RYAN, Circuit Judge.
The taxpayers, Jane Johnson and Joan Kincaid, appeal a Tax Court decision rejecting the deduction of certain legal expenses. The central issue in the case is whether Treas.Reg. § 1.212-l(k) should be read to create an exception to the general rule of [1267]*1267nondeductibility of expenses incurred to “recover” property. The proposed exception would render deductible, under I.R.C. § 212, the expense of litigation to recover title or possession of stock.
We conclude that the Tax Court’s holding denying the deductibility of these expenses should be affirmed. We hold, however, that the taxpayers are entitled to a deduction for the portion of expenses allocable to the recovery of dividends and interest.
I.
The taxpayers, two sisters, owned stock in a finance company and several banks. Both taxpayers granted their father options to buy various of these stocks at more or less advantageous prices. Their father died without exercising the options. The executor of the father’s estate then sought to exercise the options. The daughters refused to accept the tendered payment, and litigation ensued.
The daughters argued that the options were personal and therefore did not survive their father’s death, that the options were void for lack of consideration, and that the options were revocable and had in fact been revoked before the executor sought to exercise them.
Title to the finance company stock had passed to the executor prior to the litigation. Joan, who owned the finance company stock, prevailed in the Kentucky Court of Appeals, and the judgment was affirmed by the Kentucky Supreme Court. She was declared to be the owner of the stock and, in response to a supplemental complaint filed after the first case had been affirmed on appeal, she was also awarded a large sum of money, representing dividends and interest received by the executor while he held the stock.
The bank stocks, meanwhile, which were owned by both sisters, were held by the executor as the litigation went forward. Title to the bank stocks did not pass to the executor prior to the litigation, so the sisters continued to receive dividend income from their bank shares. The sisters lost their separate suits concerning the bank stocks, and the executor paid each of them for their bank stocks according to the terms of the options the sisters had granted to their father.
Both sisters deducted the legal expenses associated with all of this litigation on their 1977 tax returns. The Internal Revenue Commissioner audited their returns and determined that the expenses were not deductible. Subsequently, the Tax Court held that the expenses were nondeductible capital expenditures.
II.
This appeal presents a narrow legal issue, but one which implicates the highly artificial and thus continually vexed problem of distinguishing capital from ordinary expenditures. I.R.C. § 212 provides in part that:
In the case of an individual, there shall be allowed as a deduction all the ordinary and necessary expenses paid or incurred during the taxable year—
(1) for the production or collection of income; [or]
(2) for the management, conservation, or maintenance of property held for the production of income____
I.R.C. § 263(a) makes capital expenditures nondeductible. Capital expenditures are defined as “[a]ny amount paid out for new buildings or for permanent improvements or betterments made to increase the value of any property or estate.” As explained in Treas.Reg. § 1.263(a)-2(c), one example of a capital expenditure is “[t]he cost of defending or perfecting title to property.”
Taxpayers’ argument for the deductibility of their legal expenses rests almost solely upon the language of Treas.Reg. § 1.212-l(k), which reads as follows:
Expenses paid or incurred in defending or perfecting title to property, in recovering property (other than investment property and amounts of income which, if and when recovered, must be included in gross income), or in developing or improving property, constitute a part of the cost of the property and are [1268]*1268not deductible expenses. Attorneys’ fees paid in a suit to quiet title to lands are not deductible: but if the suit is also to collect accrued rents thereon, that portion of such fees is deductible which is properly allocable to the services rendered in collecting such rents. Expenses paid or incurred in protecting or asserting one’s right to property of a decedent as heir or legatee, or as beneficiary under a testamentary trust, are not deductible.
Id. (emphasis added).
Taxpayers contend that:
Expenses paid to acquire or dispose of property or to defend or perfect title to property (which is simply a subcategory of acquiring or disposing of property) are also capital expenditures. Woodward v. Commissioner, 379 [397] U.S. 572, 90 S.Ct. 1302, 25 L.Ed.2d 577 (1970). Expenses paid to recover property are within this category of capital expenditures unless the expenses are paid to recover investment property. Treas.Reg. § 1.212-l(k). The exception for recovering investment property is a recognition that recovery expenses are paid for the production of income under I.R.C. § 212(1), and the conservation and maintenance of property held for the production of income under I.R.C. § 212(2), rather than enhancing the value of or title to property.
III.
Treas.Reg. § 1.212-l(k) has never been authoritatively held to have the effect attributed to it by the taxpayers. Two cases cited by the taxpayers lend only slight credence to their argument. Loyd v. United States, 153 F.Supp. 416, 139 Ct.Cl. 626 (1957), cited the predecessor to § 1.212-l(k) only as an alternative basis for its holding. The Tax Court in Cruttenden v. Commissioner, 70 T.C. 191 (1978), aff'd on other grounds, 644 F.2d 1368 (9th Cir.1981), construed the regulation to allow “a deduction for expenses paid or incurred for the recovery of investment property which by its very nature is held for the production of income.” Id. at 202. However, the Ninth Circuit, in affirming the result in Cruttenden, expressly disavowed the Tax Court’s reading of this regulation and based its affirmance upon a conclusion that the regulation. did not apply. Cruttenden v. Comm’r., 644 F.2d 1368, 1378 (9th Cir. 1981).
The Tax Court, in this case, has held that the “linchpin” of the litigation was title to the stock.
Consistently since 1916 the Treasury Regulations have provided that the cost of defending or perfecting title to property constitutes a part of the cost of the property and is not a deductible expense. Congressional approval of this interpretation by reenactment of the applicable revenue statutes in successive acts in identical language gives to the regulation the efficacy of law.
Jones’ Estate v. Comm’r., 127 F.2d 231, 232 (5th Cir.1942). This circuit has acknowledged since at least 1948 that the regulations making the cost of defending or perfecting title nondeductible have “the effect of law.” Porter Royalty Pool, Inc. v. Comm’r., 165 F.2d 933, 936 (6th Cir. 1948). See also Safety Tube Corp. v. Comm’r., 168 F.2d 787, 789 (6th Cir.1948).
In both Loyd and Cruttenden, the courts found that the legal expenses in question were not attributable to disputes over title. In Loyd,
The primary purpose, or at least the motivating force, of the litigation was not the question of who had title to the property, rather it was that the proper distribution of the estate’s assets could not have been had without a prior determination by the court of the validity of the gift.
153 F.Supp. at 421. In Cruttenden, the Tax Court found that “the legal expenses paid by petitioners in no manner related to the title of the securities.” The subject of the litigation was the right to the possession of the securities.
The Court of Claims in Loyd and the Tax Court in Cruttenden presumed that construing Treas.Reg. § 1.212-l(k) to create a deduction for the expense of “recovering” investment property would not conflict [1269]*1269with the well-established rule that costs of “defending or perfecting title” are not deductible. In the instant case, however, the taxpayers adopt as their definition of “recovering” the Ninth Circuit’s language in Cruttenden, which confines “recovering property” to “instances where ownership, having been lost or relinquished, has been restored.” 644 F.2d at 1378. Thus, the taxpayers apparently contend that “recovering” property is a subset of “defending or perfecting title,” and that the regulation makes recoveries of investment property an exception to the rule of nondeductibility. Even if these cases may be considered precedential, the precedent they represent is much narrower than the rule taxpayers would have us adopt.
Taxpayers do not contend that their rule is necessarily implicit in the Code itself. Thus, in the absence of case authority, taxpayers are left with only the regulation in apparent support of their position.
IV.
The taxpayers propose that we adopt a sort of “plain meaning” approach to the language of Treas.Reg. § 1.212-l(k). The regulation provides in part that:
Expenses paid ... in recovering property (other than investment property and amounts of income which, if and when recovered, must be included in gross income) ... are not deductible____
Id. Taxpayers read this language as if it said: “Expenses paid in recovering investment property are deductible.” This is a plausible reading, but one which does not withstand close scrutiny.
In construing this language, “We are mindful of the considerable weight accorded Treasury Department Regulations.” Dorfman v. Comm’r., 394 F.2d 651, 655 (2d Cir.1968). However, there are limits to the deference we accord to them. “Regulations ‘are not absolute rules of law’ and should not be followed where they are in conflict with the ‘design’ of the applicable section of the Code.” Dorfman, 394 F.2d at 655 (quoting Mitchell v. Comm’r., 300 F.2d 533, 537-38 (4th Cir.1962)).
The Commissioner’s regulations are his interpretations of the meaning of the income tax law and where they are not contrary to the law and uniformity requires it, they may be given the force of law in the courts. But we may not unnecessarily expand the intendment of the Regulation to work an exemption from tax of income which Congress has made subject to tax.
Helvering v. Jane Holding Corp., 109 F.2d 933, 939 (8th Cir.), cert. denied, 310 U.S. 653, 60 S.Ct. 1102, 84 L.Ed. 1418; 311 U.S. 672, 61 S.Ct. 34, 85 L.Ed. 432 (1940). “The taxpayer carries the burden of showing that he comes within the provisions of a specific deduction.” Merlino v. Comm’r., 660 F.2d 415, 416 (9th Cir.1981). And as we observed in National Life and Accident Insurance Co. v. United States, 524 F.2d 559, 560 (6th Cir.1975):
A treasury regulation may implement a statute but it cannot supplement it____ The Courts and the Commissioner do not have the power to repel or amend the enactments of the legislature even though they may disagree with the result; rather it is their function to give the natural and plain meaning effect to statutes as passed by Congress.
Here, it is argued, the Tax Court has denied a deduction under this regulation in apparent disregard of the regulation’s “natural and plain meaning.” This may have been error, but, “We do not overturn a Tax Court decision absent a clear mistake of law.” Merlino, 660 F.2d at 416. We cannot give effect to the language taxpayers rely upon unless we are convinced that this will implement the statutory scheme Congress has enacted.
y.
The pertinent statutes in this case are I.R.C. § 212, allowing a deduction for expenses for the production of income, and I.R.C. § 263, disallowing a deduction for capital expenditures. I.R.C. § 211 makes § 212 “subject to” § 263 and associated provisions disallowing deductions; I.R.C. § 261 flatly provides that “no deduction shall in any case be allowed in respect of [1270]*1270the items specified in this part [which includes § 263].” See Commissioner v. Idaho Power Co. 418 U.S. 1, 17, 94 S.Ct. 2757, 2767, 41 L.Ed.2d 535 (1974). No “plain meaning” approach to Treas.Reg. § 1.212-l(k) can be acceptable unless it implements this statutory scheme.
Recently, two other circuits have discussed the regulation upon which taxpayers rely. The Second Circuit has observed:
The policy behind 26 C.F.R. § 1.212-l(k) is that expenses incurred in acquiring income-producing property — such as brokerage fees incurred in the process of acquiring stocks are “part of the cost of the property,” id., and are therefore treated as non-deductible capital expenditures. These expenditures are added to the basis of the capital asset in connection with which they are incurred, and are taken into account for tax purposes either through depreciation of the asset or through reduction of the capital gain (or augmentation of the loss) when the asset is sold.
Burch v. United States, 698 F.2d 575, 577 (2d Cir.1983) (citing Woodward v. Comm'r., 397 U.S. 572, 575, 90 S.Ct. 1302, 1305, 25 L.Ed.2d 577 (1970)).
The Ninth Circuit appears to be the only circuit to have addressed the specific language upon which the taxpayers rely here. The Cruttenden court responded to the Tax Court’s literal application of this language as follows:
The critical inquiry in interpreting this regulation is for distinctions between capital and non-capital transactions. The Tax Court’s interpretation is unsatisfactory because no such distinction, turning solely on whether property was held for the production of income, transforms a capital expenditure into an ordinary and necessary expense. To the contrary, an expenditure classified as capital in relation to property not held for the production of income would retain that capital character in relation to property held for the production of income.
If the term “investment property” means the same thing as “property held for the production of income,” as the Tax Court found, it would make no sense to except investment property from the denial of deductibility for a capital expense. Indeed, such an exception would be contrary to the dictates of section 263 of the Internal Revenue Code.
644 F.2d at 1377.
We agree that construing § 1.212 — l(k) to create a deduction for all “recoveries” of investment property would produce an inexplicable anomaly in the statutory scheme regarding capital and ordinary expenditures. Furthermore,
Uniformity among the circuits is especially important in tax cases to ensure equal and certain administration of the tax system. We would therefore hesitate to reject the view of another circuit.
First Charter Financial Corp. v. United States, 669 F.2d 1342, 1345 (9th Cir.1982). Taxpayers’ rationale for their claimed deduction amounts to nothing more than a conclusion that recovering investment property tends to produce income and to conserve income-producing property. They do not explain why, conceptually, their expenses should not be considered capital expenditures, and they fail to address the priority-setting provisions of I.R.C. §§ 211 and 261. We therefore reject taxpayers’ view (and that of the Tax Court in Cruttenden ) that we must give effect to the so-called “plain meaning” of the regulation.
The Ninth Circuit also rejected the Commissioner’s view, “that the parenthetical exception for investment property applies only in cases where the property, if and when recovered, must be included in gross income.” 644 F.2d at 1376. This view had convinced the Tax Court dissenters in Cruttenden, who, while alert to the difficulties of the regulation’s language, offered little in the way of an affirmative explication of that language:
Since section 212 applies only to investment property, there would be no purpose in providing that the expenditures for a recovery of property are not deductible under section 212 if such a rule did not apply to investment property. It may be that the purpose of the reference [1271]*1271to investment property was merely to indicate that if such property is includable in income when recovered because, for example, the loss of the property had given rise to a deduction, the expenses of such recovery may be deductible. In any event, the reference to investment property cannot justify a departure from the well-established rule that capital expenditures are not deductible under section 212, and we should adopt an interpretation of the regulation consistent with such rule.
70 T.C. at 207 (Simpson, J., dissenting). The Ninth Circuit responded:
[W]e cannot accept the Commissioner’s position that the problems in interpreting the regulation can be solved by requiring the value of “investment property” to be includable in gross income on recovery in order to escape the regulation’s bar to deductibility. Again, the crucial inquiry is for distinctions between capital and ordinary expenditures. We can see no reason why the parenthetical exception should be so worded as to allow the deduction only where the property that must be included in gross income on recovery is “investment” property. Recovery of any sort of property wherein the value of the recovered property must be included in gross income seemingly ought to qualify.
644 F.2d at 1377-78 (emphasis in original).
This is a telling response, for the cost of any “recovery” of included income would appear to be squarely within the statute’s provision regarding expenses “for the production or collection of income,” regardless of the form of the income. The Commissioner’s view is entirely consistent with the weight of authority in this circuit and elsewhere in support of the “defending or perfecting title” rule. This view is unpersuasive, however, to the extent that it fails to explain the regulation’s “investment property” language.
In sum, we are offered no persuasive explication of the regulation’s “recovering property” clause. A “plain meaning” approach yields a regulation that would supplement, rather than, implement, the Internal Revenue Code, by inexplicably making a class of capital expenditures deductible. A looser reading, which views the clause as providing only that the costs of “recovering” includable income are deductible, effectively reads the “investment property” language out of the regulation.
The Tax Court’s decision in this case, however, also effectively ignores the “investment property” language. Thus, while we are reluctant simply to ignore the regulation’s language, the deference due to the Tax Court’s decision dictates that we determine whether there is a sound legal basis for holding this portion of the regulation to be unauthorized by the Code.
VI.
The predecessor of Treas.Reg. § 1.212-l(k) was first published in December, 1942. See T.D. 5196, 1942-2 C.B. 96. A new regulation, 26 C.F.R. § 19.23(a)-15, was inserted into the treasury regulations in response to the Revenue Act of 1942, Pub.L. No. 77-753, § 121, amending § 23(a) of the Internal Revenue Code of 1939. This amendment added the predecessor to current I.R.C. § 212.
This new Code section was originally titled “Non-Trade or Non-Business Expenses,” and the purpose of the new regulations was to conform the law on the deductibility of the business expenses. Thus, the unnumbered paragraph which was the predecessor of § 1.212-l(k) began by borrowing the language on “defending or perfecting title” from the existing regulation on expenditures in the business context, then at 26 C.F.R. § 19.24-2.
Not all of the law on the deductibility of business expenses, however, was contained in the treasury regulations. The “recovering property” clause of the new regulation apparently was intended to reflect the result in the Supreme Court case of Kornhauser v. United States, 276 U.S. 145, 48 S.Ct. 219, 72 L.Ed. 505 (1928), which interpreted the business expense provisions of the Revenue Act of 1918. In Komhauser, the taxpayer had been sued by a former partner, who sought an accounting for fees paid in the form of stock to the taxpayer [1272]*1272and allegedly due the partnership. The taxpayer won, thus retaining the income, and the Court held his legal expenses to be deductible.
Although the Court’s opinion did not discuss the capital expenditure issue, the Commissioner’s brief had argued “that the fee was paid to retain title to stock and therefore should be capitalized.” Brookes, Litigation Expenses and the Income Tax, 12 Tax.L.Rev. 241, 244 (1957). Because the Court’s result was strictly inconsistent with the “defending or perfecting title” rule, “it is reasonable to suppose the court considered (the capital expenditure argument) and meant to reject it.” Id. Komhauser thus represented a significant limitation on the “defending or perfecting title” rule.
In 1942, however, the extent to which Komhauser limited the “defending or perfecting title’s rule” was unclear. Presumably, if the Court had intended to overrule the already well-established regulation, it would have done so expressly. On the other hand, because the taxpayer in Komhauser was defending his title to stock, the Commissioner could have concluded that the “defending or perfecting title” rule had been supplanted by the rule stated in Komhauser, that “where a suit or action against a taxpayer is directly connected with, or ... proximately resulted from, his business, the expense incurred is a business expense.” Komhauser, 276 U.S. at 153, 48 S.Ct. at 220. Evidently, the Commissioner chose not to resolve any inconsistencies, but to leave a version of the Komhauser holding adjacent to the “defending or perfecting title” rule in the new "non-business expense” regulation.
This historical context suggests several things about the intended meaning of the “recovering property” clause. First of all, we perceive no historical basis for the Ninth Circuit’s suggestions that “the term [recovering property] must refer to instances where ownership, having been lost or relinquished, has been restored.” Cruttenden, 644 F.2d at 1378. Komhauser was a suit “to recover” taxes paid after the government disallowed the taxpayer’s deduction for litigation expenses. The litigation itself, however, concerned fees paid to the taxpayer and for which he had been sued. He never relinquished ownership, and he only “recovered” the property in the sense that he defended the suit successfully. We conclude that “recovering property” means no more than that one has prevailed in litigation in which one’s preexisting property rights were at stake.
Similarly, there is no basis, other than the regulation’s disjunctive syntax, for presuming that “recovering property” only refers to recovering possession, where title was not in issue. In Komhauser, title was indisputably in issue. Thus, the most reasonable understanding of the clause is that it encompasses recoveries where either possession or title was in dispute. We therefore disagree with the construction apparently placed upon “recovering property” by Loyd v. United States, 153 F.Supp. 416, 139 Ct.Cl. 626 (1957) and Cruttenden v. Commissioner, 70 T.C. 191 (1978), aff'd on other grounds, 644 F.2d 1368 (9th Cir. 1981).
“Recovering property” emphatically does not mean acquiring or disposing of property, however. The Supreme Court’s opinion in Woodward v. Commissioner, 397 U.S. 572, 90 S.Ct. 1302, 25 L.Ed.2d 577 (1970), has affirmed the longstanding view of the capital nature of legal expenses associated with acquisition and disposition: “The law could hardly be otherwise, for such ancillary expenses incurred in acquiring or disposing of an asset are as much a part of the cost of the asset as is the price paid for it.” Id. at 576, 90 S.Ct. at 1305.
We also conclude that the new regulation clearly was intended to create a deduction for legal expenses associated with recoveries of “amounts of income which, if and when recovered, must be included in income.” The fees recovered by the taxpayer in Komhauser were includable income. Read narrowly, as making only the cost of such recoveries deductible, the Komhauser holding has weathered subsequent legal developments intact. As we shall discuss below, this principle is consistent with the most authoritative current interpretations of the Code.
[1273]*1273The intended meaning of the reference to “investment property” is less clear. Komhauser did involve a recovery of fees in the form of stock. One might well read the parenthetical clause as the Commissioner urged in Cruttenden, that is, as allowing deduction only of expenses for the recovery of includable income, whether in the form of cash or in some other form — such as stock or other investment property.
On the other hand, such a reading would ignore not only the syntax of the clause but the broader reading given to Komhauser by some courts. For example, in Commissioner v. Speyer, 77 F.2d 824 (2d Cir.1935), the Commissioner had determined:
that a portion of the attorney’s fees allocable to what he concedes to be the taxpayer’s capital in the awards is not deductible as an expense for the reason that it is a capital charge. This contention was rejected by the Board of Tax Appeals, which held that the full amount was deductible from gross income in accordance with the provisions of section 23(a) of the revenue act of 1928, 26 USCA section 2023(a).
Id. at 826. The Second Circuit affirmed the Board of Tax Appeals on the strength of a citation to Komhauser, concluding:
The proceedings ... were for the purpose of recovering assets for the taxpayer. It is not questioned but that the attorneys were paid the amounts claimed in connection with the transaction of this business. Such fees are deductible as an expense of the business.
Id.
Similarly, in an influential case which arose after the new regulations were published, the Supreme Court rejected the argument that the new non-business expense deduction provisions must be read “as requiring that expenses be paid for the production of income in order to be deductible.” Bingham’s Trust v. Comm’r., 325 U.S. 365, 373, 65 S.Ct. 1232, 1237, 89 L.Ed. 1670 (1945). Citing Komhauser, the Court stated:
Section 23(a)(2) is comparable and in pari materia with section 23(a)(1), authorizing the deduction of business or trade expenses. Such expenses need not relate directly to the production of income for the business. It is enough that the expense, if “ordinary and necessary,” is directly connected with or proximately results from the conduct of the business.
Id. at 373-74, 65 S.Ct. at 1237.
Thus, if we read the regulation as an attempt to reflect the state of the law circa 1942, the more plausible reading of the regulation’s reference to “investment property” is that it means just what it appears to mean. That is, the costs of recovering investment property (which apparently means nothing more nor less than property held for the production of income such as is treated by I.R.C. § 212) are deductible, regardless of whether that property is includable in gross income upon recovery.
VII.
Read in their historical context, the first two clauses of Treas.Reg. § 1.212-l(k), the “defending or perfecting title” clause and the “recovering property” clause, are entirely self-contradictory. Because the phrase “recovering property” encompasses not only recoveries of possession but all cases of “defending or perfecting title,” and because the references to “investment property” and includable income together encompass the entire scope of I.R.C. § 212, the second clause, if given full effect, would deprive the first clause of any effect.
In the Komhauser case, the only limitation placed upon the deductibility of recoveries of investment property was a proximate relationship with the business enterprise. In the non-business context, this limitation drops out, leaving no workable limitation on the principle whatsoever. As the Second Circuit has observed, in upholding the “defending or perfecting title” language in the regulation:
Obviously, in every case a taxpayer who succeeds in retaining his property against an unjustified attack “conserves” it in a certain sense. But “conservation” must have a narrower meaning than this. The statute and the regulation exist side-by-side and it would make no sense to [1274]*1274say that every case which falls within the regulation also comes within the statute.
Galewitz v. Comm’r., 411 F.2d 1374, 1378—79 (2d Cir.1969). Similarly, the Fifth Circuit has observed:
Of course, when property is held for the production of income, any expenditure which relates to the perfection or defense of the taxpayer’s title to the property can, in one sense, be said to have been an expenditure for the conservation of the property. On the other hand, it seems equally as sound to say that any such expenditure is more properly an element in the cost of acquiring or preserving the legal right to the income produced by the property, and therefore must be considered a capital expenditure.
Brown v. Comm’r., 215 F.2d 697, 699 (5th Cir.1954).
We cannot uphold both the nondeductibility of the cost of defending or perfecting title and the deductibility of the costs of recovering investment property. Both rules, however, are apparently authoritative. The “recovering property” rule has been in the regulations for over forty years and reflects Supreme Court precedent. See Bingham's Trust v. Comm’r., 325 U.S. 365, 65 S.Ct. 1232, 89 L.Ed. 1670 (1945); Kornhauser v. United States, 276 U.S. 145, 48 S.Ct. 219, 72 L.Ed. 505 (1928). On the other hand, the “defending or perfecting title” rule has been in the regulations for roughly seventy years and was well-accepted by the same courts which followed the Kornhauser rule. See, e.g., Levitt & Sons, Inc. v. Nunan, 142 F.2d 795, 797 (2d Cir.1944).
While the rules may have been equally authoritative in 1942, however, when this regulation was first published, their respective fates in the ensuing four decades have been strikingly different. While both rules appear side-by-side in the regulations, the “defending or perfecting title” language has been applied repeatedly to facts to which the other language could readily have been applied. See, e.g. Burch v. United States, 698 F.2d 575 (2d Cir.1983); Galewitz v. Comm’r., 411 F.2d 1374 (2d 1969); Lucas v. Comm'r., 388 F.2d 472 (1st Cir.1967); Brown v. Comm’r., 215 F.2d 697 (5th Cir.1954); Safety Tube Corp. v. Comm’r., 168 F.2d 787 (6th Cir.1948). At the same time, the “recovering property” language has been almost completely ignored. Once one sets aside Cruttenden v. Commissioner, 70 T.C. 191 (1978), aff'd on other grounds, 644 F.2d 1368 (9th Cir. 1981), the reasoning of which was disapproved by the Ninth Circuit, and Loyd v. United States, 153 F.Supp. 416, 139 Ct.Cl. 626 (1957), where the “investment property” language was quoted almost as an afterthought, it becomes apparent that no court has ever relied upon the “recovering property” language in allowing a deduction.
Furthermore, while in recent years the courts have seen the need to limit the “defending or perfecting title” rule, they have not limited the rule in the degree or in the manner suggested by the regulation’s “recovering property” clause. In Woodward v. Commissioner, 397 U.S. 572, 90 S.Ct. 1302, 25 L.Ed.2d 577 (1970), the Supreme Court observed:
In one sense, any lawsuit brought against a taxpayer may affect his title to property — money or other assets subject to lien. The courts, not believing that Congress meant all litigation expenses to be capitalized, have created the rule that such expenses are capital in nature only where the taxpayer’s “primary purpose” in incurring them is to defend or perfect title.
Id. at 576, 90 S.Ct. at 1305 (footnote omitted). In response to Woodward, which established the “origin of the claim” test for cases of acquisition or disposition, the Tax Court in Reed v. Commissioner, 55 T.C. 32 (1970), adopted the “origin of the claim” test for “defense or perfection of title” cases.
The courts have consistently interpreted the Tax Code to dictate that the costs of defending or perfecting title are non-deductible capital expenditures. The only limitation upon this rule recognized by recent cases following Woodward and Reed is that litigation expenses may be deducted [1275]*1275when the “origin of the claim” is not the defense or perfection of title. This rule is not limited by a sweeping exception for recoveries of investment property. Because, as originally written, the regulation was intended to create such an exception, we cannot construe this language narrowly to refer to anything less than all investment property. Instead, we must conclude that, whatever authority the regulation appeared to have when first published, the “investment property” language is not authorized by any current interpretation of the Internal Revenue Code.
The current status of the includable income language is less clear. There is precedent, albeit somewhat dated precedent, in this circuit apparently denying a deduction for the costs of recovering income on the ground that “the gist of the controversy is the right to the asset which produced the income.” Safety Tube Corp. v. Comm’r., 168 F.2d 787, 790 (6th Cir. 1948). The Safety Tube Court distinguished Komhauser on the spurious ground that “the litigation here struck at the very ownership of the patent itself.” Id. More recently, however, in applying the Supreme Court’s “origin of the claim” rule, the Tax Court has recognized that:
Situations arise where the litigation is rooted, in part, in the defense or perfection of title (nondeductible) and,- in part, in the collection of income or for the management of a property (deductible). In these circumstances, section 1.212-l(k), Income Tax Regs., contemplates that an allocation should be made____
Boagni v. Comm’r., 59 T.C. 708, 713 (1973).
We consider Safety Tube to be distinguishable. First, in Komhauser, the litigation concerned title to includable income, whereas in Safety Tube the litigation concerned title to a patent which had produced royalties. Second, the “origin of the claim” test is more flexible than the “primary purpose” test applied by this court in Safety Tube. Where the “primary purpose” of litigation was to determine ownership of a capital asset, it became irrelevant whether another motivation for the suit was to recover includable income. The “origin of the claim” test, however, permits courts to recognize that litigation may have more than one origin.
We therefore approve of the regulation’s language to the extent that it permits a deduction for recoveries of “amounts of income which, if and when recovered, must be included in gross income,” as that language was interpreted by the Tax Court in Boagni.
VIII.
The Tax Court held that the origin of the claim in both the finance company stock litigation and the bank stock litigation was title to the stock. Accordingly, the Tax Court denied any deduction for the legal expenses associated with recovery of the stock. The Tax Court gave no effect to the “investment property” language of Treas.Reg. § 1.212-l(k). Because we have concluded that the “investment property” language of the regulation is not the law, we agree with the Tax Court’s resolution of this issue.
The Tax Court also failed to give effect to the regulation’s includable income language. The Commissioner apparently concedes the authority of Boagni v. Commissioner, 59 T.C. 708 (1973), which approved the includable income language and proceeded to allocate legal expenses between those attributable to the recovery of income derived from a capital asset and those attributable to litigation over title to the capital asset. Only the latter expenses were considered nondeductible. The Commissioner seeks to distinguiish Boagni, however, arguing that “no part of taxpayers’ 1977 litigation was rooted in the collection of accrued income.”
The Commissioner’s argument apparently rests upon the fact that in Boagni there were two discrete legal proceedings brought by different plaintiffs, whereas in this case taxpayers’ claims for dividends and interest (none of which was owing when the litigation began) were subsumed in their suits over title to the stock. We find this distinction unpersuasive. The origin of taxpayers’ claims was an attempt to collect income precisely to the extent that [1276]*1276stock dividends and interest were recoverable in their suits. The best objective evidence of a concern over income would not be the filing of a separate suit, which, in any event, was not required here. Instead, the best evidence is the amounts of income actually recovered.
We conclude that the Tax Court clearly erred in failing to apply the includable income language in the regulation and that this is a case, like Boagni, in which an allocation between the expenses of defense or perfection of title and the expenses for the collection of income is both appropriate and feasible. Because the Tax Court failed to perform an allocation, remand to the Tax Court is necessary for determination of the appropriate method of allocation.
IX.
We AFFIRM the Tax Court decision insofar as it denied deduction for the expenses of recovering title to stock.
We REVERSE the Tax Court decision insofar as it denied deduction for the cost of recovering income. We remand to the Tax Court for determination of the proper method of allocating between deductible and nondeductible expenses.