delivered the opinion of the Court.

This case raises questions concerning the interpretation of that part of § 811 (c) of the Internal Revenue Code which for estate tax purposes requires including in a decedent’s gross estate the value of all the property the decedent had transferred by trust or otherwise before his death which was “intended to take effect in possession or enjoyment at or after his death . . . .” Estate of Spiegel v. Commissioner, post, p. 701, involves questions which also depend upon interpretation of that provision of § 811 (c). After argument and consideration of the cases at the October 1947 Term, an order was entered restoring them to the docket and requesting counsel upon reargument particularly to discuss certain questions *634broader in scope than those originally presented and argued. Journal Supreme Court, June 21, 1948, 296-298. Those additional questions have now been fully treated in briefs and oral arguments.

This case involves a trust executed in 1924 by Francois Church, then twenty-one years of age, unmarried and childless. He executed the trust in New York in accordance with state law. Church and two brothers were named co-trustees. Certain corporate stocks were transferred to the trust with grant of power to the trustees to hold and sell the stocks and to reinvest the proceeds. Church reserved no power to alter, amend, or revoke, but required the trustees to pay him the income for life. This reservation of life income is the decisive factor here.

At Church’s death (which occurred in 1939) the trust was to terminate and the trust agreement contained some directions for distribution of the trust assets when he died. These directions as to final distribution did not, however, provide for all possible contingencies. If Church died without children and without any of his brothers or sisters, or their children, surviving him, the trust instrument made no provision for disposal of the trust assets. Had this unlikely possibility come to pass (at his death there were living, five brothers, one sister, and ten of their children) the distribution of the trust assets would have been controlled by New York law. It has been the Government’s contention that under New York law had there been no such surviving trust beneficiaries the corpus would have reverted to the decedent’s estate. This possibility of reverter plus the retention by the settlor of the trust income for life, the Government has argued, requires inclusion of the value of the trust property in the decedent’s gross estate under our holding in Helvering v. Hallock, 309 U. S. 106.

*635The Hallock case held that where a person while living makes a transfer of property which provides for a reversion of the corpus to the donor upon a contingency terminable at death, the value of the corpus should be included in the decedent’s gross estate under the “possession or enjoyment” provision of § 811 (c) of the Internal Revenue Code.1 In this case, the Tax Court, relying upon its former holdings2 declared that “The mere possibility of reverter by operation of law upon a failure of the trust, due to the death of all the remaindermen prior to the death of decedent, is not such a possibility as to come within the Hallock case.” This holding made it unnecessary for the Tax Court to decide the disputed question as to whether New York law operated to create such a reversionary interest. The United States Court of Appeals for the Third Circuit, one judge dissenting, affirmed on the ground that it could not identify a clear-cut mistake of law in the Tax Court’s decision. 161 F. 2d 11. The United States Court of Appeals for the Seventh Circuit in the Spiegel case found that under Illinois law there was a possibility of reverter and reversed the Tax Court, holding that possible reversion by operation of law required inclusion of a trust corpus in a decedent’s estate. Commissioner v. Spiegel’s Estate, 159 F. 2d 257. Other United States courts of appeal have held the same.3 *636Because of this conflict we granted certiorari in this and the Spiegel case.

Counsel for the two estates have strongly contended in both arguments of these cases that the law of neither New York nor Illinois provides for a possibility of reverter under the circumstances presented. They argue further that even if under the law of those states a possibility of reverter did exist, it would be an unjustifiable extension of the Hallock rule to hold that such a possibility requires inclusion of the value of a trust corpus in a decedent’s estate. The respondent in this case pointed out the extreme improbability that the decedent would have outlived all his brothers, his sister, and their ten children. He argues that the happening of such a contingency was so remote, the money value of such a reversionary interest was so infinitesimal, that it would be entirely unreasonable to hold that the Hallock rule requires an estate tax because of such a contingency. But see Fidelity-Philadelphia Trust Co. v. Rothensies, 324 U. S. 108, 112.

Arguments and consideration of this and the Spiegel case brought prominently into focus sharp divisions among courts, judges and legal commentators, as to the intended scope and effect of our Hallock decision, particularly whether our holding and opinion in that case are so incompatible with the holding and opinion in May v. Heiner, 281 U. S. 238, that the latter can no longer be accepted as a controlling interpretation of the “possession or enjoyment” provision of §811 (c).4 May v. Heiner held that the corpus of a trust transfer need not *637be included in a settlor’s estate, even though the settlor had retained for himself a life income from the corpus. We have concluded that confusion and doubt as to the effect of our Hallock case on May v. Heiner should be set at rest in the interest of sound tax and judicial administration. Furthermore, if May v. Heiner is no longer controlling, the value of the Church trust corpus was properly included in the gross estate, without regard to the much discussed state law question, since Church reserved a life estate for himself. For reasons which follow, we conclude that the Hallock and May v. Heiner holdings and opinions are irreconcilable. Since we adhere to Hal-lock, the May v. Heiner interpretation of the “possession or enjoyment” provisions of § 811 (c) can no longer be accepted as correct.

The “possession or enjoyment” provision appearing in § 811 (c) seems to have originated in a Pennsylvania inheritance tax law in 1826.5 As early as 1884 the Supreme Court of Pennsylvania held that where a legal transfer of property was made which carried with it a right of possession with a reservation by the grantor of income and profits from the property for his life, the transfer was not intended to take effect in enjoyment until the grantor’s death: “One certainly cannot be considered, as in the actual enjoyment of an estate, who has no right to the profits or incomes arising or accruing therefrom.” Reish, Adm’r v. Commonwealth, 106 Pa. 621, 526. That court further held that the “possession or enjoyment” clause did not involve a mere technical question of title, but that the law imposed the death tax unless one had parted *638during his life with his possession and his title and his enjoyment. It was further held in that case that the test of “intended” was not a subjective one, that the question was not what the parties intended to do, but what the transaction actually effected as to title, possession and enjoyment.

Most of the states have included the Pennsylvania-originated “possession or enjoyment” clause in death tax statutes, and with what appears to be complete unanimity, they have up to this day, despite May v. Heiner, substantially agreed with this 1884 Pennsylvania Supreme Court interpretation.6 Congress used the “possession or enjoyment” clause in‘death tax legislation in 1862, 1864, and 1898. 12 Stat. 432, 485; 13 Stat. 223, 285; 30 Stat. 448, 464. In referring to the provision in the 1898 Act, this Court said that it made “the liability for taxation depend, not upon the mere vesting in a technical sense of title to the gift, but upon the actual possession or enjoyment thereof.” Vanderbilt v. Eidman, 196 U. S. 480, 493. And five years before the 1916 estate tax statute incorporated the “possession or enjoyment” clause to frustrate estate tax evasions, 39 Stat. 756, 780, this Court had affirmed a judgment of the New York Court of Appeals sustaining the constitutionality of its state inheritance tax in an opinion which said: “It is true that an ingenious mind may devise other means of avoiding an inheritance tax, but the one commonly used is a transfer with reservation of a life estate.” Matter of Keeney, 194 N. Y. 281, 287, 87 N. E. 428, 429; Keeney v. New York, 222 U. S. 525. And see Helvering v. Bullard, 303 U. S. 297, 302, *639where the foregoing quotation was repeated with seeming approval.

From the first estate tax law in 1916 until May v. Heiner, supra, was decided in 1930, trust transfers which were designed to distribute the corpus at the settlor’s death and which reserved a life income to the settlor had always been treated by the Treasury Department as transfers “intended to take effect in possession or enjoyment at . . . his death.” The regulations had so provided and millions of dollars had been collected from taxpayers on this basis. See e. g., T. D. 2910, 21 Treas. Dec. 771 (1919); and see 74 Cong. Rec. 7078, 7198-7199 (March 3, 1931). This principle of estate tax law was so well settled in 1928, that the United States Court of Appeals decided May v. Heiner in favor of the Government in a one-sentence per curiam opinion. 32 F. 2d 1017. Nevertheless, March 2, 1931, this Court followed May v. Heiner in three cases in per curiam opinions, thus upsetting the century-old historic meaning and the long standing Treasury interpretation of the “possession or enjoyment” clause. Burnet v. Northern Trust Co., 283 U. S. 782; Morsman v. Burnet, 283 U. S. 783; McCormick v. Burnet, 283 U.S. 784.

March 3, 1931, the next day after the three per curiam opinions were rendered, Acting Secretary of the Treasury Ogden Mills wrote a letter to the Speaker of the House explaining the holdings in May v. Heiner and the three cases decided the day before. He pointed out the disastrous effects they would have on the estate tax law and urged that Congress “in order to prevent tax evasion,” immediately “correct this situation” brought about by May v. Heiner and the other cases. 74 Cong. Rec. 7198, 7199 (1931). He expressed fear that without such action the Government would suffer “a loss in excess of one-third of the revenue derived from the Federal estate tax, with *640anticipated refunds of in excess of $25,000,000.” The Secretary's surprise at the decisions and his apprehensions as to their tax evasion consequences were repeated on the floor of the House and Senate. 74 Cong. Rec. supra. Senator Smoot, Chairman of the Senate Finance Committee, said on the floor of the Senate that this judicial interpretation of the statute “came almost like a bombshell, because nobody ever anticipated such a decision.” 74 Cong. Rec. 7078. Both houses of Congress unanimously passed and the President signed the requested resolution that same day.7

February 28, 1938, this Court held that neither passage of the resolution nor its later inclusion in the 1932 Revenue Act was intended to apply to trusts created before its passage. Hassett v. Welch, Helvering v. Marshall, 303 U. S. 303. Accordingly, if the corpus of the Church trust executed in 1924 is to be included in the settlor’s estate without this Court’s involvement in the intricacies of state property law, it must be done by virtue of the possession and enjoyment section as it stood without the language added by the joint resolution.

Crucial to the Court’s holding in May v. Heiner was its finding that no interest in the corpus passed at the settlor’s death because legal title had passed from the settlor irrevocably when the trust was executed; for this reason the *641grantor’s reservation of the trust income for his life8— one of the chief bundle-of-ownership interests — -was held not to bring the transfer within the category of transfers “intended to take effect in . . . enjoyment at . . . his death.” This Court had never before so limited the possession or enjoyment section.9 Thus was formal legal title rather than the substance of a transaction made the sole test of taxability under §811 (c). For from the viewpoint of the grantor the significant effect of this transaction was his continued enjoyment and retention of the income until his death; the important consequence to the remaindermen was the postponement of their right to this enjoyment of the income until the grantor’s death.

The effect of the Court’s interpretation of this estate tax section was to permit a person to relieve his estate from the tax by conveying its legal title to trustees whom he selected, with an agreement that they manage the estate during his life, pay to him all income and profits from the property during his life, and deliver it to his *642chosen beneficiaries at death. Preparation of papers to defeat an estate tax thus became an easy chore for one skilled in the “various niceties of the art of conveyancing.” Klein v. United States, 283 U. S. 231, 234. And by this simple method one could, despite the “possession or enjoyment” clause, retain and enjoy all the fruits of his property during life and direct its distribution at death, free from taxes that others less skilled in tax technique would have to pay. Regardless of these facts May v. Heiner held that such an instrument preserving the beneficial use of one’s property during life and providing for its distribution and delivery at death was “not testamentary in character.” May v. Heiner, supra at 243. Cf. Keeney v. New York, supra at 535, 536.

One year after May v. Heiner, this Court decided Klein v. United States, supra. There the grantor made a deed conveying property to his wife for her life with provisions that if she survived him she should “by virtue of this conveyance take, have, and hold the said lands in fee simple,” but the fee was to “remain vested in” him should his wife die first. This Court pointed out that in general and under the law of Illinois where the deed was made, vesting of title in the grantee “depended upon the condition precedent that the death of the grantor happen before that of the grantee.” Thus, since it was found that under Illinois law -legal title to the land had been retained by the husband, it was held that the value of the land should be included in his gross estate under the “possession or enjoyment” section. The Court did not cite May v. Heiner.

In 1935, this Court decided Helvering v. St. Louis Trust Co., 296 U. S. 39, and Becker v. St. Louis Trust Co., 296 U. S. 48. In each of these cases the Court again, as in May v. Heiner, delved into the question of legal title under rather subtle property law concepts and decided that the legal title of the trust properties there, *643unlike the situation in the Klein transfer, had passed irrevocably from the grantor. This passage of bare legal title was held to be enough to render the possession or enjoyment section inapplicable. These cases were expressly overruled by Helvering v. Hallock.

Helvering v. Hallock was decided in 1940. Three separate trusts were considered in the Hallock case. These three trusts as those considered in the St. Louis Trust and Becker cases, had been executed with provisions for reversion of the trust properties to the grantors should the grantors outlive the beneficiaries. The trusts had been executed in 1917, 1919, and 1925. In the Hallock case this Court was again asked to limit the effect of § 811 (c) by emphasis upon the formal passage of legal title. By such concentration on elusive legal title, the Court was invited to lose sight of the plain fact that complete enjoyment had been postponed. We declined to limit the effectiveness of the possession or enjoyment provision of § 811 (c) by attempting to define the nature of the interest which the decedent retained after his inter vivos transfer. We called attention to the snares which inevitably await an attempt to restrict estate tax liability on the “niceties of the art of conveyancing” at p. 117. We declared that the statute now under consideration “taxes not merely those interests which are deemed to pass at death according to refined technicalities of the law of property. It also taxes inter vivos transfers that are too much akin to testamentary dispositions not to be subjected to the same excise,” p. 112, and inter vivos gifts “resorted to, as a substitute for a will, in making dispositions of property operative at death,” p. 114.

As pointed out by the dissent in Hallock, we there directly and unequivocally rejected the only support that could possibly suffice for the holdings in May v. Heiner. That support was the Court’s conclusion in May v. Heiner *644that retention of possession or enjoyment of his property was not enough to require inclusion of its value in the gross estate if a trust grantor had succeeded in passing bare legal title out of himself before death. In Hallock we emphasized our removal of that support by declaring that § 811 (c) “deals with property not technically passing at death but with interests theretofore created. The taxable event is a transfer inter vivos. But the measure of the tax is the value of the transferred property at the time when death brings it into enjoyment,” pp. 110-111.

Moreover, the Hallock case, p. 114, stands plainly for the principle that “In determining whether a taxable transfer becomes complete only at death we look to substance, not to form . . . However we label the device [if] it is but a means by which the gift is rendered incomplete until the donor’s death” the “possession or enjoyment” provision applies.

How is it possible to call this trust transfer “complete” except by invoking a fiction? Church was sole owner of the stocks before the transfer. Probably their greatest property value to Church was his continuing right to get their income. After legal title to the stocks was transferred, somebody still owned a property right in the stock income. That property right did not pass to the trust beneficiaries when the trust was executed; it remained in Church until he died. He made no “complete” gift effective before that date, unless we view the trust transfer as a “complete” gift to the trustees. But Church gave the trustees nothing, either partially or completely. He transferred no right to them to get and spend the stock income. And under the teaching of the Hallock case, quite in contrast to that of May v. Heiner, passage of the mere technical legal title to a trustee is not necessarily crucial in determining whether and when a gift becomes “complete” for estate tax purposes. Looking to substance and not merely to form, *645as we must unless we depart from the teaching of Hallock, the inescapable fact is that Church retained for himself until death a most valuable property right in these stocks — the right to get and to spend their income. Thus Church did far more than attach a “string” to a remotely possible reversionary interest in the property, a sufficient reservation under the Hallock rule to make the value of the corpus subject to an estate tax. Church did not even risk attaching an unbreakable cable to the most valuable property attribute of the stocks, their income. He simply retained this valuable property, the right to the income, for himself until death, when for the first time the stock with all its property attributes “passed” from Church to the trust beneficiaries. Even if the interest of Church was merely “obliterated,” in May v. Heiner language, it is beyond all doubt that simultaneously with his death, Church no longer owned the right to the income; the beneficiaries did. It had then “passed.” It never had before. For the first time, the gift had become “complete.”

Thus, what we said in Hallock was not only a repudiation of the reasoning which was advanced to support the two cases (St. Louis Trust and Becker) that Hallock overruled, but also a complete rejection of the rationale of May v. Heiner on which the two former cases had relied. Hallock thereby returned to the interpretation of the “possession or enjoyment” section under which an estate tax cannot be avoided by any trust transfer except by a bona fide transfer in which the settlor, absolutely, unequivocally, irrevocably, and without possible reservations, parts with all of his title and all of his possession and all of his enjoyment of the transferred property. After such a transfer has been made, the settlor must be left with no present legal title in the property, no possible reversionary interest in that title, and no right to possess or to enjoy the property then or thereafter. *646In other words such a transfer must be immediate and out and out, and must be unaffected by whether the grantor lives or dies. See Shukert v. Allen, 273 U. S. 545, 547; Smith v. Shaughnessy, 318 U. S. 176. We declared this to be the effect of the Hallock case in Goldstone v. United States, 325 U. S. 687, 690, 691. There we said with reference to § 811 (c) in connection with our Hallock ruling: . . It thus sweeps into the gross estate all property the ultimate possession or enjoyment of which is held in suspense until the moment of the decedent’s death or thereafter. . . . Testamentary dispositions of an inter vivos nature cannot escape the force of this section by hiding behind legal niceties contained in devices and forms created by conveyancers.” And see Fidelity-Philadelphia Trust Co. v. Rothensies, supra, and Commissioner v. Estate of Field, 324 U. S. 113.

It is strongly urged that we continue to regard May v. Heiner as controlling and leave its final repudiation to Congress. Little effort is made to defend the May v. Heiner interpretation of “possession or enjoyment” on the ground that it truly reflects the congressional purpose, nor do we think it possible to attribute such a purpose to Congress. There is no persuasive argument, if any at all, that trusts reserving life estates with remainders over at grantors’ deaths are not satisfactory and effective substitutes for wills. In fact, the purpose of this settlor as expressed in his trust papers was to make “provision for any lawful issue” he might “leave at the time of his death as well as provide an income for himself for life.” This paper, labeled a trust, but providing for all the substantial purposes of a will, was intended to and did postpone until the settlor’s death the right of his relatives to possess and enjoy his property. There may be trust instruments that fall more clearly within the class intended to be treated as substitutes for wills by the “possession or enjoyment” clause, but we doubt it.

*647The argument for continuing the error of May v. Heiner is not on the merits but is advanced in the alleged interest of tax stability and certainty, stare decisis and a due deference to the just expectations of those who have relied on the May v. Heiner doctrine. Special stress is laid on Treasury regulations which since the Hassett v. Welch holding in 1938 have accepted the May v. Heiner doctrine and have not provided that the value of a trust corpus must be included in the decedent’s gross estate where a grantor had reserved the trust income. It is even argued that Congress in some way ratified the May v. Heiner doctrine when it passed the joint resolution and that if not, the decision in the Hassett and Marshall cases set at rest all questions as to the soundness of the May v. Heiner interpretation. We find no merit in these contentions.

What was said in the Hallock opinion on the question of stare decisis would appear to be a sufficient answer to that contention here. The Hallock opinion also answers the argument as to recent Treasury regulations, all of which were made by the Treasury under compulsion of this Court’s cases. Furthermore, the history of the struggle of the Treasury to subject such transfers as this to the estate tax law, a history shown in part in the Hassett v. Welch opinion, has served to spotlight the abiding conviction of the Treasury that the May v. Heiner statutory interpretation should be rejected. In view of the struggle of the Treasury in this tax field, the variant judicial and Tax Court opinions, our opinion in the Hallock case and others which followed, it is not easy to believe that taxpayers who executed trusts prior to the 1931 joint resolution felt secure in a belief that May v. Heiner gave them a vested interest in protection from estate taxes under trust transfers such as this one. And so far as this trust is concerned, Treasury regulations required the value of its corpus to be included in the gross estate when it was *648made in 1924, and most of the period from then up to the settlor’s death in 1939.

Moreover, the May v. Heiner doctrine has been repudiated by the Congress and repeatedly challenged by the Treasury. It certainly is not an overstatement to say that this Court’s Hallock opinion and holding treated May v. Heiner with scant respect. We said Congress had “displaced” the May v. Heiner construction of § 811 (c); in overruling the St. Louis Trust cases we pointed out that those cases had relied in part on the “Congressionally discarded May v. Heiner doctrine”; we thought Congress “had in principle already rejected the general attitude underlying” the May v. Heiner and St. Louis Trust cases; and finally our Hallock opinion demolished the only reasoning ever advanced to support the May v. Heiner holding. And in the Hallock case, trusts created in 1917, 1919, and 1925 were held subject to the estate tax under the provisions included in § 811 (c). What we said and did about May v. Heiner in the Hallock case took place in 1940, two years after Hassett v. Welch had held that the 1931 and 1932 amendments could not be applied to trusts created before 1931. Certainly, May v. Heiner cannot be granted the sanctuary of stare decisis on the ground that it has had a long and tranquil history free from troubles and challenges.

Nor does the joint resolution or the opinion in the Hassett v. Welch and Helvering v. Marshall cases, decided together, support an argument that the May v. Heiner doctrine be left undisturbed. It would be impossible to say that Congress in 1931 intended to accept and ratify decisions that hit the Congress like a “bombshell.”10 *649And in Hassett v. Welch the Government did not ask this Court to reexamine or overrule May v. Heiner or the three per curiam cases that relied on May v. Heiner. In fact, the government brief argued that May v. Heiner on its facts was distinguishable from Hassett v. Welch. The government brief also pointedly insisted that its position in Hassett v. Welch did “not require a reexamination of the three per curiam decisions of March 2, 1931.” It was the Government’s sole contention in the Hassett and Marshall cases that the 1932 reenactment of the joint resolution was not limited in application to trusts thereafter created, but was intended to make the new 1932 amendment applicable to past trust agreements. That contention was rejected. The holding was limited to that single question.

The plain implications of the Hallock opinion recognize that the Hassett and Marshall cases did not reaffirm the May v. Heiner doctrine. In the Marshall case the trust, created in 1920, contained a provision that should the settlor outlive the trust beneficiary, the trust corpus would revert to the settlor. That is the very type of provision which we held in Hallock would require inclusion of its value in the settlor’s estate. Since the Hallock case did not overrule the Marshall case involving a trust created in 1920, it must have accepted the Marshall and Hassett cases as deciding no more than that the value of the trust properties there could not be included in the de*650cedent’s gross estate where the Government’s sole reliance was on a retroactive application of the 1931 and 1932 amendments to the estate tax law.

That the Hallock opinion did not treat the Hassett and Marshall cases as having reaffirmed this Court’s interpretation of the pre-1931 possession or enjoyment clause is further emphasized by the effect of the Hallock case on the type of trust in McCormick v. Burnet, 283 U. S. 784, a trust created before 1931. The United States Court of Appeals in that case had held that the trust property should be included in the decedent’s estate chiefly because of the trust provision that the corpus should revert to the settlor in the event that she outlived her three children. 43 F. 2d 277. This Court in its per curiam opinion reversed the Court of Appeals and held that the McCormick corpus need not be included in the decedent’s estate. Our Hallock case held directly the contrary, for since Hallock, the McCormick corpus would have to be taxed under the pre-1931 language of §811 (c). In so interpreting the pre-1931 language in the Hallock case, we necessarily rejected the contention made there that the Congress by passage of the resolution and this Court by the Hassett and Marshall opinions had accepted as correct the May v. Heiner restrictive interpretation of §811 (c). It is plain that this Court in the Hallock case considered that the Has-sett and Marshall cases held no more than that the 1931 and 1932 amendments were prospective, and that neither the congressional resolution nor the Hassett and Marshall cases were designed to give new life and vigor to the May v. Heiner doctrine.11

*651The reliance of respondent here on the Hassett and Marshall cases is misplaced. We hold that this trust agreement, because it reserved a life income in the trust property, was intended to take effect in possession or enjoyment at the settlor’s death and that the Commissioner therefore properly included the value of its corpus in the estate.

Reversed.

concurring in No. 3, Spiegel v. Commissioner, post, p. 701, and dissenting in No. 5, Commissioner v. Church, ante, p. 632.

As these tax decisions may have an influence on subsequent decisions beyond the limited area of the issues decided, I have thought it advisable to state my position for whatever light it may throw. I agree with the judg*652ment directed by the Court in Spiegel v. Commissioner and with so much of the opinion as rests solely upon the controlling effect of the possibility of reverter under the law of Illinois. As I disagree with Church v. Commissioner, decided today, I cannot accept so much of the opinion in the Spiegel case, p. 705, as seems to put reliance upon the fact that the settlor as trustee retains any “possession or enjoyment” of the trust, other than a possibility of reverter. I am opposed to the view expressed in the dissent written by Mr. Justice Burton that the settlor’s intent rather than the effect of his acts is the touchstone to determine the taxability of his property for estate tax purposes.

So far as Commissioner v. Church is concerned, I do not believe that May v. Heiner, 281 U. S. 238, should be overruled. The Joint Resolution of March 3, 1931, therefore, stands as the determinative factor in reaching a conclusion as to the taxability of the Church estate. Hassett v. Welch, 303 U. S. 303, decided that the Resolution was not retroactive. Consequently, the Church estate is not subject to an estate tax because of the reservation of a life estate.

We are asked to accept an overruling of May v. Heiner, supra, and also, I think, of Reinecke v. Northern Trust Co., 278 U. S. 339, not to mention the incidental fall of Hassett v. Welch, supra, on the one side, or, on the other hand, to limit the rule as to the possibility of reverter in Helvering v. Hallock, 309 U. S. 106, and the numerous cases that follow its teaching, to reverters expressly reserved in the documents. Legislation indicates a purpose to promote gifts as a desirable means for early distribution of property benefits. In reliance upon a long-settled course of legislative and judicial construction, donors have made property arrangements that should not now be upset summarily with no stronger reasons for doing so than that former courts *653and the Congress did not interpret the legislation in the same way as this Court now does. Judicial efforts to mold tax policy by isolated decisions make a national tax system difficult to develop, administer or observe. For more than thirty years Congress has legislated upon this problem and this Court has interpreted the enactments so that now what seems to me a reasonably fair interpretation of tax liability under § 811 (c) of the Internal Revenue Code, as now written, has been worked out. Relying upon the desirability of stare decisis under the decisions concerning § 811 (c), I would leave such changes as may seem desirable to the Congress, where general authority for that purpose rests.

(1) A provision including in a decedent’s estate the value at time of death of interest in any transfer by trust “in contemplation of or intended to take effect in possession or enjoyment at or after his death” has been in the federal estate tax law since the Income Tax Act of 1916.1 It will be noted that the phrase relating to a transfer “in contemplation of or intended to take effect in possession or enjoyment at or after his [settlor’s] death” has not changed. It was construed by this Court, at first, to apply to those circumstances where something passed *654from the “possession, enjoyment or control of the donor at his death.” Reinecke v. Northern Trust Co., 278 U. S. 339, 348. “Of course it was not argued that every vested interest that manifestly would take effect in actual enjoyment after the grantor’s death was within the statute.” Shukert v. Allen, 273 U. S. 545, 547. When, after the execution of a trust, the settlor “held no right in the trust estate which in any sense was the subject of testamentary disposition,” this Court was of the opinion that the gift was not intended to take effect in possession or enjoyment at the donor’s death. Helvering v. St. Louis Union *655Tr. Co., 296 U. S. 39, 43; Helvering v. City Bank Farmers Trust Co., 296 U. S. 85, 88; Burnet v. Northern Trust Co., 283 U. S. 782; Morsman v. Burnet, 283 U. S. 783; McCormick v. Burnet, 283 U. S. 784; May v. Heiner, 281 U. S. 238. A reserved power of appointment or change is, in a sense, a testamentary power over the corpus. Reinecke v. Northern Trust Co., supra, at 345; Porter v. Commissioner, 288 U. S. 436.

Klein v. United States, 283 U. S. 231, brought doubt into the above conception of the meaning of the phrase in question. That trust was to A for life and on condition that A survive the donor to A in fee simple. It was the death of the donor that “brought the larger estate into being . . . and effected its transmission from the dead to the living,” this Court said in upholding the tax on the trust property. This was construed by four members of the Court to mean that the donor’s death “operating upon his gift inter vivos not complete until his death, is the event which calls the statute into operation.” Mr. Justice Stone, dissenting in the later case of Helvering v. St. Louis Trust Co., supra, 46. The two positions, one that only power in the settlor at the time of death to cause the property to be transferred from him to another by will or by descent or to select beneficiaries through appointment brought the property formerly transferred within the reach of the words “intended to take effect in possession or enjoyment at or after his death,” the Reinecke concept, and the other that, in addition, every possibility of reversion of the transferred interest to the settlor must be barred by the trust instrument, the dissenter’s ground in Helvering v. St. Louis Trust Co., were fully discussed in the majority and dissenting opinions in Helvering v. Hallock, 309 U. S. 106.2 The latter *656position was accepted as the sound interpretation by us and I adhere to that view for the reasons stated in the Court’s opinion in Helvering v. Hallock. Cf. Eisenstein, Estate Taxes and the Higher Learning of the Supreme Court, 3 Tax Law Rev. 395. That interpretation has gained strength from the fact that Congress has not repudiated it as inconsistent with the legislative purpose and by other judgments by this Court applying the principles of the Hallock case in accordance with this statement. Fidelity-Philadelphia Trust Co. v. Rothensies, 324 U. S. 108; Commissioner v. Estate of Field, 324 U. S. 113. Possession or enjoyment of property as heretofore applied has meant from the standpoint of the taxability of the transferor’s estate, at least, that the death of the transferor perfects the right of the transferee and cuts off any possibility of reverter to the transferor left by the instruments of transfer. If the transferor *657reacquired the property by inheritance or by purchase, other factors would enter. Before the Joint Resolution even the reservation of a life estate was insufficient to preserve possession or enjoyment in the transferor as nothing passed at his death. When words such as “possession or enjoyment” used in a section of a revenue statute with their many possible shades and ambiguities of meaning have been given definition through the course of legislation and litigation, a change by courts should be avoided.3 By the Resolution such a reservation or that of power of appointment was also made the source of an estate tax.

Prior cases have involved trust instruments where the settlor specifically reserved remainders, reverters or contingent powers of appointment. In these cases the value at death of the entire corpus of the trusts was taxed. This was because in each case there was a contingency through which completed gifts of the entire corpus to the beneficiaries might fail before the death of the settlor with the result that the settlor would again control the transfer of the corpus.'4 In such circumstances, I take it as settled that the property is taxable on the event of the settlor’s death under §§ 810 and 811 (c). Cf. 324 U. S. at 111.

The trust instruments in the present cases of the Spiegel and Church estates do not specifically provide for such possibility of reverter or for regaining control of the devolution of the property. The issue raised by these cases is whether a like possibility of reverter springing not from the instrument but by operation of law through the failure of all beneficiaries named in the trust instru*658ment shall have the same effect. All named beneficiaries in these two trusts might die before the settlors without surviving issue. Thus, depending upon the controlling state law, the settlors might repossess the estates.5

To lay bare the heart of the problem, it seems helpful to put aside certain phases of possible congressional intention and possible statutory meaning, as not involved or heretofore decided for sound reasons.

A. It was not the purpose of Congress at any time in dealing with the inclusion of transfers of property in trust to have the whole value, at the donor’s death, of the total of all gifts made during life, included in the settlor’s *659estate for estate tax purposes.6 The words of the statute show this. See note 1, supra. Gifts in trust are taxable only where an interest remains in the donor. Therefore a gift by A to a trust company to hold in trust for B during B’s life and at B’s death to C, his heirs, devisees or assigns is not taxable under §811 (c). Reinecke v. Northern Trust Co., supra, 347-48. Before the amendment of 19317 the retention of an estate for life in the settlor did not subject the trust to estate tax where the remainder was taken by beneficiaries without regard to future action by the settlor.8

B. The Joint Resolution of 1931 made no change in the language of the subsection of the estate tax relating to the inclusion in estates of interests in trusts intended to take effect in possession or enjoyment at or after death. Neither the resolution nor the discussion on the floor of either house suggested a change in the words of the section to define what is meant by an interest intended to take effect after death. Congress aimed at the retention of life interests, not at this Court’s determinations of the meaning of “possession or enjoyment.” Those words were left untouched and an addition was made providing for the inclusion in the estate of interests where the settlor had retained the possession or enjoyment of the property or a right to income or the power to designate the beneficiaries. See note 1, supra. Therefore the words relating to intention, death, possession or enjoyment have the same meaning now as they did *660before the 1931 amendment was adopted.9 The doctrine of May v. Heiner that the statute, as written when that case was, handed down, did not cover reservations of life interests and powers of designation was legislatively changed by adding the words of the Joint Resolution. See in accord Helvering v. Hallock, supra. When Hallock there refers to the doctrine of May v. Heiner discarded by Congress, it is the doctrine of May v. Heiner that a settlor might reserve a life interest that was meant. Hallock did not say or imply, as I read it, that the May v. Heiner doctrine, which is supported by Reinecke and Shukert v. Allen, as to when “possession or enjoyment” passes from a donor was changed by the Resolution. These cases had held that something must pass from the settlor. The only difference wrought by Hallock on this concept of possession and enjoyment was to apply the Klein rule that the enlargement of the remainder estate did effect a transmission from the dead to the living.

Assuming that Congress might have legislated so that the added words would apply to the estates of all who died after the passage of the Joint Resolution, Congress definitely manifested an intention that the amendments were not to apply to trusts created prior to the Resolution though the settlor might die subsequently thereto. This whole matter is discussed thoroughly and, I think, unanswerably in Hassett v. Welch, 303 U. S. 303, and I can add nothing to the argument. Attention, however, *661should be called to the statements on the floor of the House by members of the Committee on Ways and Means at the time of the passage of the Joint Resolution.10 Mr. Hawley, Chairman of the Committee, answering a question as to the nature of the Resolution said, “It provides that hereafter no such method shall be used to evade the tax.”

Mr. Garner of the same Committee stated:

“The Committee on Ways and Means this afternoon had a meeting and unanimously reported the resolution just passed. We did not make it retroactive for the reason that we were afraid that the Senate would not agree to it. But I do hope that when this matter is considered in the Seventy-second Congress we may be able to pass a bill that will make it retroactive.”

And in answer to a question, he reiterated, “I have strong hopes that the next Congress will make it retroactive.” Congress never took any subsequent action and this Court’s interpretation of the meaning of “intended to take effect in possession or enjoyment” remained the same. The addition to the section made by the Joint Resolution made certain future gifts inter vivos, which would theretofore have been free of estate tax, subject to such a tax.

C. As a corollary to the foregoing section B, it is clear to me also that Congress by the Joint Resolution made no change in the statute for the purpose of bringing trusts into an estate merely because the actual use of the estate or its income by the cestui que trust was postponed until the death of the donor. Shukert v. Allen, supra.

D. It is impossible for me to look upon the Spiegel or Church trust as closely akin to a will. The decisive *662difference is that a will may be changed at any time during life, while these trusts obliterated any power in the settlors to change or modify the devolution. Only the chances of death, wholly beyond their control, might put the disposition again in their hands. Further, during life the settlors must handle the trusts for the benefit of all beneficiaries. They were not free to do as they pleased as would have been the case of a will. Of course, if the settlor had made similar provision for the objects of his bounty by will, in effect at death, the result to the takers would have been the same; or if, in the Spiegel case, the father had annually given his children the same sums that the trust earned, their economic position would have been the same for that year but the children could not look forward with certainty to their annual income from the trust. Without the trust, the beneficiaries’ income would have been subject to the wish of the settlor. It needs no argument or illustration to show that a father’s gift from his income is a very different thing from an irrevocable gift of principal to a child.

Returning to the issue in these present cases, the difference between them and Helvering v. Hallock and its progeny is that here the possibility of reverter arises by operation of law whereas in them the possibility arises out of the terms of the trust. That difference I do not think is material as to taxability under § 810 and § 811 (c). Granting that in early interpretations of the sections this Court might logically have determined that remote possibilities of reverter did not interfere with the beneficiaries’ complete possession and enjoyment of the gift during the lifetime of the donor, the balance of experience and precedent, since Helvering v. Hallock, tips the scale the other way in my judgment. It is important, though not decisive, since we are not justified in pushing every rule to its logical extreme, that this conclusion is a *663logical outgrowth of the Hallock rule. Since we know it is the purpose of Congress to put an estate tax on gifts intended to take effect at or after death, the interpretation of those words should be broad enough to accomplish the purpose effectually. “Intended to take effect,” in that view, has for me the meaning of an intention to abide by the legal result of the terms of the trust.

I recognize that this interpretation has possibilities of variation in result through the employment of technicalities of property law. The addition of a phrase may make the difference between a completed or an incom-pleted gift. To make the intention of the settlor the determinative factor creates equal difficulties. Nor am I unmindful of this Court’s effort, in which I joined, in the Hallock case to find a harmonizing principle for the difficulties engendered by § 811 (c). In that case the principle applied was that a tax lies against an estate when the death of the grantor brings a larger estate into being for the beneficiary. This does accomplish uniformity in the interpretation of the section of federal law. Hallock attempted nothing more. It leaves its application to particular trusts dependent upon state determination of when a settlor has divested himself of every possible interest in the res of a trust.11 We are *664dealing with a statute and Congress is fully competent to correct any misunderstanding we may have of the congressional intention.

(2) The foregoing leads to the conclusion in the Spiegel case that this estate must, pay a federal estate tax on the trust res unless that res, under the law of Illinois, would have passed to the heirs at law or the legatees of the last descendant of the settlor. If under Illinois law the estate returned to the settlor on his surviving all his descendants, the tax is due. The possibilities of this happening in this case are extremely remote but a trust might have been created by a young son for an aged mother to pay her the income for life and at the settlor’s death to pay her the principal.

The Court of Appeals concluded (159 F. 2d at 259) that “If none of the beneficiaries survived the settlor, and that was a possibility, then the trust failed, and the trustees would hold the bare naked title to the corpus as resulting trustees for the settlor.” There is no Illinois case holding squarely on this point, and in the absence of such a determination by a state court we do not interfere with a reasonable decision of the circuit which embraces Illinois. Helvering v. Stuart, 317 U. S. 154, 164; MacGregor v. State Mutual Life Assurance Co., 315 U. S. 280. The rule followed by the Court of Appeals accords with that generally accepted. Restatement, Trusts § 411; 3 Scott, Trusts §411; 2 Bogert, Trusts and Trustees § 468; Harris Trust & Savings Bank v. Morse, 238 Ill. App. 232; Lill v. Brant, 6 Ill. App. 366, 376.12

*665The taxpayer relies upon cases wherein the language of wills was construed in order to create vested remainders. These cases, however, do not overturn the firmly settled principle that where an express trust fails for lack of a beneficiary, a resulting trust in favor of the settlor arises by operation of law.13 To vest property under a will or deed is desirable. To vest property under a trust may not be. It is more reasonable to return trust property to the settlor on failure of the trust than to have it go to the heirs of the beneficiary.

From a reading of the trust instrument involved in the instant case, it is manifest that the settlor did not intend to grant his children the power to dispose of their respective shares should they predecease the settlor with*666out issue. The settlor specifically named as beneficiaries of the trust his children and grandchildren. That he intended to restrict the trust to these two classes of beneficiaries is evidenced by the provision of the instrument that in the event of the death of a child without issue that child’s share was to be added to the shares of the settlor’s surviving children. His retention of the trusteeship and failure to grant the power of disposition to his children in his lifetime negative any intention of the settlor to exclude the possibility of a reversion of the trust property to himself.

No error appears in the conclusion of the Court of Appeals on this point.

(3) Finally, the situation in the Church case must be dealt with. The trust was created in New York by a resident of New York who died a resident of New Jersey. Two of three trustees were at all times residents of New York where the stocks and accounts of the trust were kept. From what is before me, I would assume that the New York law would control as to the possibility of the retention of an interest by the settlor. This produces a variant from the Spiegel case. The determination of New York law will be made by a circuit that does not include that state. This, I think, is not significant in determining the course to be followed.

As the Court of Appeals for the Third Circuit made its decision on the authority of the Dobson rule, 161 F. 2d 11, it did not consider the effect of Hassett v. Welch, 303 U. S. 303. As May v. Heiner stands, in my opinion, trusts, like the Church trust, created prior to the passage of the Joint Resolution of March 3, 1931, are not includable in the gross estate of a settlor for federal estate tax purposes unless there is a possibility of reverter to the settlor by operation of the controlling state law. To determine this question, I would vacate the judgment *667of the Third Circuit and remand the case to that court to determine the state law.

I would affirm No. 3, Spiegel v. Commissioner; I would vacate No. 5, Commissioner v. Church.

dissenting.*

By fitting together my agreement with portions of the dissenting concurrence and my disagreement with a part of the comprehensive dissenting opinions of my brother Burton, I could indicate, substantially, my views of these cases. But such piecing together would make a Joseph’s coat. Therefore, even at the risk of some repetition of what has been said by others, a self-contained statement on the basic issues of these cases will make for clarity. Particularly is this desirable where disharmony of views supports a common result — a result the upsetting of which by Congress is almost invited.

I.

In the Spiegel case, No. 3, the decedent made a settlement by the terms of which he reserved no interest for himself, and it is not suggested that the form of the settlement disguised an attempted evasion of the estate-tax law. The corpus of the decedent’s estate is found to be subject to the estate tax on the basis of Helvering v. Hallock, 309 U. S. 106, as supplemented by Fidelity-Philadelphia Trust Co. v. Rothensies, 324 U. S. 108, Commissioner v. Estate of Field, 324 U. S. 113, and Goldstone v. United States, 325 U. S. 687. On that basis it is now decided that if there is a possibility, due to the terms of the instrument or by operation of law, however remote, that settled property may return to the set-*668tlor, the entire trust property must be included in the gross estate for purposes of the federal estate tax. Thus, under the Court's decision tax liability may be incurred by the discovery of a gossamer thread of possession or enjoyment, which has no value. Nevertheless the entire trust corpus is included in the gross estate and taxed as if the settlor really had possession or enjoyment of the property. Such a result not only creates unanticipated hardship for taxpayers; it is also an unrealistic interpretation of § 811 (c) of the Internal Revenue Code. Since such an unrealistic interpretation is not a judicial duty whereas its avoidance is, I am compelled to conclude that Spiegel did not transfer an interest in property “intended to take effect in possession or enjoyment at or after death” within the meaning of § 811 (c) and that the trust corpus settled by him in his lifetime was no part of his gross estate.

This case is brought under the decisions of Hallock and the three subsequent cases only by a disregard of the vital differences between the interest created by the Spiegel indenture and the arrangements before this Court in the four cases upon which reliance is placed.

1. In 1920, Spiegel transferred securities to himself and another person as co-trustees, the income to be paid equally to Spiegel’s three named children during his lifetime. If any of the children died before the settlor, the share of that child was to go to his issue, if any, otherwise to the settlor’s other children. The instrument provided further that upon the settlor’s death the corpus, together with any accumulated income, should be divided “equally among my said three (3) children, and if any of my said children shall have died, leaving any child or children surviving, then the child or children of such deceased child of mine shall receive the share” of the trust to which his or her parent would have been entitled. *669If any of the settlor’s three children died without leaving surviving children, that share was to go to the two remaining children. When the trust was established Spiegel was 47 years old, and his three children were aged 25, 15, and 13. At his death twenty years later the children were still living and there were three grandchildren. Upon the assumption that there would have been a re-verter to Spiegel by operation of Illinois law in the event that all his children predeceased him without leaving “surviving children,” the value of this remote contingency was determined mathematically to be worth $4,000.1

2. In the Helvering v. Hallock series, supra, each of the several donors created a trust giving an estate to another but providing that the property would revert to the donor if the donee predeceased him. The donor’s death in each case was the operative fact which established final and complete dominion as between the donor and the donee according to the terms of the instruments. Until the former’s death the donor was, as it were, competing with the donee for the ultimate use and enjoyment of the property. We there held that the particular form of conveyancing words is immaterial if the net effect is that transferred property will revest in a donor who survives the donee. Except on a contingency of Illinois law so remote as to be nonexistent in the practical affairs of life, the property would never revert to Spiegel. His death no doubt would finally determine which children or grandchildren would have the ultimate enjoyment of the trust corpus settled upon his children, but in the real world the property could never come back to him as a windfall. His death did not determine contingencies *670from which he could benefit. His death merely definitively closed the class of beneficiaries and fixed the quantum of each child’s share.

Contrary to the suggestion in the concurring opinion in this case — a suggestion accepted by the majority opinion — the Court of Appeals did not find that Spiegel retained an interest because he had not provided for all contingencies. It included the settled property in the gross estate on the theory that every trust carries as it were the seed of its own destruction through failure of the trust, thereby generating a resulting trust. It said, “If none of the beneficiaries survived the settlor, and that was a possibility, then the trust failed, and the trustees would hold the bare naked title to the corpus as resulting trustees for the settlor.” 159 F. 2d at 259. But this mode of argument would have swept into the gross estate a conveyance in trust in fee to any of Spiegel’s children in 1920 since the failure of the trust for any conceivable reason presumably would not turn the trust property into an outright gift to the trustees.

The trust indenture is a comprehensive arrangement for the children and their offspring to take care of the contingencies of mortality among the children and their offspring. Provisions such as were made in the Spiegel case are precisely the kind of arrangement made by an ancestor for his children and children’s children by which he settles property upon them with a view to the contingencies of successive generations and reserves no interest in himself. Nothing was reserved in the settlor except what feudal notions about seisin may have reserved. But feudal notions of seisin are no more pertinent in tax cases when they lead to imposition of an estate tax than when they lead away from it. At the very basis of the decision of the Hallock case was the insistence that these “unwitty diversities of the law of property derive [d] from medieval concepts as to the necessity of a continuous *671seisin. . . . are peculiarly irrelevant in the application of tax measures now so largely directed toward intangible wealth.” Helvering v. Hallock, supra at p. 118. The metaphysical remoteness of the present settlor’s interest at the time the trust was created is clearly shown by the fact that it depended upon the highly unlikely event that all the children in existence at the time of the conveyance would die and would die childless. Even this remote possibility evaporated long before the settlor died. And certainly the only tenable construction of the statute is that not only must there have been a transfer of the sort designated in §811 (c) but the settlor’s interest must also persist up to the time of his death. Cf. Estate of Miller, 40 B. T. A. 138; see Griswold, Cases and Materials on Federal Taxation 145 (1940).

3. The three later decisions invoked by the Court bear no resemblance to the situation presented by the Spiegel case and give no justification for the ruling now made. In Fidelity-Philadelphia Trust Co. v. Rothensies, supra, the settlement provided for a life estate in the settlor, life estates in the two daughters, and a reversion in the settlor unless the daughters had issue. See Brief for Respondent, p. 8, Fidelity-Philadelphia Trust Co. v. Rothensies, supra; Goldstone v. United States, 325 U. S. 687, 693, n. 3. The birth of the grandchildren which cut off the settlor’s interest did not occur until after the death of the settlor. Since, therefore, the taxability is to be determined at death, it followed that the value of the trust property was to be included in the gross estate. The sole controversy was whether deduction should be allowed for the mother’s and daughters’ life interests and for a contingent gift to unborn children.2 Likewise in the Estate of Field case it was conceded that the settlor retained until death a substantial interest — the right to *672reduce or cancel the interest of life tenants and a reversion of the corpus to himself if he survived these tenants. In the Estate of Field case too the controversy concerned the basis on which the estate was to be assessed — whether the value of the life tenancies was to be deducted from the corpus. The Goldstone case was in effect another Hallock case, the insurance being payable upon the donor’s death to the wife but with a reserved right in the donor if she predeceased him.

The birth of grandchildren in Spiegel’s lifetime destroys all resemblance between his case and the cases just discussed. On the least favorable reading of the trust instrument — whereby the grandchildren would have to survive not only their parents but also the settlor — the possibility that the settlor would regain the property was extremely tenuous. Reading the trust instrument in a customary and not in a hostile spirit, the grandchildren would merely have to survive their parents and not the settlor for their interest to become indefeasible. Thus the remote contingency of reacquisition by the settlor vanishes.3

To be sure, in both the Fidelity-Philadelphia Trust Co. and the Estate of Field cases there is generality of *673language about indifference regarding the remoteness or uncertainty of the decedent’s “reversionary interest.” But in both cases as we have seen there was no question that the trust instrument itself purposely reserved in the settlor an interest which in its context was substantial. The talk of uncertainty and remoteness was merely a way of indicating that where the settlor himself had reserved an interest terminable only by his death, it was not for the law to make nice calculations as to the chance he was giving himself to regain the property. In these two cases the settlor thought the reserved interest had significance and of course the law gave that significance monetary value. Spiegel contrariwise designed to retain nothing and his estate should not be held to include property of which he divested himself many years before his death.

4. But even the gossamer thread which binds the majority together in subjecting the Spiegel trust corpus to an estate tax is visible only to their mind’s eye. The gossamer thread is the remote possibility that at the time of Spiegel’s death there would be a reverter of the trust property to him. But that possibility depends entirely upon its recognition by the law of Illinois. It is at best a dubious assumption that such a reverter exists under Illinois law. My brother Burton’s argument in disproof is not lightly to be dismissed. At best, however, this Court’s guess that Illinois law would enforce such a reverter may be displaced the day after tomorrow by the Illinois Supreme Court’s authoritative rejection of the guess. If tax liability is to hang by a gossamer thread, the Court ought to be sure that the thread is there. Since only the courts of Illinois can definitively *674inform us about this, it would seem to me common sense to secure an adjudication from them if some appropriate procedure of Illinois, like the Declaratory Judgment Act, is available.4 To justify at all the Court’s theory, the rational mode of disposing of the case would be to remand it to the Court of Appeals for the Seventh Circuit in order to allow that court to decide whether in fact a procedure is available under Illinois law for a ruling upon the point of Illinois law which is made the basis of this Court’s decision, since the correctness of this Court’s assumption is at best doubtful. Cf. Thompson v. Magnolia Petroleum Co., 309 U. S. 478, 483-484; Spector Motor Service, Inc. v. McLaughlin, 323 U. S. 101. A determination so made would conclusively fix the interests actually held by the parties to the instrument and at the same time leave to the federal courts the tax consequences of these interests. Blair v. Commissioner, 300 U. S. 5, 9-14; Freuler v. Helvering, 291 U. S. 35.

II.

The reach of the Church case, No. 5, extends far beyond the proper construction of the tax statute.5 It concerns the appropriate attitude of this Court toward a series of long-standing unanimous decisions by this Court. More than that, it involves the respect owed by this Court to the expressed intention of Congress.

*675The short of the matter is this. More than eighteen years ago this Court by a unanimous ruling found that Congress did not mean to subject a trust corpus transferred by a decedent in his lifetime to the estate tax imposed by the Revenue Act of 1918 merely because the settlor had reserved the income to himself for life. May v. Heiner, 281 U. S. 238. At the earliest opportunity, in three cases having minor variations but presenting the same issue, the Treasury invited the Court’s reconsideration of its decision. But the Court, after having had the benefit of comprehensive briefs and arguments by counsel specially competent in fiscal matters, unanimously adhered to its ruling in May v. Heiner. Burnet v. Northern Trust Co., 283 U. S. 782; Morsman v. Burnet, 283 U. S. 783; McCormick v. Burnet, 283 U. S. 784. These decisions, now cast aside, were shared in by judges of whom it must be said without invidiousness that they were most alert in recognizing the public interest and resourceful in protecting it. There were brave men before Agamemnon. If such a series of decisions, viewed in all their circumstances, as that which established the rule in May v. Heiner, is to have only contemporaneous value, the wisest decisions of the present Court are assured no greater permanence.

In fairness, attention should be called to the fact that in joining the Court’s decisions laying down, and adhering to, the May v. Heiner ruling, Mr. Chief Justice Hughes, Mr. Justice Holmes, Mr. Justice Brandéis, and Mr. Justice Stone were not denied argument which the Government has now urged upon us. But it is also fair to the Government to point out that it has not of its own accord asked this Court to overrule the four decisions rendered eighteen years ago. It was only after the case was ordered for reargument and a series of questions was formulated by the Court which shed doubt upon the continued vitality of May v. Heiner, that the Government *676suggested that the decision be cast into limbo. 68 Sup. Ct. 1524. No doubt stare decisis is not “a universal, inexorable command.” Brandeis, J., dissenting in Washington v. Dawson & Co., 264 U. S. 219, 238. But neither is it a doctrine of the dead hand. In the very Hallock case relied upon so heavily in these cases the Court said, “We recognize that stare decisis embodies an important social policy. It represents an element of continuity in law, and is rooted in the psychologic need to satisfy reasonable expectations.” 309 U. S. at 119. And one of the most recent reliances on stare decisis for decision was expressed with such firmness as to manifest allegiance to principle, not utilization of an ad hoc argument.6 We are not dealing here with a ruling which cramps the power of Government; we are not dealing with a constitutional adjudication which time and experience have proved a parochial instead of a spacious view of the Constitution and which thus calls for self-correction by the Court without waiting *677for the leaden-footed process of constitutional amendment. We are dealing with an exercise of this Court’s duty to construe what Congress has enacted with ample powers on its part quickly and completely to correct misconstruction.

Those powers were promptly invoked in this case. Because the Treasury was dissatisfied with the meaning given by this Court to the estate-tax provision, the very next day after the three decisions reaffirming May v. Heiner were handed down, the Treasury appealed to Congress for relief and Congress gave relief. The true significance of today’s decision in the Church case is not to be found in the Court’s failure to respect stare decisis. The extent to which judges should feel in duty bound not to innovate is a perennial problem, and the pull of the past is different among different judges as it is in the same judge about different aspects of the past. We are obligated, however, to enforce what is within the power of Congress to declare. Inevitable difficulties arise when Congress has not made clear its purpose, but when that purpose is made manifest in a manner that leaves no doubt according to the ordinary meaning of English speech, this Court, in disregarding it, is disregarding the limits of the judicial function which we all profess to observe.

The Treasury no doubt was deeply concerned over the emphatic reaffirmation of May v. Heiner. The relief sought from Congress was formulated by the fiscal and legal expert who had that very day failed in persuading this Court to overrule May v. Heiner. What relief did the Treasury seek from Congress? Did the Secretary of the Treasury ask Congress to rewrite § 302 (c) of the Revenue Act of 1926, now §811 (c) of the Internal Revenue Code, so as to sterilize May v. Heiner ? Certainly not. Not one word was altered of the language of the provision which this Court felt compelled to construe *678as it did in May v. Reiner. What the Treasury proposed and what Congress granted was a qualifying addition to the statute as construed in May v. Heiner whereby trust settlements reserving a life interest in the settlor were to be included in a decedent’s gross estate, but only in the case of settlements made after this qualification became operative, that is, after March 3, 1931. Such, in the light of the legislative history, was the inescapable meaning of what Congress did, and the only thing it did, to qualify the reading which this Court four times felt constrained to place upon the mandate of Congress in the imposition of the estate tax. The history is recounted in Hassett v. Welch, 303 U. S. 303, again without a dissenting voice. This history is so crucial to the exercise of the judicial process in this case, that it bears repetition.

When the Joint Resolution of March 3, 1931, was adopted, it was clear that it was to be only of prospective effect. Its sponsors specifically declared:

“Entirely apart from the refunds that may be expected to result, it is to be anticipated that many persons will proceed to execute trusts or other varieties of transfers under which they will be enabled to escape the estate tax upon their property. It is of the greatest importance therefore that this situation be corrected and that this obvious opportunity for tax avoidance be removed. It is for that purpose that the joint resolution is proposed.” 74 Cong. Rec. 7198 and 7078.

And there was good reason for not making it retroactive:

“We did not make it retroactive for the reason that we were afraid that the Senate would not agree to it. But I do hope that when this matter is considered in the Seventy-second Congress we may be able to pass a bill that will make it retroactive.” 74 Cong. Rec. 7199.

*679These statements on the floor by those in charge of the Resolution are controlling, as much as though they had been submitted in a Committee Report, for they were the authoritative explanation of the Resolution’s purpose and meaning. In fact, Representative Schafer of Wisconsin had stated that unless the sponsors explained the bill he would object, thus preventing its acceptance as a resolution. 74 Cong. Rec. 7198.

When the section was reenacted by the 72d Congress as § 803 (a) of the Revenue Act of 1932, it remained in the pre-lkfay v. Heiner language with the Joint Resolution of March 3, 1931, added in slightly different phrasing. 47 Stat. 279. This section was interpreted in 1938 by a unanimous Court as not applying to a reserved life estate created in 1924. Hassett v. Welch, 303 U. S. 303. The briefs filed by the Government in that case again contained much of the same data now found to demand a contrary result.7 On the same day this Court also decided Helvering v. Bullard, 303 U. S. 297, which held the Joint Resolution applicable to reserved life estates created after the passage of the Resolution. It quoted the same language from Matter of Keeney, 194 N. Y. 281, 287, now quoted by the majority, thus indicating that it appreciated the tax-avoidance problem and would have interpreted § 803 (a) retroactively had Congress indicated that it intended to tax reserved life estates created before March 3, 1931.8 It *680is especially difficult to say that in Hassett v. Welch, supra, the Court considered only the language added by the Joint Resolution and not the section in its entirety, since it phrased the issue before it in this' way:

“The petitioners ask us to hold that § 302 (c) of the Revenue Act of 1926 as amended by the Joint Resolution of Congress of March 3, 1931, and § 803 (a) of the Revenue Act of 1932, includes in the gross estate of a decedent, for estate tax, property which, before the adoption of the amendments, was irrevocably transferred with reservation of a life estate to the transferor . . . .” 303 U. S. at 304.

If May v. Heiner had not been accepted as authoritative, it would have been pointless to decide that the amendment to § 302 (c) of the Revenue Act of 1926 did not operate retroactively. See Learned Hand, J., in Helvering v. Proctor, 140 F. 2d 87, 89 (C. A. 2d Cir.).

Of course the Government did not attack May v. Heiner in Hassett v. Welch, supra. Having been rebuffed three times by this Court in its efforts to secure its overruling and having resorted to Congress to nullify its effect, the whole claim of the Government in Hassett v. Welch was that Congress had, as it were, overruled May v. Heiner by the Resolution of March 3, 1931, not only prospectively, but retrospectively. That construction of the Resolution of 1931 had to be rejected in the light of the legislative history of the Resolution. The unanimity of *681the Court’s decision in Hassett v. Welch confirms the inevitability of the decision. And the considerations that led the Government not to attack May v. Heiner in Hassett v. Welch likewise led the Government not to ask the Court to overrule May v. Heiner in this litigation until propelled to do so by this Court’s order for reargument. These considerations were of the same nature, except reenforced by another decade’s respect for May v. Heiner by the Treasury in the actual administration of the revenue law.

Congress has made no change in this section since 1932 and the identical language was carried over as § 811 (c) of the Internal Revenue Code in 1939. There has been no amendment to this language in the Code. Although the sponsors of the Joint Resolution in the House expressed the hope that the next Congress would make the Resolution’s provisions retroactive, nothing of the sort was done. See 74 Cong. Rec. 7199, partially quoted ante at p. 678. Nor did the Treasury remind any subsequent Congress of this unfinished business, despite the fact that it urged amendment of other provisions of the estate-tax law.9

*682The Court during the past decade, in an impressive body of decisions, has given effect to legislative history under circumstances far less compelling than the story here summarized. See the massive body of cases collected in Appendix A, post, p. 687. Moreover, in the face of the legislative history set out above, even an overruling of the five cases in which this precise issue was decided would not give this Court a free hand. For the subsequent actions of Congress make the meaning announced in May v. Heiner and reaffirmed four times as much a part of the wording of the statute as if it had been written in express terms. See Note, 59 Harv. L. Rev. 1277, 1285. An interpretation that “came like a bombshell” certainly had the attention of the Congress. Its failure to alter the language indicates that it accepted that interpretation. See the cases collected in Appendix B, post, p. 690. Due regard for this Court’s function precludes it from ignoring explicit legislative intention even to “yield results more consonant with fairness and reason.” Anderson v. Wilson, 289 U. S. 20, 27; see Cardozo, The Nature of the Judicial Process 14 (1921). What the Treasury could not induce the House to do because the Senate would not vote for it we should not now, eighteen years later, bring to pass simply because our action in this case does not depend upon that body’s concurrence.

*683No comparable legislative history was flouted in Helvering v. Hallock, 309 U. S. 106. It is one thing to hold that Congress is not charged either with seeking out and reading decisions which reach conflicting views in the application of a sound principle or with taking steps to meet such decisions. This is the meaning of our holding in the Hallock case.10 It is quite a different thing to *684say that a statute does not acquire authoritative content when a decision interpreting it has been called to the attention of the public and of Congress and has engendered professional controversy, and when Congress, after full debate, has not merely refused to undo the effect of the decision but' has seen fit to modify it only partially. Helvering v. Griffiths, 318 U. S. 371; United States v. South Buffalo R. Co., 333 U. S. 771, 773-785; cf. Apex Hosiery Co. v. Leader, 310 U. S. 469, 487-489. That is this case.11

*685The opinion of the majority in the Hallock case did not, either explicitly or by implication, declare that May v. Heiner was no longer the accepted interpretation of the pre-1931 part of the language in § 811 (c). When we spoke of what had been “Congressionally discarded” — a reference, incidentally, made to answer the argument that Congress had legislatively recognized the distinction between the Klein12 and the St. Louis Trust13 cases — we meant just what Congress meant, that where a settlor created a trust after May 3, 1931, in which he reserved a life estate, the property transferred would be included in the gross estate. It is significant that only one14 of the many circuit judges who have dealt with the Hallock opinion has thought that it overruled May v. Heiner or that the interpretation there announced was to be changed. Commissioner v. Hall’s Estate, 153 F. 2d 172 (C. A. 2d Cir.); Helvering v. Proctor, 140 F. 2d 87 (C. A. 2d Cir.); Commissioner v. Church’s Estate, 161 F. 2d 11 (C. A. 3d Cir.); United States v. Brown, 134 F. 2d 372 (C. A. 9th Cir.). The contention that the Hallock case overruled May v. Heiner was, one would have supposed, conclusively answered by Judge Learned Hand in Helvering v. Proctor, supra at pp. 88-89:

“The opinion of the majority in Helvering v. Hal-lock, supra, did not explicitly, or by inference from *686anything said, declare that May v. Heiner, supra . . . was no longer law. We do not forget that in a note on page 120 of 309 U. S. . . . Frankfurter, J., spoke of the 'Congressionally discarded May v. Heiner doctrine;’ but it would be quite unwarranted from that to infer that the court meant to overrule that 'doctrine,’ and the note was added for quite another purpose. ... it cannot properly be interpreted as holding that the amendment was a legislative interpretation that May v. Heiner, supra, had been wrongly decided. Perhaps it was wrongly decided; perhaps the amendment is evidence that it was; but the Supreme Court did not say so, or indicate that it thought so. It is true that Roberts, J. in his dissent found no difference (309 U. S. at page 127 . . .) between that decision and Helvering v. St. Louis Union Trust Co., supra, 296 U. S. 39 . . . and apparently thought that consistently, May v. Heiner, supra, must also fall, but the majority did not share his opinion.

''Helvering v. Hallock, supra, 309 U. S. 106 . . . was concerned with quite another situation. The settlor had provided that, if he survived his wife— who had a life estate — the remainder went to him; but if she survived him, the remainder went to her. All that was decided was that, when that was the intent, it made no difference what was the form of words used. It was enough that the settlor’s death cut off an interest which he had reserved to himself upon a condition then determined; that made the remainder a part of his estate. ... If therefore May v. Heiner, supra, 281 U. S. 238 . . . is to be overruled, we do not see how Helvering v. Hallock, supra, can be thought to contribute to that result; it must be overruled by a new and altogether *687independent lift of power, which it is clearly not ours to exercise. Furthermore, if the Commissioner is right, Helvering v. Hallock, supra, 309 U. S. 106 . . . also overruled Hassett v. Welch, 303 U. S. 303 . . . sub silentio. That decision had held that the amendment to § 302 (c) did not operate retroactively; and it would not have been necessary to discuss that question, nor would the actual result have been the same, if May v. Heiner, supra, 281 U. S. 238 . . . had not been law.”

I would reverse Spiegel v. Commissioner, No. 3, and affirm Commissioner v. Estate of Church, No. 5.

Appendix A

DECISIONS DURING THE PAST DECADE IN WHICH LEGISLATIVE HISTORY WAS DECISIVE OF CONSTRUCTION OF A PARTICULAR STATUTORY PROVISION

United States v. Durkee Famous Foods, Inc., 306 U. S. 68; United States v. Towery, 306 U. S. 324; Kessler v. Strecker, 307 U. S. 22; United States v. Maher, 307 U. S. 148; United States v. One 1936 Model Ford, 307 U. S. 219; Sanford v. Commissioner, 308 U. S. 39; Palmer v. Massachusetts, 308 U. S. 79; Valvoline Oil Co. v. United States, 308 U. S. 141; Haggar Co. v. Helvering, 308 U. S. 389; American Federation of Labor v. Labor Board, 308 U. S. 401; Kalb v. Feuerstein, 308 U. S. 433; Morgan v. Commissioner, 309 U. S. 78; South Chicago Coal & Dock Co. v. Bassett, 309 U. S. 251; Amalgamated Utility Workers v. Consolidated Edison Co. of New York, 309 U. S. 261; Germantown Trust Co. v. Commissioner, 309 U. S. 304; Sheldon v. Metro-Goldwyn Pictures Corp., 309 U. S. 390; United States v. City and County of San Francisco, 310 U. S. 16; Sunshine Anthracite Coal Co. v. Adkins, 310 U. S. 381; United States v. American Trucking Assns., 310 U. S. 534; United States v. Dickerson, 310 U. S. 554; Helvering v. Northwest Steel Rolling Mills, Inc., 311 U. S. 46; Neuberger v. Commissioner, 311 U. S. 83; Milk Wagon Drivers’ Union v. Lake Valley Farm Products, 311 U. S. 91; Helvering v. Janney, 311 U. S. 189; Taft v. Helvering, 311 U. S. 195; Hines v. Davidowitz, 312 U. S. 52; United States v. Gilliland, *688312 U. S. 86; Palmer v. Webster & Atlas National Bank, 312 U. S. 156; United States v. Cooper Corp., 312 U. S. 600; Helvering v. Enright, 312 U. S. 636; Maguire v. Commissioner, 313 U. S. 1; Helvering v. Campbell, 313 U. S. 15; Shamrock Oil & Gas Corp. v. Sheets, 313 U. S. 100; Phelps Dodge Corp. v. Labor Board, 313 U. S. 177; Helvering v. William Flaccus Oak Leather Co., 313 U. S. 247; Sampayo v. Bank of Nova Scotia, 313 U. S. 270; Baltimore & Ohio R. Co. v. Kepner, 314 U. S. 44; Parker v. Motor Boat Sales, Inc., 314 U. S. 244; Textile Mills Securities Corp. v. Commissioner, 314 U. S. 326; Gray v. Powell, 314 U. S. 402; District of Columbia v. Murphy, 314 U. S. 441; Illinois Natural Gas Co. v. Central Illinois Public Service, 314 U. S. 498; Duncan v. Thompson, 315 U. S. 1; Cudahy Packing Co. v. Holland, 315 U. S. 357; United States v. Local 807 of International Brotherhood of Teamsters, 315 U. S. 521; Stonite Products Co. v. Melvin Lloyd Co., 315 U. S. 561; Labor Board v. Electric Vacuum Cleaner Co., 315 U. S. 685; Miles v. Illinois Central R. Co., 315 U. S. 698; United States to the use of Noland Co. v. Irwin, 316 U. S. 23; Mishawaka Rubber & Woolen Manufacturing Co. v. S. S. Kresge Co., 316 U. S. 203; Kirschbaum Co. v. Walling, 316 U. S. 517; Helvering v. Cement Investors, Inc., 316 U. S. 527; Marine Harbor Properties, Inc. v. Manufacturers Trust Co., 317 U. S. 78; Braverman v. United States, 317 U. S. 49; Riggs v. Del Drago, 317 U. S. 95; Ex parte Kumezo Kawato, 317 U. S. 69; State Bank of Hardinsburg v. Brown, 317 U. S. 135; Pfister v. Northern Illinois Finance Corp., 317 U. S. 144; United States v. Wayne Pump Co., 317 U. S. 200; Parker v. Brown, 317 U. S. 341; Walling v. Jacksonville Paper Co., 317 U. S. 564; Harrison v. Northern Trust Co., 317 U. S. 476; United States v. Hess, 317 U. S. 537; United States v. Monia, 317 U. S. 424; Ziffrin, Inc. v. United States, 318 U. S. 73; Palmer v. Hoffman, 318 U. S. 109; Overstreet v. North Shore Corp., 318 U. S. 125; Robinette v. Helvering, 318 U. S. 184; Smith v. Shaughnessy, 318 U. S. 176; Helvering v. Sabine Transp. Co., 318 U. S. 306; Federal Security Adm’r v. Quaker Oats Co., 318 U. S. 218; United States v. Swift & Co., 318 U. S. 442; Ecker v. Western Pac. R. Co., 318 U. S. 448; Fred Fisher Music Co. v. M. Witmark & Sons, 318 U. S. 643; Jersey Central Power & Light Co. v. Federal Power Commission, 319 U. S. 61; National Broadcasting Co. v. United States, 319 U. S. 190; Boone v. Lightner, 319 U. S. 561; Schneiderman v. United States, 320 U. S. 118; Hirabayashi v. United States, 320 U. S. 81; Roberts v. United States, 320 U. S. 264; United States v. Dotterweich, 320 U. S. 277; Crescent Express Lines v. United States, *689320 U. S. 401; Colgate-Palmolive-Peet Co. v. United States, 320 U. S. 422; United States v. Laudani, 320 U. S. 543; United States v. Myers, 320 U. S. 561; McLean Trucking Co. v. United States, 321 U. S. 67; Brotherhood of Railroad Trainmen, Enterprise Lodge, No. 27 v. Toledo, P. & W. R. Co., 321 U. S. 50; B. F. Goodrich Co. v. United States, 321 U. S. 126; Davies Warehouse Co. v. Bowles, 321 U. S. 144; Hecht Co. v. Bowles, 321 U. S. 321; Cornell Steamboat Co. v. United States, 321 U. S. 634; Labor Board v. Hearst Pvolications, 322 U. S. 111; Carolene Products Co. v. United States, 323 U. S. 18; Smith v. Davis, 323 U. S. 111; United States v. Rosenwasser, 323 U. S. 360; Western Union Telegraph Co. v. Lenroot, 323 U. S. 490; Hartford-Empire Co. v. United States, 323 U. S. 386; Central States Electric Co. v. City of Muscatine, 324 U. S. 138; Gemsco v. Walling, 324 U. S. 244; Canadian Aviator v. United States, 324 U. S. 215; Connecticut Light & Power Co. v. Federal Power Commission, 324 U. S. 515; A. H. Phillips, Inc. v. Walling, 324 U. S. 490; Brooklyn Sav. Bank v. O’Neil, 324 U. S. 697; Federal Trade Commission v. A. E. Staley Mfg. Co., 324 U. S. 746; Jewell Ridge Coal Corp. v. Local No. 6167, United Mine Workers of America, 325 U. S. 161; Elgin, J. & E. R. Co. v. Burley, 325 U. S. 711; Interstate Commerce Commission v. Parker, 326 U. S. 60; Markham v. Cabell, 326 U. S. 404; John Kelley Co. v. Commissioner, 326 U. S. 521; Roland Electrical Co. v. Walling, 326 U. S. 657; Mabee v. White Plains Pub. Co., 327 U. S. 178; Duggan v. Sansberry, 327 U. S. 499; United States v. Carbone, 327 U. S. 633; Williams v. United States, 327 U. S. 711; Federal Trade Commission v. A. P. W. Paper Co., 328 U. S. 193; Hust v. Moore-McCormack Lines, 328 U. S. 707; United States v. Sheridan, 329 U. S. 379; Oklahoma v. United States Civil Service Commission, 330 U. S. 127; United States v. United Mine Workers of America, 330 U. S. 258; United Brotherhood of Carpenters & Joiners of America v. United States, 330 U. S. 395; American Stevedores, Inc. v. Porello, 330 U. S. 446; Interstate Commerce Commission v. Mechling, 330 U. S. 567; United States v. Ogilvie Hardware Co., 330 U. S. 709; McCullough v. Kammerer Corp., 331 U. S. 96; Ayrshire Collieries Corp. v. United States, 331 U. S. 132; Williams v. Austrian, 331 U. S. 642; Jones v. Liberty Glass Co., 332 U. S. 524; Fong Haw Tan v. Phelan, 333 U. S. 6; Hilton v. Sullivan, 334 U. S. 323; United States v. National City Lines, 334 U. S. 573; United States v. Zazove, 334 U. S. 602; United States v. Congress of Industrial Organizations, 335 U. S. 106; Shapiro v. United States, 335 U. S. 1; Ahrens v. Clark, 335 U. S. 188.

*690Appendix B

OPINIONS DURING THE PAST DECADE RESTING UPON THE RULE THAT THE REENACTMENT OP A STATUTE CARRIES GLOSS OP CONSTRUCTION PLACED UPON IT BY THIS COURT

Electric Storage Battery Co. v. Shimadzu, 307 U. S. 5, 14; Rasquin v. Humphreys, 308 U. S. 54; Apex Hosiery Co. v. Leader, 310 U. S. 469; Brooks v. Dewar, 313 U. S. 354; Helvering v. Griffiths, 318 U. S. 371; Walling v. Halliburton Oil Well Cementing Co., 331 U. S. 17; Commissioner v. Munter, 331 U. S. 210; Francis v. Southern Pacific Co., 333 U. S. 445; United States v. South Buffalo R. Co., 333 U. S. 771; cf. Federal Communications Comm’n v. Columbia Broadcasting System of California, 311 U. S. 132, 132-133; see Mr. Justice Black dissenting in Washingtonian Publishing Co. v. Pearson, 306 U. S. 30, 42; see Mr. Chief Justice Stone dissenting in Girouard v. United States, 328 U. S. 61, 70.

dissenting.

Except for its important reservation to the settlor of a right to the net income of the trust during the settlor’s life, the deed of trust in this case1 is largely comparable to the trust instrument in the Spiegel case, 335 U. S. 701. *691Both speak for themselves as complete transfers in praesenti. Neither was made in contemplation of death.

The evidence of the factual intent of this settlor, likewise, is comparable to that of the settlor in the Spiegel *692case. In fact, the affirmative evidence that the settlor intended to make a transfer complete and absolute in praesenti is stronger here than in the Spiegel case. This settlor avowedly sought to protect not only his family *693but also himself against the possibility of his further disposal of his interest in the corpus of the trust. The remoteness of any possibility of a reverter, arising by operation of law, is comparable here to the remoteness of the alleged possibility of a reverter in the Spiegel case. Two other features of this case, however, require separate consideration.

First. It is the law of New York that must determine here whether the possibility of a reverter, either to the settlor or to his estate, arose by operation of law from the deed of trust. As this case came up from New Jersey, in the Third Circuit, we have no announcement of the law of New York from the United States Court of Appeals for the Second Circuit which includes New York. Furthermore, when the United States Court of Appeals for the Third Circuit rendered its judgment in favor of the taxpayer, it did so in express reliance upon the opinion of the Tax Court and the Tax Court, in turn, did not elucidate the law of New York.

While I rest my conclusion in favor of affirmance upon the absence of the factual intent which, as stated in my dissent in the Spiegel case, I believe is required by *694§811 (c) of the Internal Revenue Code, a substantial argument might be made for affirmance on the ground that, under the law of New York, no possibility of a reverter arose from this trust by operation of law.2 A substantial argument might also be made for affirmance on the ground that the alleged possibility of a reverter, here and also in the Spiegel case, should be disregarded on the doctrine of de minimis non curat lex.

Second. In the opinion of the Court in the instant case, the judgment below is reversed, however, without facing any of the above grounds for its affirmance. This is done by overruling May v. Heiner, 281 U. S. 238, and that action has carried with it the foundation for this Court’s opinion in Hassett v. Welch, 303 U. S. 303. The effect of such reversal is to place this trust, which was executed in 1924, in the same position as though it had been executed after the Joint Resolution of March 3, 1931, 46 Stat. 1516-1517. That Resolution made the federal estate tax applicable to property transferred thereafter by any deed of trust that reserved to the transferor a right to the possession or enjoyment of or the income from the trust property during his life. There is no doubt but that the transfer in the instant case would have been subject to the estate tax if the deed of trust had been executed after, instead of before, the Resolution of March 3, 1931. The legislative history of that Resolution demonstrates, however, that it was not intended to be retroactive. Its prospective character also carried *695with it at least a congressional recognition of the existence of some basis for making a distinction between prior and future transfers of the type described.

After the execution of the instant trust in 1924 — and certainly between March 3, 1931 and the death of the settlor on December 11, 1939 — there was little, if any, reason for him to consider making further disposition of his reserved rights in order to protect his estate from the federal estate tax. Between 1924 and 1939, there were handed down by this Court its decisions in May v. Heiner, supra, on April 14, 1930; Morsman v. Burnet, 283 U. S. 783, and its two companion cases, on March 2, 1931; and Hassett v. Welch, supra, on February 28, 1938. In those of the above decisions which were rendered before March 3, 1931, this Court unanimously and unequivocally held that the federal estate tax was not to be applied to a trust merely because of the retention thereunder of a right in the settlor to receive the income of the trust during his life. The entry made by this Court in each of the companion cases decided March 2, 1931, expressly stated a doubt as to the constitutional authority of Congress to enact a law which would apply the estate tax retroactively to transfers that already had been made. The action of Congress on March 3, 1931, reflected that doubt. The seven Justices who participated in the case of Hassett v. Welch, supra, in 1938, refrained from expressing doubt as to the state of the law before March 3, 1931. In that case the Court reviewed carefully the legislative history that was material to the case and also the administrative interpretation which had been given to the statute. The Court concluded as follows (at pp. 314-315):

“In view of other settled rules of statutory construction, which teach that a law is presumed, in the absence of clear expression to the contrary, to operate prospectively; that, if doubt exists as to the *696construction of a taxing statute, the doubt should be resolved in favor of the taxpayer, we feel bound to hold that the Joint Resolution of 1931 and § 803 (a) of the Act of 1932 apply only to transfers with reservation of life income made subsequent to the dates of their adoption respectively.

“Holding this view, we need not consider the contention that the statutes as applied to the transfers under consideration deprive the respondents of their property without due process in violation of the Fifth Amendment.”

Thus, up to the time of the settlor’s death in 1939, he never was given reason, at least by this Court, to suspect that the property which he had included in his 1924 deed of trust would be-added to his gross estate for federal estate tax purposes.3

The issue as originally presented in May v. Heiner was solely one of statutory interpretation and there were persuasive arguments for deciding the case either way. However, the unanimous decision of this Court in that case changed the status of that issue. Thereafter, the statute carried the meaning ascribed to it by this Court. *697Such an acceptance of the effect of May v. Heiner was expressly stated in the three per curiam companion decisions of March 2, 1931. This acceptance also has been evidenced in some degree by the failure of Congress, at any time, to set forth a contrary view on its part as to the meaning of its original language. Congress merely added new language to change the effect of that interpretation for the future. The Treasury Department conformed its regulations and practices to the reasoning of May v. Heiner. This Court further acceded to this view in 1938 in Hassett v. Welch and in the companion case of Helvering v. Marshall, 303 U. S. 303, when it affirmed the respective lower court judgments in those cases. The lower courts had held that certain pre-1931 comparable trusts, executed in 1920 and 1924, were not subject to the federal estate tax.4 Today, with ten addi*698tional years of administrative practice in conformity with it, the rule of May v. Heiner should be substantially less subject to reversal than in 1938. The doctrine of stare decisis, with full recognition of its appropriate limitations as expressed in Helvering v. Hallock, 309 U. S. 106, 119—122, weighs strongly against a reversal of May v. Heiner now. The problem presented here is just such a one as was said not to exist in the Hallock case. The problem here is one of rejecting a settled statutory construction. This Court’s reversal of the May v. Heiner construction of the estate tax statute as to pre-March 3, 1931 trusts does retroactively, in 1948, what this Court and Congress respectively declined to attempt in 1931. Since 1931, countless taxpayers doubtless have relied upon and benefited by the interpretation announced in May v. Heiner. They had no more right to such benefits than has the taxpayer in this case. If the Government, after this reversal, issues regulations to relieve, in all fairness, settlors who, in demonstrated reliance upon the decisions of this Court and upon the practice of the Treasury Department, have not disposed of their reserved rights under pre-March 3, 1931 trusts like the present one, such special regulations will further emphasize the unique unfairness of enforcing the present decision against the taxpayer in the instant case.

By reversing May v. Heiner this Court repudiates the finality of its 1930 and 1931 decisions interpreting the *699pre-1931 legislation. It holds that the statutory interpretation then announced by this Court of final resort is not final, except as to the parties to the respective cases in which the original judgments are res judicata. After reliance by the Judicial, Legislative and Executive branches of the Government for 18 years upon this authoritative statutory construction, a reversal of it can be justified only by extraordinary circumstances. I fail to find such circumstances, either in the merits of the decision, in the nature of the issue or in the relative importance to the general public of a reversal as against an affirmance of the original interpretation of this tax statute. The statutory interpretation established in May v. Heiner has a peculiarly limited application because its interpretation of the statute in relation to future trusts was cut off on March 3, 1931. Passage of time will soon eliminate transfers made prior to that date by settlors who are yet to die or who have died and whose estates may still be forced to include such transfers for federal estate tax purposes. The 1931 legislation plus the passage of time would thus have disposed of May v. Heiner without the injustices that will now arise from its reversal.

Value is added to the fully considered decisions of this Court by our own respect for them. Faith is justifiable that this Court will exercise extreme self-restraint in using its power of self-reversal. While that power is essential in appropriate cases and is an inherent part of this Court’s finality of jurisdiction, each case that suggests its use should be scrutinized with the utmost care. In the instant case I find arguments to suggest and support, but not to require, a construction of the statute contrary to that originally given in May v. Heiner. I find nothing sufficient to justify the reversal of this Court’s original construction 18 years after this Court approved it unanimously and 17 years after this Court unanimously reaf*700firmed that approval. Likewise, I find nothing in the intervening decisions of this Court that forces this reversal upon us.5 For these reasons, I believe that the judgment of the Court of Appeals should be affirmed on the authority of May v. Heiner and Hassett v. Welch, and upon the principles stated in my dissent in the Spiegel case, post, p. 708.