This is primarily a suit for refund of interest, collected on delayed tax payments, which the plaintiffs would deny to the Government because of the special history of their prior tax litigation. The case has its origin in the income tax returns filed by Frank Erickson, the taxpayer,1 for the years 1937-1950 (exclusive of 1941). In January 1951, on the ground that the returns were fraudulent, the Commissioner of Internal Revenue made jeopardy assessments of tax, fraud penalty, and deficiency interest for each of the years 1937-1946 (except 1941) in large sums varying in amount from over $36,000 to almost $500,000. A deficiency notice was sent in March 1951 (appending a statement of the amount assessed by jeopardy, including interest). In June 1951, the taxpayer sought from the Tax Court a redetermination of his income taxes for those years; he attached a copy of the notice received from the Commissioner, but the Commissioner did not himself file any notice of the amount of the jeopardy assessment (or of any abatement).

In July 1955, the Commissioner made similar jeopardy assessments, again invoking fraud, for the years 1947-1950, in amounts ranging from over $518,000 to over $1,801,000. Notices of these deficiencies were sent in August 1955, including statements of the jeopardy assessments (but this time without mention of interest). In September 1955, the taxpayer commenced two Tax Court proceedings — one for 1947 and another for 1948-1950 — to test these later jeopardy assessments. Again, he attached copies of the notices and statements he had received from the Commissioner, but the latter did not file with the Tax Court any notice of the amount of the jeopardy assessments.

In the years between 1951 and November 1958, the defendant collected all the amounts assessed by way of jeopardy for 1937-1946, together with additional (i. e., delinquency) interest in substantial sums. The jeopardy assessment for 1947 was also collected, as well as a part of that for 1948.

The three Tax Court proceedings challenging these jeopardy assessments were not concluded until 1958. In September 1957, taxpayer's counsel offered to settle the three cases by an agreement that the liability for deficiency in tax and for penalties for each taxable year was equal to a specified dollar amount. The letter added that statutory interest would also be payable. The Government responded by forwarding proposed settlement stipulations incorporating figures which differed only slightly from those suggested by counsel. These stipulations were executed by both sides and filed with the Tax Court on August 27, 1958.

Two days later, on August 29, 1958, the Tax Court entered a decision in each of the three cases, "pursuant to the written stipulation of settlement," deciding the amount of the deficiencies and the overpayments for the various years, in the sums indicated in the stipulations. The court's decisions referred to taxes and to additions to the tax under Section 293(b) (fraud penalty) and Section 294 (d) (2) (penalty for substantial underestimate of estimated taxes) of the 1939 Code, but did not mention or refer to interest.

In the fall of 1958, following these Tax Court decisions, the Commissioner issued to the taxpayer Certificates of Overassessment for various of the taxable years involved in the Tax Court proceedings. Those overassessments which had not been collected were abated; of the remainder, the major portion was credited against unpaid liabilities of the taxpayer, and the rest was refunded. In addition, interest on the overpayments, in the amount of $266,039.87, was paid to the taxpayer.

Through the credits for overpayments, the Commissioner also collected amounts of tax, penalty, and interest (deficiency and delinquency) which he considered to be due under the Tax Court decisions but which had not yet been collected from the taxpayer. In December 1958 and January 1959, the taxpayer filed claims for refund of the interest — not the tax or penalties — obtained by the Government. No action having been taken by the Treasury on these claims, the plaintiffs instituted this suit in July 1959.

The claim for refund of the interest collected by the Commissioner is a strictly technical one, held together by alternative chains of reasoning. The first position (which has two facets) is based on the view that the Tax Court had jurisdiction to determine interest. The argument proceeds in the following steps: (1) Where a Tax Court proceeding is begun pursuant to a statutory notice issued with respect to a jeopardy assessment, that court has complete jurisdiction (under Sections 272 and 273 of the 1939 Code) to determine the full "amount which should have been assessed," including interest; the three Tax Court determinations involved here did not decide that interest was due but on the contrary determined, erroneously or not, that interest was not owing; these Tax Court decisions became final and binding; nevertheless, the interest was collected by the Government, although it exceeded the amount finally determined by the Tax Court as the "amount which should have been assessed"; therefore, the taxpayer is entitled to a refund of the interest and is not precluded by any of the statutory prohibitions on refund suits after a Tax Court ruling, regardless of whether claims for refund were timely filed. (2) If, however, the Tax Court's decisions are read as not reaching the issue of whether interest was included in the "amount which should have been assessed" by the Commissioner, then the defendant is precluded by res judicata from retaining the interest because the Tax Court proceedings could have decided the interest issue (which was an integral part of the cause of action) in the Government's favor, but did not do so; and since the taxpayer filed timely refund claims he is entitled to recover.

The alternative theory assumes that the Tax Court did not have jurisdiction to determine interest, but argues that in any event the jeopardy assessments in this case could not support the collection of interest because the Commissioner failed to follow the statutory requirement (Section 273(c) of the 1939 Code) that he notify the Tax Court of the amounts of the jeopardy assessments.

The defendant meets these elaborate technical contentions on their merits but also presents an initial non-technical argument which, if upheld, would be dispositive. It is that the taxpayer is estopped from seeking interest in this refund suit because he specifically agreed, as part of the settlement which eventuated in the Tax Court decisions and the Commissioner's subsequent actions, that interest was to be payable. Counsel's letter of September 25, 1957, proposing a settlement, said that, "It is understood, of course, that in addition to the foregoing [stated sums of tax deficiency and penalties for the various years] statutory interest will be payable". The stipulations in the two Tax Court cases in which deficiencies were determined expressly stated that, "effective upon the entry of decision pursuant to this stipulation," the taxpayer waived all restrictions on the assessment and collection of the deficiencies and penalties "plus statutory interest." As its orders recognized, the Tax Court obviously acted, when it rendered its decisions two days after the filing of the stipulations, on the basis of the parties' agreement, including the agreement as to interest. The Commissioner followed the stipulations when he collected interest, after the Tax Court decisions. He also followed the parties' understanding when he paid taxpayer $266,039.87 as interest on the overpayments determined by the Tax Court — although the court's orders contained no express finding or determination on interest. The taxpayer never withdrew his agreement on interest, as embodied in the stipulations, by any formal document, and certainly did not do so before the Tax Court decisions became final on November 29, 1958. Neither side objected to the failure of the court to make an express determination on interest, and neither side sought review.

In these circumstances it is apparent that, if we can properly consider estoppel in this case, that defense should be taken into account. The general nature of the claim would normally require us to do so. An action for refund of taxes is essentially governed by equitable principles. Lewis v. Reynolds, 284 U.S. 281, 52 S. Ct. 145, 76 L. Ed. 293 (1932); Stone v. White, 301 U.S. 532, 534-535, 57 S. Ct. 851, 81 L. Ed. 1265 (1937); Ryan v. Alexander, 118 F.2d 744, 748 (C.A. 10, 1941); Blansett v. United States, 283 F.2d 474, 479, 480-81 (C.A.8, 1960); Western Maryland Railway Co. v. United States, 131 F. Supp. 873, 889-890 (D.Md., 1955). That is not to say either that the taxpayer must always show that he has in fact overpaid his tax, or that statutes preventing the Government from retaining the money (despite the absence of an overpayment) should not be respected. McEachern v. Rose, 302 U.S. 56, 58 S. Ct. 84, 82 L. Ed. 46 (1937); Lyddon & Co. (America), Inc. v. United States, 158 F. Supp. 951, 141 Ct. Cl. 545, 551, cert. denied, 358 U.S. 832, 79 S. Ct. 55, 3 L. Ed. 2d 70 (1938). But the principle that equity is woven into the warp and woof of a refund suit does mean that the court should apply so fundamental an equitable concept as estoppel, when the circumstances warrant, unless some Congressional enactment or valid regulation requires a different result. See R. H. Stearns Co. of Boston, Mass. v. United States, 291 U.S. 54, 61-62, 54 S. Ct. 325, 78 L. Ed. 647 (1934); Building Syndicate Co. v. United States, 292 F.2d 623, 626 (C.A.9, 1961); Guggenheim v. United States, 77 F. Supp. 186, 111 Ct. Cl. 165, 182 (1948), cert. denied, 335 U.S. 908, 69 S. Ct. 411, 93 L. Ed. 441 (1949); Daube v. United States, 5 F. Supp. 769, 78 Ct. Cl. 754, 763-764 (1934); Naumkeag Steam Cotton Co. v. United States, 2 F. Supp. 126, 135, 76 Ct. Cl. 687, 702-711 (1933), cert. denied, 289 U.S. 749, 53 S. Ct. 694, 77 L. Ed. 1495 (1933).2

The plaintiffs answer that, whatever may be the general rule in refund cases, we cannot consider equitable estoppel in this case because the defendant did not set it up in its pleadings as an affirmative defense — as demanded by our Rule 15(b) (following Rule 8(c) of the Federal Rules of Civil Procedure). Ordinarily, estoppel, like laches, should be pleaded at the outset. Crocker v. United States, 127 F. Supp. 568, 573, 130 Ct. Cl. 567, 575 (1955). But here the course of the proceedings was such that it is just and proper to consider the pleadings amended (under our Rule 18(b), "Amendments To Conform to the Evidence") to state the defense of estoppel. During the trial the plaintiffs' counsel was aware, although he objected to consideration of the defense, that the defendant contemplated raising estoppel (Tr. 37 ff.). His objections to certain evidence on this ground (and others) were overruled, and therefore it must have seemed quite possible that estoppel would be held to be in the case. True, plaintiffs' counsel did not introduce the evidence to counteract this defense which he would probably have offered had estoppel been pleaded. But in his brief and at the bar he told us what that evidence would have been; and if we consider it on the estoppel issue no injustice will be done to either party. See Moore's Federal Practice (2d ed.) §§ 12.23, 15.13, and 15.14.3 The equitable foundation of refund suits moves us to give an equitable application in this case to the rules of pleading and of amending pleadings.

Considering the defense of estoppel as properly before us, we hold that it must be sustained. In his settlement offer, taxpayer expressly proposed that interest would be payable; the settlement was consummated on that basis; the Tax Court orders were premised on, and induced by, stipulations envisaging the collection of interest; the Commissioner acted on that basis, both in collecting interest from taxpayer and in refunding overpayments of interest; the taxpayer did not take any formal written action disavowing his agreement until after the Commissioner had acted and the Tax Court decisions had become final in November 1958. Taxpayer tells us that his counsel orally indicated to representatives of the defendant in October 1958, before the Tax Court orders had become final, that he intended to sue for most of the interest assessed and not refunded. This, he says, was sufficient warning to the Government that his position had changed. But such informal oral statements, even if strongly asserted, would not change the reliance which the Government had already placed on the prior written agreement and could rightfully continue to place, at least until the agreement was formally repudiated in writing. The decisions, cited above, which have applied estoppel in tax refund suits show that where the Government has relied on requests or agreements of the taxpayer, the latter cannot recover by thereafter changing his position. "The doctrine [of estoppel] particularly applies where the acts complained of are induced and brought about by the very party complaining." Daube v. United States, supra, 5 F.Supp. at 773, 78 Ct. Cl. at 764 (1934). Nothing in the statutes on which plaintiffs rely for their right to recover overrides that pervasive principle which is controlling here.

If we disregard estoppel and delve into the merits of taxpayer's major claim, we still must reject it. The first of the two alternative arguments for recovery, outlined above, assumes that the Tax Court had jurisdiction to determine interest. This may or may not be so — we do not decide the issue — but it is quite plain that the Tax Court, in the three cases involving this taxpayer, thought that it had no such power and therefore did not make any determination, express or implied, as to interest, one way or the other. In Fuller v. Commissioner, 20 T.C. 308, 313, 318, decided in 1953 (some five years prior to taxpayer's cases) that court held, in a jeopardy assessment case, that it had no jurisdiction "over matters relating to interest." That view was undoubtedly reflected in the subsequent orders in the taxpayer's cases which contained no express finding or determination, in any form, as to the amount of interest which should have been assessed. The plain fact is that the Tax Court did not consider or take account of interest in determining (under Section 273(i) ) the "amount which should have been assessed," and did not decide that interest was or was not due. It was not that the court found no interest due because it felt itself without authority to order it paid; it simply did not reach the interest question at all.

Assuming with taxpayer (but not deciding) that the Tax Court erred in its renunciation of jurisdiction, the question then becomes whether its refusal to pass upon the defendant's right to interest nevertheless precludes the defendant from now asserting that right because interest should have been adjudicated as part of the causes of action before the Tax Court — even though that court considered itself deprived of power. The principles of res judicata do not extend that far. The rule is that, where a court erroneously holds that a particular claim cannot be enforced in that action but only in a separate suit, the judgment in the first action does not preclude enforcement of the claim in another litigation. See A.L.I., Restatement of the Law of Judgments, sec. 67 ("Where First Court Makes Erroneous Ruling That Recovery Can Be Had only in a Different Action"); Ash Sheep Co. v. United States, 252 U.S. 159, 170, 40 S. Ct. 241, 64 L. Ed. 507 (1920). Under that sensible rule, the Tax Court's refusal on jurisdictional grounds to consider interest, although it may have been wrong, would not bar the Government from defending this suit on the basis that interest was properly assessed.

Taxpayer's second, alternative, argument (as we understand it) is that, even if the Tax Court does not have general jurisdiction to determine interest in jeopardy assessment cases, still the Commissioner's collection of interest was invalid because he did not comply with the mandate of Section 273(c) that he "notify" the Tax Court of the amount of the jeopardy assessment. If, as in this case, a jeopardy assessment is made before issuance of the statutory notice of deficiency, the later statutory notice, which is required to be filed by the taxpayer on beginning his case in the Tax Court, will always contain a statement concerning the jeopardy assessment. In that situation the Commissioner's normal practice is not to file with the court a notice of the amount of the jeopardy assessment, but to rely on the statement set forth in the statutory deficiency notice which the taxpayer must file. We do not and need not pass upon the question whether this practice fully complies, for all purposes, with the requirement in Section 273(c). But we do decide that where, as in this case, the Tax Court has been apprised by the taxpayer of the Commissioner's statement of the amounts of the jeopardy assessment (tax and additions), failure of the Government itself to notify the court directly of those amounts does not invalidate the jeopardy assessment, or deprive the Tax Court of jurisdiction, or prevent the Commissioner from collecting interest otherwise due. There is no reason to suppose that the notification requirement of Section 273(c), even if it does order the Government itself to file a notice, has the enormous potency which the taxpayer would give it.

Aside from its two arguments based on the jurisdiction of the Tax Court over interest and on the Commissioner's failure to file a notice of the amounts of the jeopardy assessments — contentions which we have just rejected — the taxpayer does not deny that the interest collected by the Commissioner was owing. In any event, it is plain that under the 1939 Code interest was due on these deficiencies to the date of assessment (deficiency interest) and on all assessed amounts until paid (delinquency interest). See Sections 273, 292, 294, and 297 of the 1939 Code; United States v. Glasser, 287 F.2d 433 (C.A.7, 1961); Ginsburg v. United States, 278 F.2d 470 (C.A.1, 1960); Symonides v. Crenshaw, 45 A.F.T.R. (RIA) 1171 (E.D.Va., 1953); Brinck v. United States, 173 F. Supp. 286 (D.Mont., 1959); Grimsley v. Patterson, 183 F. Supp. 729 (N.D.Ala., 1957); United States v. Miyamoto, 5 A.F.T.R.2d (RIA) 1154 (D.Haw., 1960); but see Crolich v. United States, 144 F. Supp. 109 (S.D.Ala., 1956).4

The plaintiffs' minor claim is a separate one involving income taxes for 1942. For that year the taxpayer filed his return on or about March 15, 1943, showing a liability of $22,242.60, of which $62.40 was paid at the source. He made timely payments of $11,090.10, and the balance ($11,090.10) was abated on September 1, 1943, under Section 6 of the Current Tax Payment Act of 1943, 57 Stat. 126. That Act, which put federal income tax collection on a current "pay as you go" basis, forgave the major portion of the 1942 tax, as of September 1, 1943.5

Thereafter, as part of his jeopardy assessment on the ground of fraud for all the years from 1937 to 1946 (except 1941), the Commissioner found that taxpayer's 1942 return was fraudulent and assessed additional amounts of tax, fraud penalties, and interest; in determining the deficiency in tax for 1942, the Commissioner used, as a starting point, the reported tax balance due of $22,180.20, unreduced by the abatement of $11,090.10. The Tax Court decision for that year, based on the parties' compromise, found a deficiency in tax of $64,802.78 and in addition to the tax for fraud (under Section 293(b) ) of $32,401.39. The unpaid portions of these deficiencies were collected, after the Tax Court's decision, as part of the Commissioner's adjustments made in October and November 1958. The total amount unpaid for 1942, and collected at that time, was $24,214.43.

In addition, the Commissioner, on September 25, 1958, without prior notice, made another assessment of $21,422.12 for 1942, representing a reassessment of the previously abated part of the 1942 tax ($11,090.10), together with deficiency interest of $10,332.02. This additional assessment was collected (in October 1958) from the overpayment for 1938. No notice of deficiency with respect to this assessment was issued. The Commissioner sought this extra tax for 1942 because Section 6 of the Current Tax Payment Act of 1943 provided that its forgiveness feature was not to be applicable in any case "in which additions to the tax for such taxable year [1942] are applicable by reason of fraud." In the Commissioner's view, the Tax Court's decision of August 29, 1958, determining fraud penalties for 1942, required that the previous discharge of part of the 1942 tax should be set aside. The plaintiffs sue to recover the $21,422.12 thus collected.

We find it unnecessary to decide whether the additional assessment for 1942 made by the Commissioner on September 25, 1958, was a deficiency assessment, within Section 271 of the 1939 Code, requiring a notice to the taxpayer and an opportunity to seek further relief in the Tax Court before collection could be made.6 For on the facts proved here we hold that the erasure of the prior abatement for 1942 because of taxpayer's fraud could and should have been presented in the pending Tax Court proceeding; that the Commissioner should not have waited until after the Tax Court decision to raise this problem; and that the Tax Court decision is res judicata as to taxpayer's liability for income tax and fraud additions for 1942.

The issues of fraud and of additions to the tax because of fraud were directly before the Tax Court in the proceeding relating to 1942 (Docket No. 34883). The parties' stipulation agreed that there were deficiencies in tax of $64,802.78 and in additions under Section 293(b) (pertaining to fraud) of $32,401.39. The Tax Court so determined. Moreover, the Commissioner was quite aware of the taxpayer's fraud long before the Tax Court decision on August 29, 1958, and even before the Tax Court proceeding was begun. On January 23, 1951, he made a jeopardy assessment for 1942, on fraud grounds, in the amount of $52,821.29 in tax, $26,410.64 in fraud penalties, and $24,895.47 in interest (a total of $104,127.40). Significantly, in making this calculation the defendant used the balance reported by taxpayer in his 1942 return ($22,180.20), not the amount reduced by the abatement; this suggests that at that time the Commissioner was already invoking the provision of the Current Tax Payment Act of 1943 which precluded abatement under that Act for taxpayers guilty of fraud for 1942. Taxpayer's Tax Court proceeding challenging this assessment (as well as those for other years) was initiated on June 5, 1951. There is no doubt that in that proceeding the Commissioner could have explicitly asked the court to set aside, because of the taxpayer's fraud, the prior abatement of a large portion of the 1942 tax. Section 272(a), (b), (e), and (g) and Section 273 of the 1939 Code.

It necessarily follows, we think, that the Tax Court was the forum in which there should have been presented and decided all questions (relating to taxpayer's income tax for 1942) of the amount of the deficiency in tax, of fraud, and of additions to the tax because of fraud. In a situation such as this there is special force to the familiar statement that the Tax Court's jurisdiction, once it attaches, extends to the entire subject of the correct tax for the particular year.7 The cause of action then before the court encompassed all phases of the taxpayer's income tax for 1942, including the fraud which was already known.

It follows, too, that the Tax Court's decision of August 1958 with respect to 1942 is binding and res judicata in the present case, even though the defendant may not have raised in that proceeding any question as to the prior abatement under the Current Tax Payment Act. That decision bars further litigation not only on those aspects of the 1942 tax which were actually raised but also on the issues which could have been presented. American Woolen Co. v. United States, 18 F. Supp. 783, 21 F. Supp. 125, 1021, 85 Ct. Cl. 101, 112, 115, 117, 119 (1937-38), cert. denied, 304 U.S. 581, 58 S. Ct. 1055, 82 L. Ed. 1544 (1938); T. W. Warner Co. v. United States, 15 F. Supp. 160, 83 Ct. Cl. 495, 511-512 (1936); Lehigh Valley Trust Co. v. United States, 34 F. Supp. 839, 841 (E.D. Pa., 1939); Lenny v. Williams, 143 F. Supp. 29, 34 (N.D.Ohio, 1956); Griswold, Res Judicata in Federal Tax Cases, 46 Yale L.J. 1320, 1328-30, 1357 (1937).8

This conclusion is not affected by the fact that the Tax Court decision was entered on the basis of an agreement by the parties to compromise the case. Tax Court judgments resting on such stipulations are res judicata for the years involved, although they may not be effective for other years. United States v. International Building Co., 345 U.S. 502, 506, 73 S. Ct. 807, 97 L. Ed. 1182 (1953). We have applied the doctrine of res judicata against a taxpayer where the prior decision of the Tax Court was based on a compromise settlement. Maher v. United States, 172 F. Supp. 689, 690, 145 Ct. Cl. 701, 704 (1959). A District Court has applied it against the Government in a case very close to the present one. Lenny v. Williams, supra. There is no reason to depart from these holdings. The Tax Court's decision establishing the tax owing for 1942 is not open to re-litigation. Since the Commissioner collected a sum in excess of the proper amount computed in accordance with that Tax Court ruling, the plaintiffs are permitted by Section 322(c) (2) of the 1939 Code to bring this suit for refund.

Accordingly, the plaintiffs are not entitled to recover on their major demand for interest, and the petition will be dismissed as to those claims. But plaintiff Frank Erickson is entitled to recover on his claim for the extra $21,422.12 collected for 1942, with interest as provided by law, and judgment will be entered to that effect; the amount of recovery will be determined pursuant to Rule 38(c).

JONES, Chief Judge, and DURFEE, LARAMORE and WHITAKER, Judges, concur.