KAPLAN, J.

The indictments for the homicides in the present cases would fail if a "year and a day" rule were applied, since in each of the cases the victim died more than that period of time after the criminal act. More particularly: In Commonwealth vs. Randolph Lewis (and three companion cases), the defendants were indicted on September 8, 1978, for the murder of Richard Poleet, death having occurred on May 30, 1978, as a result, so the Commonwealth claimed, of the defendants' assault upon him on April 19, 1976. (Two of the defendants had been found guilty on May 4, 1977, of other crimes arising from the same incident.) In the consolidated case, Commonwealth vs. Lanier W. Phillips, Jr., the assault occurred on October 22, 1977, the victim Daniel Mark Duffault died on April 27, 1979, and indictment for murder was handed up on June 13, 1979.

On motions of the defendants in these cases to dismiss the indictments, judges of the Superior Court rendered reports which inquired, in effect, whether the prosecutions might lawfully be maintained. The reports were to be expected, and are appropriate, as severe doubts about the modern day justification for the year and a day rule were expressed by us in the recent case of Commonwealth v. Golston, 373 Mass. 249 (1977), cert. denied, 434 U.S. 1039 (1978).

Any discussion of the subject begins with the antique statute 6 Edw. 1, c. 9 (1278), which declares as to the private form of prosecution for murder called "appeal," that it shall not be abated so soon as it has been heretofore; but if the appellor declare the "Deed" and the time when it was done, and with what weapon the victim was slain, the appeal shall stand and not abate "if the Party shall sue within the Year and the Day after the Deed done." "Deed" was read as meaning the death, not the act of assault leading to the death, and evidently it was from the date of the death that the period was measured. See the discussion in Commonwealth v. Ladd, 402 Pa. 164, 166-169 (1960). See also Commonwealth v. Macloon, 101 Mass. 1, 8 (1869); Louisville, Evansville, St. Louis R.R. v. Clarke, 152 U.S. 230, 239-241 (1894). If "Deed" were read as referring to the act, then the act might not have eventuated in a death within the year, and so there would be no basis for an appeal. On the other hand, it would be hard to understand the statute to say that appeal would lie within a year and a day of the act, provided death also occurred within that time. Possibly the draftsman gave no thought to the problem of an interval between the act and the death. At any rate, according to the generally accepted interpretation, the statute operated like an ordinary statute of limitations.

The "appeal" became obsolete long before it was formally extinguished in 1819 by the Stat. 59 Geo. 3, c. 46, and our real interest is in Crown prosecutions for homicide. By the Eighteenth Century, and indeed much earlier, we find a general assumption that a homicide could be prosecuted as such only if the death occurred within a year and a day of the act; this was distinct from any question of the period of limitations for commencing a prosecution. See Commonwealth v. Ladd, supra at 169. Here we have the year and a day rule that is questioned in the present cases. Blackstone states this rule separately from the one about appeals deriving from the statute of Edward I (4 W. Blackstone, Commentaries 197-198, 310-311 [1769]), but it is natural to suppose, although it has not been demonstrated, that there was a connection between the two. The standard, if perhaps unhistorical, explanation of the rule, often repeated in the books, is that, in the condition of medical science until recent times, it would have been hard to establish convincingly a line of causation between an act and a relatively distant death, and it was thus plausible to make the presumption ("conclusive" as well as arbitrary) that a death more than a year removed from the assault or similar antecedent arose from a natural rather than the criminal cause. See State v. Brown, 21 Md. App. 91, 94-95 (1974). Occasionally it has been surmised that the rule was linked in some way to the early function of the jury as reporters of the happenings of the vicinage who required no aid from witnesses — but the jury would not have had knowledge sufficient to trace cause to effect over a sizeable interval of time. See State v. Sandridge, 5 Ohio Op. 3d 419, 420 (C.P. 1977). Again we find a suggestion that the rule was intended simply to soften the old brutal law regarding homicides. See Commonwealth v. Evaul, 5 Pa. D. C. 105, 106 (Q.S. Phila. County 1924).

These reasons are far from persuasive today as justification for the rule. In particular the rule appears anachronistic upon a consideration of the advances of medical and related science in solving etiological problems as well as in sustaining or prolonging life in the face of trauma or disease. See Commonwealth v. Golston, 373 Mass. 249, 254-255 (1977); State v. Brown, supra at 97; People v. Snipe, 25 Cal.App.3d 742, 747 (1972). Thus the relatively short time limit is seen as not only capricious but as senselessly indulgent toward homicidal malefactors. It is reckoned a sufficient safeguard for defendants that the prosecution, quite apart from the rule, must establish the connection between act and death by proof beyond a reasonable doubt. If, nevertheless, the rule can be said still to exist in a large number of jurisdictions in this country (see Annot., 60 A.L.R.3d 1323 [1974]), this is accounted for by the perdurability of statutes in some of them stating the rule, by the fact that over past years there have been few occasions on which the issue has been raised and presented squarely to the courts for decision, and by a tendency to regard so old a dogma as peculiarly suitable for interment by Legislatures not courts. Yet there has been some movement in the courts as we shall see.

The parties in the present actions have debated whether there has been a year and a day rule in Massachusetts and whether it is right to abolish such a rule by judicial decision. If the law is a prediction of what the courts will do, as Holmes averred, then, without entering upon refinements of the problem of our "reception" of English law, we can assert with confidence that the rule formed part of the law of the Commonwealth after the Revolution. The authority of Blackstone would itself have been influential. A dictum by Parker, C.J., recognizing the rule appeared in 1824 ( Commonwealth v. Parker, 2 Pick. 550, 558 [1824]), and the remark or its substance has then been repeated several times in dicta. Commonwealth v. Macloon, supra. Commonwealth v. Robertson, 162 Mass. 90, 97 (1894). Commonwealth v. Snell, 189 Mass. 12, 18 (1905). Commonwealth v. Vanetzian, 350 Mass. 491, 493 n. 1 (1966). Commonwealth v. Pinnick, 354 Mass. 13, 15 n. 1 (1968). But there has not been a holding. Meanwhile some other courts were expressing discomfort with the rule, while suggesting that the Legislatures were the proper instruments of reform. A court questioned whether the rule really covered all varieties of homicide and decided that it did not, that its province should not be extended to the kind of homicide there at bar. New York held that its statutes fully defined the crimes denounced, and as the statutes on homicide omitted mention of the year and a day rule, the rule was not in effect. People v. Brengard, 265 N.Y. 100, 105-108 (1934); People v. Legeri, 239 App. Div. 47 (N.Y. 1933). Pennsylvania construed the rule as being one of evidence, not substance, and a poor one, at that; so, even if the rule had been "received" in the common law of the State, it could be wiped out by decision and the decision applied "retroactively" to the homicide in the case before the court: a change of a mere point of evidence was thought not to invoke the policy against ex post facto laws. Commonwealth v. Ladd, 402 Pa. 164 (1960). An intermediate appellate court in New Jersey found that the rule was part of the common law of the State, and was substantive, not merely evidential or procedural in character, but proceeded to abrogate it because its only cognizable reason — the one about medical competence — could no longer be accepted; and the abrogation was given effect in that case. State v. Young, 148 N.J. Super. 405 (1977).

In this condition of creeping inanition of the rule, we noted in Commonwealth v. Golston, supra at 255, that none of our references to it in past cases had been essential to decision, and we added, "We take this occasion to announce that if the point comes before us we shall feel free to reexamine the justification for the rule." This statement invited direct challenge of the rule which is finally made in the present cases; it might also have stimulated a legislative solution, but that has not occurred.

Not long after our Golston decision, the highest court of New Jersey reviewed the cited ruling of the intermediate court. State v. Young, 77 N.J. 245 (1978). All seven Justices agreed that the rule should be characterized as substantive and that it had been absorbed into the common law of the jurisdiction. Five Justices also agreed that it should be abolished by decision, of whom four preferred that the decision should apply "prospectively" to future acts to avoid ex post facto difficulties, while one would have gone so far as to apply it to past acts.

We share the view that the rule is no longer supportable in reason, and that its relegation to the shades of history may be accomplished by court decision. As to the ex post facto problem, characterization of the rule as "evidential" or "substantive" does not strike us as the key here. Nor can the matter really turn here on ideas of justified reliance by a defendant. Rather the constitutional provisions should be seen as intended to discourage badly motivated or erratic action improper in a lawgiver. See Commonwealth v. Davis, 380 Mass. 1, 16 n. 22 (1980). Accordingly, we think a fair solution is to say that the old rule continues to apply to acts antedating a declaration by the court that the rule was vulnerable. The result reached in New Jersey is consistent with this thought, as the declaration there came only when the court spoke in the very decision. In the situation at bar, a declaration quite sufficient for the purpose was made when we handed down our opinion in Golston on August 26, 1977. Cf. Whitney v. Worcester, 373 Mass. 208, 225 (1977); Commonwealth v. Barnes, 369 Mass. 462, 468 (1976). We believe the selection of such a cleavage date comports with the ex post facto clauses.

Justice Conford in State v. Young, supra, 77 N.J. at 255-257, suggested that, ideally, the year and a day should be replaced by some definite interval from act to death which would reflect modern enhanced scientific capabilities but at the same time give due regard to the undesirability of keeping a threat of prosecution for homicide overhanging an actor for an indefinite period. See also State v. Brown, 21 Md. App. 91, 97 (1974). Such a task of adjustment is characteristically for the Legislature, but if not undertaken by that branch, may fall to the courts. Meanwhile the full measure of proof of causation will be as much as a defendant may claim as a matter of right.

The reports, questioning whether the prosecutions for homicide may be maintained, are answered in the negative as to the defendants in the Lewis and companion cases, and in the affirmative as to the defendant in the Phillips case. The cases are remanded for appropriate further proceedings.

So ordered.