970). See also Commonwealth v. Owens, 441 Pa. 318, 323 n.4, 271 A.2d 230, 232 n.4 (1970).
¶10Nevertheless, a close reading of recent decisions of the United States Supreme Court dealing with inferences reveals that the Court has decided cases consistently with a reasonable doubt standard. Barnes, supra;Turner, supra;Leary, supra;United States v. Romano, 382 U.S. 136, 86 S. Ct. 279 (1965) ; United States v. Gainey, 380 U.S. 63, 85 S. Ct. 754 (1965). See Tot v. United States, 319 U.S. 463, 63 S. Ct. 1241 (1943). Cf. W. LaFave & A. Scott, Handbook on Criminal Law 149 (1972).
¶11In both Barnesand Turner, the Supreme Court did not conclude the inferences in question were constitutional until satisfied that the standard of “beyond a reasonable doubt” was met. It is true that in striking down inferences, the Supreme Court has not used a reasonable doubt standard. E.g., Leary; Romano; Tot; see Turner. For two reasons, however, the rationale of these inference cases is not here controlling. First, the approach of the Court when testing the constitutionality of inferences has developed over time. Each decision has added more precision to the articulation of its *126standard. Those cases employing standards other than beyond a reasonable doubt were decided before either Turneror Barnes. Second, the earlier tests of “rational connection” or “more likely than not” are clearly less rigorous than a reasonable doubt standard. Compare, Romano; Gainey; Tot (all using “rational connection” test), with Leary(“more likely than not” standard). See Turner. Therefore, if an inference cannot pass muster under one of these less stringent tests, then it must of necessity fail when measured by the reasonable doubt standard. See Barnes v. United States, supra at 841-44, 845-46, 93 S. Ct. at 2361-62, 2363 (1973).
¶12United States v. Gainey, supra, is distinguishable. It was an early ease in the development of the Supreme Court’s thinking with respect to the constitutionality of presumptions and inferences. There, the Court held constitutional a statutory inference that permitted presence at the site of an illegal still to be sufficient evidence that the person present was “carrying on” the business of Illegal distillation. The majority had little difficulty reaching its result, since the offense was “one of the most comprehensive of the criminal statutes designed to stop the production and sale of untaxed liquor.” Id. at 67, 85 S. Ct. at 757. The Supreme Court did not utilize either a “more-likely-than-not” or a “reasonable doubt” standard, but instead, following Tot v. United States, 319 U.S. 463, 63 S. Cf., 1241 (1943), concluded that there was a rational connection between the fact presumed and the facts proved. In view of the embracive nature of the criminal statute in Gainey, the statutory inference there might well satisfy today’s more rigorous reasonable doubt standard. See Barnes; Turner.
¶13The teaching of those cases in which the Supreme Court held inferences invalid, as measured by a standard other than reasonable doubt, serves only as the first step toward deciding whether an inference is valid. If proof of a crime charged or an essential element of a crime depends upon an inference, then that inference must guarantee beyond a reasonable doubt that the inferred fact follows from the proved facts.
¶14 The trial court instructed the jury in the following fashion:
¶15“The unexplained possession of recently stolen property by the defendant [appellant], if you find as a fact that he did have possession thereof, can give rise in the jury’s mind to an inference that he was the thief, but such is an inference of fact and not of *127law, and you must pass upon it as part of the evidence against the defendant.
¶16“You cannot draw such an inference that the defendant was the thief unless you can conclude, with substantial assurance, that such inferred fact is more likely than not to flow from the proof of his possession of the checks.”
¶17From this instruction the jury could have believed it permissible to convict appellant if satisfied that it was only “more likely than not” that appellant was the thief. Although the trial court did admonish the jury that it was compelled to find appellant guilty beyond a reasonable doubt, the challenged instruction quite possibly could have induced the jury to eonvict on a lesser standard of proof. The practical effect of the challenged instruction was to submit the ease to the jury on an alternative theory. In such a case, if any of the theories are unconstitutional, the judgment of sentence must be reversed. See Leary v. United States, 395 U.S. 6, 31-32, 89 S. Ct. 1532, 1545-46 (1969) ; Stromberg v. California, 283 U.S. 359, 51 S. Ct. 532 (1931).
¶18 “[T]he Due Process Clause protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to constitute the crime with which he is charged.” In re Winship, 397 U.S. 358, 364, 90 S. Ct. 1068, 1073 (1973). See note 3 of majority opinion.
¶19 See Commonwealth v. Shaffer, 447 Pa. 91, 116-17, 288 A.2d 727, 741 (1972) (Pomeboy, T., concurring and dissenting). The inference that the possessor of recently stolen goods was the thief of those goods was at issue in Shaffer. I agree with the views stated there by Mr. Justice Pomeboy. He found “it difficult to believe *128. . . that it is more likely than not that the possessor was himself . . . the thief, or that . . . [the inference] is the equivalent of proof beyond a reasonable doubt of the elements of [larceny].” Id. at 117, 288 A.2d at 741 (concurring and dissenting opinion). If an inference cannot conform to the requirements of the less stringent “more-likely-than-not” test, then it must necessarily fail to satisfy the reasonable doubt standard. See note 3 supra.