GALLAGHER, Associate Judge:

The trial court ordered the District of Columbia (the District) to pay the fees and expenses of the attorneys who represented appellees, the successful taxpayer-litigants in District of Columbia v. Green, D.C.App., 310 A.2d 848 (1973) (Green I), and District of Columbia v. Green, D.C.App., 348 A.2d 305 (1975) (Green II). The District appeals, presenting this court with the issue of whether an award of attorneys’ fees 1 fits within either of the two recognized exceptions to the general American rule which, in the absence of statutory authorization, prohibits an award of attorneys’ fees to a successful litigant.

In Green I the appellees successfully sued the District to prevent an illegal tax assessment as applied to certain single-family residential properties and to enforce the District’s compliance with the District of Columbia Administrative Procedure Act’s rulemaking provisions in changing the assessment level applicable to such residential properties. As a result of that suit, 77,485 single-family residential property taxpayers received real property tax bills for fiscal year 1974 based at a level of assessment of 55% of fair market value, rather than the illegally adopted 60%, which the District had attempted to use. 2

Green II arose from the District’s attempted compliance with Green I in adjusting the affected taxpayers’ bills to reflect properly the 55% assessment level, rather than 60%. The same taxpayer-litigants from Green I again successfully sued the *580 District, this time because of the District’s computational method for adjusting the tax bills. As a result of this suit the trial court ordered the District to use appellees’ more precise method of calculation, resulting in higher tax bills for some of the affected single-family residential taxpayers, and in lower tax bills for others in the same class. We affirmed. Green II, supra.

In this case the trial court considered the issue of whether appellees, after two such successful suits against the District, could recover an award for their attorneys’ fees and expenses. After noting the general prohibition of awards of attorneys’ fees to prevailing litigants under the American rule, the trial court decided that the appel-lees were entitled to such an award under the so-called common fund or common benefit exception. 3

The common fund exception began with Trustees v. Greenough, 105 U.S. 527, 26 L.Ed. 1157 (1882), in which the Supreme Court recognized

the historic power of equity to permit the trustee of a fund or property, or a party preserving or recovering a fund for the benefit of others in addition to himself, to recover his costs, including his attorneys’ fees, from the fund or property itself or directly from the other parties enjoying the benefit. That rule has been consistently followed. Central Railroad & Banking Co. v. Pettus, 113 U.S. 116, 5 S.Ct. 387, 28 L.Ed. 915 (1885); Harrison v. Perea, 168 U.S. 311, 325-26, 18 S.Ct. 129, 42 L.Ed. 478 (1897); United States v. Equitable Trust Co., 283 U.S. 738, 51 S.Ct. 639, 75 L.Ed. 1379 (1931); Sprague v. Ticonic National Bank, 307 U.S. 161, 59 S.Ct. 777, 83 L.Ed. 1184 (1939); Mills v. Electric Auto-Lite Co., 396 U.S. 375, 90 S.Ct. 616, 24 L.Ed.2d 593 (1970); Hall v. Cole, . . . [412 U.S. 1, 93 S.Ct. 1943, 36 L.Ed.2d 702 (1973)]; ....

Alyeska Pipeline Co. v. Wilderness Society, 421 U.S. 240, 257-58, 95 S.Ct. 1612, 1621, 44 L.Ed.2d 141 (1975) (hereinafter cited as Alyeska). (Further citations and a footnote omitted.) In Mills v. Electric Auto-Lite Co., 396 U.S. 375, 391-97, 90 S.Ct. 616, 24 L.Ed.2d 593 (1970), the Supreme Court extended the “common fund” doctrine to include cases where a common “benefit” was derived but no fund was created. The Court there discussed fully the principle and rationale of the original exception and its extension.

A primary judge-created exception has been to award expenses where a plaintiff has successfully maintained a suit, usually on behalf of a class, that benefits a group of others in the same manner as himself. ... To allow others to obtain full benefit from the plaintiff’s efforts without contributing equally to the litigation expenses would be to enrich the others unjustly at the plaintiff’s expense. .

The fact that this suit has not yet produced, and may never produce, a monetary recovery from which the fees would be paid does not preclude an award based on this rationale. Although the earliest cases recognizing a right to reimbursement involved litigation that had produced or preserved a “common fund” for the benefit of a group, nothing in these cases indicates that the suit must actually bring money into the court as a prerequisite to the court’s power to order reimbursement of expenses. . . . [The foundation of such power] “is part of the original authority of the chancellor to do equity in a particular situation.” Sprague v. Ticonic Nat. Bank, 307 U.S. 161, 166, 59 S.Ct. 777, 83 L.Ed. 1184 (1939).

Mills v. Electric Auto-Lite Co., supra at 392-93, 90 S.Ct. at 625. The most recent *581 Supreme Court discussion of the common fund and benefit doctrine 4 is in Alyeska, supra. The Court there related in summary fashion the characteristics of past common benefit decisions: “the classes of beneficiaries were small in number and easily identifiable. The benefits could be traced with some accuracy, and there was reason for confidence that the costs could be shifted with some exactitude to those benefitting.” Alyeska, supra, 421 U.S. at 265 n. 39, 95 S.Ct. at 1625. 5

Against this background, we must first determine whether the trial court erred in its so-called “quasi-application” of the common benefit exception. The trial court ordered the fees to be paid from the District’s general public funds, since, in its opinion, the results in both Green I and Green II benefitted all District taxpayers because of “the ultimate effect on the District of Columbia taxing system, . . . ” and the “extraordinary benefits from ‘preserving . . . democracy, not only in the immediate impact of results achieved here but in their implications for the future conduct of the [Respondents’] 6 affairs’ in taxing policies and in ‘open-disclosure’ forums.” (Opinion and Order, April 30, 1976, quoting in part from Hall v. Cole, 412 U.S. 1, 8, 93 S.Ct. 1943, 36 L.Ed.2d 702 (1973)). 7 The vindication of important public rights, a laudable purpose, to be encouraged and rewarded, was not the aim of the taxpayers in this case. Appellees stated repeatedly their purely selfish, monetary purpose (Brief for Appellees at 1, 6) and conceded, “[i]t was not a suit to vindicate anyone’s rights.” Id. at 17. Any public rights vindicated by their suits were too remote to justify an award whose burden would be borne by District taxpayers who, for example, own no real property. Furthermore, we do not think the benefits could be traced to the entire taxpaying population of the District nor that the costs could be shifted with any exactitude to that population.

In Hall v. Cole, supra, the successful litigant vindicated for each union member certain First Amendment rights to free speech — an intangible benefit, but common *582 to all union members — and thus there was a direct and immediate impact upon each union member as a result of that suit. The potential benefit to each union member was precisely the same — the expanded scope of protected free speech applied to all equally, although the actual benefit realized may depend on the extent to which each union member exercised that vindicated right. The cost of that benefit was also equally apportioned amongst the beneficiaries, since the award was paid out of the union treasury to which all had contributed and from which disbursements were for the benefit of all.

In this litigation, however, the direct and immediate benefits of Green I and Green II were to 77,485 single-family residential property taxpayers and to 37,073 single-family residential property taxpayers, respectively, in the form of reduced property tax bills for fiscal year 1974. The benefits to every other taxpayer are much more difficult to gauge, or even to identify. 8 Furthermore, by ordering the award to be paid from the general treasury of the District, the costs are not being shifted with any exactitude to those benefitting. 9 Instead of charging costs to the beneficiaries in proportion to their benefits received, the trial court in making the award from the general treasury necessarily assumes that all taxpayers benefitted equally and the costs were borne equally. Under the facts of this case — particularly the fact that taxpayers who own no property would be bearing part of the burden of the award 10 — this court cannot accept that assumption. We hold, therefore, that the common benefit exception does not authorize an award of attorneys’ fees out of the District public treasury under the circumstances presented here. 11

Next we consider the application of the common benefit exception to Green I and to *583 Green II. Due to differences in the nature of these two actions and in the evidence of record concerning each, the propriety of an award of attorneys’ fees for each case under Alyeska’s criteria is considered separately. The method and costs of administering the awards are considered subsequently.

II.

The first criterion to apply to Green I is the “small” size of the class. Unfortunately, very few of the earlier Supreme Court cases applying this doctrine specify how many class members there were. For example, in Hall v. Cole, supra, the Supreme Court awarded the successful litigant attorneys’ fees against the Seafarers International Union of North America-Atlantic Gulf, Lakes and Inland Waters without noting how many union members there were. In Mills v. Electric Auto-Lite Co., supra, the Court awarded attorneys’ fees to the successful litigants against the corporation, but did not specify the number of shareholders benefitted. There are a number of post-Alyeska federal circuit decisions which have considered the application of the common benefit exception. Those decisions which have awarded attorneys’ fees under this exception generally involved unions or shareholder derivative suits and usually have not specified the precise number of class beneficiaries. See, e. g., Ross v. International Brotherhood of Electrical Workers, 544 F.2d 1022 (9th Cir. 1976); Usery v. Teamsters Local 639, 177 U.S.App.D.C. 222, 543 F.2d 369 (1976), cert. denied, Teamsters Local 639 v. Marshall, 429 U.S. 1123, 97 S.Ct. 1159, 51 L.Ed.2d 573 (1977); Harrison v. United Transportation Union, 530 F.2d 558 (4th Cir. 1975), cert. denied, 425 U.S. 958, 96 S.Ct. 1739, 48 L.Ed.2d 203 (1976); McDonald v. Oliver, 525 F.2d 1217 (5th Cir. 1976), cert. denied sub nom. International Longshoremen’s Local 795 v. McDonald, 429 U.S. 817, 97 S.Ct. 61, 50 L.Ed.2d 78 (1976); National Treasury Employees Union v. Nixon, 172 U.S.App.D.C. 217, 521 F.2d 317 (1975) (approximately 3Vh million employees as class members) (hereinafter cited as NTEU v. Nixon).. Thus, the maximum class size specifically noted in a post-Alyes-ka common benefit case is 3V2 million beneficiaries. NTEU v. Nixon, supra. Green I involves 77,485 single-family residential taxpayers as the class of direct beneficiaries. The size of the benefitted class in Green I appears to be wieldy enough administratively to satisfy the common benefit rule. The precedents do not indicate otherwise.

The next criterion to satisfy is the susceptibility to identification of the members of the benefitted class. As discussed previously, at least two of the pre-Alyeska Supreme Court decisions — Hall v. Cole, supra, and Mills v. Electric Auto-Lite Co., supra —and all of the post-Alyeska circuit court decisions that have awarded attorneys’ fees to the prevailing litigant under the common benefit exception have involved successful suits against a union or shareholder derivative suits. 12 The awards have been made from the union or corporate treasury in each of the cases and thus had the effect of indirectly burdening each of the benefitted class members with a proportionate share of the attorneys’ fees. In Green I the identification of class members is not so automatic, but presumably the beneficiaries can be identified from records of single-family homeowners for fiscal year 1974. 13 Compared to NTEU v. Nixon, supra, a post- Alyeska decision in which 3V2 million federal government workers were considered a sufficiently ascertainable class, the class of 77,485 affected taxpayers of Green I should present no insurmountable problems. In contrast, two post-Alyeska decisions which rejected the application of the common benefit doctrine in part because of the difficul *584 ty in identifying the class members involved factual situations in which there were no records or other information available. See Burroughs v. Bd. of Trustees of Pension Trust Fund for Operating Engineers, 542 F.2d 1128, 1132 (9th Cir. 1976), cert. denied, 429 U.S. 1096, 97 S.Ct. 1113, 51 L.Ed.2d 543 (1977) (no records maintained which would have revealed either the number or identities of benefitted persons); Satoskar v. Indiana Real Estate Commission, 517 F.2d 696, 698 (7th Cir. 1975), cert. denied, 423 U.S. 928, 96 S.Ct. 276, 46 L.Ed.2d 256 (1975) (class of resident aliens in Indiana presenting problems of indefiniteness).

The next criterion to be applied to Green I is whether the benefits can “be traced with some accuracy.” Alyeska, supra, 421 U.S. at 265 n.39, 95 S.Ct. 1612. The class of 77,485 single-family residential taxpayers benefitted directly from appellees’ “successful suit in that their real property tax bills were based, for Fiscal Year 1974, at a level of assessment of 55% instead of 60%, this latter the percentage that would have quietly gone into effect by July 1, 1973 had it not been for . . . [appellees’] action.” Opinion of Trial Court at 36. The total tax savings to the class as a result of their action was estimated at $3,775,298.16. 14 Although the appellees note that the average taxpayer savings thus was approximately $50, 15 the fact is that each taxpayer in the class actually benefitted in different amounts — and the amounts may well be substantially different, depending upon the fair market value of each property at the time of assessment for fiscal year 1974. Along with records of those benefitted taxpayers’ identities, however, there assumedly would be available such fair market values at that time in order to trace with greater accuracy the precise benefit to each taxpayer. Thus, the precise benefit in monetary savings, as a result of appellees’ actions, could be traced directly to the taxpayer benefitting from that savings — demonstrating that this criterion from Alyeska is satisfied.

The last factor found in Alyeska focuses on whether there is “reason for confidence that the costs could be shifted with some exactitude to those benefitting.” 16 Alyeska, supra at 265 n.39, 95 S.Ct. at 1625. In Hall v. Cole, supra, and Mills v. Electric Auto-Lite Co., supra, as well as in post- Alyeska circuit court decisions applying the common benefit doctrine to award counsel fees, the award has been made against the union or corporate treasuries (depending upon which is involved in the suit). The beneficiaries were the union members, who had presumably contributed dues to the union treasury, or shareholders, who had presumably contributed something of value to the corporate treasury. Thus, by awarding fees to be paid out of the treasuries, most beneficiaries, if not all, would have borne the costs of their benefits.

There may, of course, be some taxpayers who have sold their properties and are difficult, or impossible, to locate for purposes of contribution. This fact alone should not defeat an award here where the appellees have still benefitted a large class of taxpayers, who can be located. The shifting of the costs with some exactitude to the beneficiaries would still be accomplished. 17 Thus, the fact that not all beneficiaries can be located and charged for their proportionate share of the costs certainly *585 should not mean that the appellees’ attorneys receive nothing, but instead simply that they may receive a partially diminished award. 18 In Green I, since the precise amount of each beneficiary’s benefits is available, and since each beneficiary is identifiable, the amount of costs which should be borne by each is determinable. 19

The Green II award presents a slightly different situation. Green II resulted in the District’s using “a method of calculation designed to reach greater, and fairer, mathematical exactitude in the computation of real property taxes.” Opinion of Trial Court at 19. By use of that method, rather than the one originally used by the District, approximately 37,073 single-family residential taxpayers were taxed about $13,468.18 less and 29,788 such taxpayers were taxed about $9,431.14 more. Green II, supra at 309. Under the District’s former method, undercharges of up to $35.98 and overcharges of up to $72.14 would have been made to that group of taxpayers. Id. Thus, the difference presented by Green II is that only a subclass of the affected taxpayers received benefits from the litigation.

Applying the common benefit criteria from Alyeska to Green II, the first issue is whether the class of beneficiaries is small in number in the sense of being wieldy administratively. Since the entire class of 77,485 single-family residential taxpayers who benefitted from Green I was found to be sufficiently wieldy, then a fortiori a subclass of that group — the 37,073 taxpayers benef itting from Green II —is wieldy enough.

The second issue is whether this group is “easily identifiable.” The District in an attempt to comply with Green I had sent out adjusted tax bills for fiscal year 1974, using the required 55% level of assessment. Green II, however, affirmed the trial court’s order requiring the District “to recompute assessed values, correct the tax roll, and reissue supplementary tax bills or refunds . . . ” using the more correct method of calculation. Green II, supra at 306. Consequently, there must be records of both the erroneous and corrected tax bills, which, when examined, would reveal the identities of those taxpayers whose bills were reduced. Those beneficiaries would seem to be reasonably susceptible of identification.

Since it is certain that as a result of appellees’ efforts in Green II, reduced tax bills were sent to the taxpayers identified above, and since the same records used for identification should reveal the precise amount of the savings (i. e., the difference between the first bill and the second), the benefits could be traced with more than “some” accuracy — the next Alyeska criterion.

To calculate how to shift the costs with some exactitude to those benefitting, the same formula discussed above in connection with Green I may be considered, among other methods. 20 This would result in indi *586 vidualized amounts payable by each beneficiary proportionate to his derived benefits. 21 As in Green I, the fact that not all of the beneficiaries can now be located for payment should not defeat appellees’ recovery, but may require the trial court’s attention on remand. 22

Concededly in the Green I and II suits no common fund has literally been created, segregated, and come into custody of this court. This court’s course would be plainer if there were some such fund consisting of the illegally imposed and collected taxes, 23 from which to order an award. The fact that appellees prevented the collection of the illegally imposed taxes, rather than required their refund after collection, should not, in principle, defeat their recov-ery. 24 Thus, the exception might be applied to this case on the basis of the tax savings realized by the affected taxpayers and the trial court might exercise jurisdiction over the District to require the collection of the appropriate fees from the benefiting taxpayers, 25 unless the trial court ascertains, on remand, that such course of action is prohibited or unwarranted.

With this foundation, a remand to the trial court for further exploration is indicated. Our conclusion is that the equities favor an award of reasonable counsel fees, if at all feasible. There are two crucial questions remaining on remand: (1) how to collect the proportionate share of attorneys’ fees from each beneficiary; and (2) whether it is practicable administrative *587 ly to do so. At this point, there is not sufficient information before us to determine that a suitable approach is not available which would neither be legally prohibited, nor confiscatory financially, nor intolerably burdensome from the administrative standpoint.

One possible method of collection would require imposing a surcharge on the bene-fitting taxpayers’ next real property tax bills. This method has some attraction, but there may be other methods available which are better from both legal and administrative perspectives. But more importantly, the legality of such a surcharge remains to be fully explored and decided at the trial court level. The District noted that such a method is not authorized by any specific statute in the D.C.Code, but neither is such a method specifically prohibited and the trial court must explore in the first instance whether its equitable powers are broad enough to accomplish this proposed method, or any others which may develop on remand. The remand should not commence with the assumption that those who received the monetary benefits are to be excluded from bearing any measure of the costs of the proposed award for counsel. 26

The second issue on remand concerns the imposition of administrative costs in carrying out the selected method of collection. If upon further exploration it should appear that to collect a sum equivalent to the amount of the proposed award 27 would result in a financial and administrative cost so burdensome as to make it realistically prohibitive, then the trial court would be required to conclude, however reluctantly, that the scales of equity were too far out of balance to permit the award. 28 “In equity, as nowhere else, courts eschew rigid absolutes and look to the practical realities and necessities inescapably involved in reconciling competing interests . . . .” Lemon v. Kurtzman, 411 U.S. 192, 201, 93 S.Ct. 1463, 1469, 36 L.Ed.2d 151 (1973). It hardly needs saying, however, that the problem should be approached with the recognition that a certain amount of administrative cost and inconvenience will necessarily be entailed. 29 That is assumed beforehand. The question is whether the scales will become unbalanced to the extent that the administrative and financial costs would be so far out of proportion to the benefit sought that, in the name of good sense, the effort should be abandoned. 30 If upon further proceedings this should appear, we do not doubt that the able trial judge will face up to it.

Remanded with instructions.

. Appellant only challenges that portion of the lower court’s order awarding compensation for the services of counsel and the related services of paraprofessionals, but not the portion of that award for costs attributable to disbursements for transcripts, depositions, and the like. Brief for Appellant at 6.

. The litigation in Green I only affected a subgroup of all single-family residential property taxpayers. The District’s illegally adopted and administered plan of reassessment used a “stair step” approach and thus was intended to raise the assessment levels to 60% from 55% over a three-year period — with another, separate group of single-family residential properties affected each year. Thus, assessment levels had been raised for only two years — affecting about 33,000 such properties in fiscal year 1973 and about 44,485 such properties in fiscal year 1974. Another 19,000 such properties remained to be assessed upward in fiscal year 1975, and thus were unaffected by Green I for fiscal year 1974. For a fuller discussion of the history of this litigation, see Green I and Green II, supra, and District of Columbia v. Keyes, D.C.App., 362 A.2d 729 (1976) (denying refunds of that portion of taxes assessed and collected for fiscal year 1973, which had been declared invalid for fiscal year 1974 in Green I).

. The other exception to the American rule— prohibiting awards of attorneys’ fees to successful litigants — permits an award of attorney’s fees to the prevailing litigant “when the losing party has ‘acted in bad faith, vexatiously, wantonly, or for oppressive reasons. . . ’ F. D. Rich Co. [v. U. S. ex rel. Industrial Lumber Co.], 417 U.S. [116], at 129, 94 S.Ct. 2157, 40 L.Ed.2d 703 (citing Vaughan v. Atkinson, 369 U.S. 527, 82 S.Ct. 997, 8 L.Ed.2d 88 (1962); cf. Universal Oil Products Co. v. Root Refining Co., 328 U.S. 575, 580, 66 S.Ct. 1176, 90 L.Ed. 1447 (1946).” Alyeska Pipeline Co. v. Wilderness Society, 421 U.S. 240, 258-59, 95 S.Ct. 1612, 1622, 44 L.Ed.2d 141 (1975).

. Hereinafter this exception will be referred to solely as the common benefit doctrine for the sake of convenience.

. This characterization must be read in the context of the discussion — a criticism of Justice Marshall’s concept of a private attorney general doctrine as proposed in his dissenting opinion. Thus, immediately after listing those factors, Justice White (for the majority) commented on the difficulty of applying them to Alyeska’s facts, which presented substantially different problems than are presented here. In particular he pointed out the difficulty of gauging benefits and apportioning costs among the beneficiaries of the Alyeska suit. There the class of beneficiaries consisted of the entire United States population and the intangible nature of the benefits received — “ensuring the proper functioning of our system of government,” Wilderness Society v. Morton, 161 U.S. App.D.C. 446, 456, 495 F.2d 1026, 1036 (1974) (en banc), rev’d, Alyeska, supra — made measuring the benefits virtually impossible. Not being able to measure benefits would make allocation of the burden of the costs amongst beneficiaries also unattainable. Green I and Green II, however, involve a significantly smaller class of beneficiaries as well as readily determinable benefits and apportionable costs. See discussion infra.

. This refers to appellants here.

. In a suit seeking an award of attorneys’ fees and expenses from the District’s treasury, the court must be strict in its application of the common benefit exception — ensuring that each taxpayer receives a clearly recognizable and traceable, personal benefit. See, e. g., Mandel v. Hodges, 54 Cal.App.3d 596, 127 Cal.Rptr. 244 (1976); Knoff v. City and County of San Francisco, 1 Cal.App.3d 184, 81 Cal.Rptr. 683 (1969); City of Hammond v. Darlington, 241 Ind. 536, 162 N.E.2d 619, rehearing denied, 173 N.E.2d 662 (Ind.1961); State ex rel. Burk v. Oklahoma City, 522 P.2d 612 (Okl.1974); Weiss v. Bruno, 83 Wash.2d 911, 523 P.2d 915 (1974). In each of those cases, attorneys’ fees and expenses were awarded from the local jurisdiction’s general public funds to the successful taxpayer-litigants under the common benefit exception, because the benefits of each litigation were monetary and inured to all of the local taxpayers within the respective taxing jurisdictions. Each suit had resulted in tax savings to all such taxpayers either from preventing the illegal disbursement of taxpayer funds or from procuring for the public treasury additional funds or property to which the local government was entitled.

. For a discussion of similar problems, see Burbank v. Twomey, 520 F.2d 744, 749 (7th Cir. 1975) (where the class of beneficiaries consisting of all Illinois prisoners, and the benefit, a change in prison procedures for the imposition of discipline, were considered too indefinite to permit recovery under the common benefit theory); Satoskar v. Indiana Real Estate Comm’n, 517 F.2d 696, 698 (7th Cir. 1975), cert. denied, 423 U.S. 928, 96 S.Ct. 276, 46 L.Ed.2d 256 (1975) (where the class of beneficiaries, consisting of all resident aliens of Indiana, and the benefit, the right of all such persons to obtain real estate licenses in Indiana, were too indefinite to justify an award of attorneys’ fees under the common benefit theory).

. “If it is to be asserted that all citizens have in fact benefited from the vindication of constitutional principles, the ‘common benefit’ [exception] would merge into the ‘private attorney general’ approach, which . . has been rejected by the Court in Alyeska Pipeline.” Satoskar v. Indiana Real Estate Comm’n, supra at 698. See also Skehan v. Bd. of Trustees of Bloomsburg State College, 538 F.2d 53, 56 (3d Cir. 1976) (rejection of common benefit exception where general public derived benefit from requiring public institutions to act in accordance with the demands of due process). But see Serrano v. Priest, 18 Cal.3d 728, 135 Cal. Rptr. 345, 557 P.2d 929 (1976), cert. denied sub nom. Clowes v. Serrano, - U.S. -, 97 S.Ct. 2951, 53 L.Ed.2d 1079 (1977) (a post- Aiyeska decision upholding the equitable power of California courts to award reasonable attorney’s fees under the private attorney general theory for vindication of public rights under California’s constitution).

. We agree with appellants that, “[a]s a member of a benefited class of taxpayers, each such taxpayer [referring to taxpayers, like appellees, who realized tax savings as a result of Green I and II] might arguably be compelled to contribute to the cost of this lawsuit on the theory that this court’s decision to effect gave rise in a fund inuring to his benefit, but that circumstance hardly justifies this court’s assessment of attorneys’ fees against respondents [appellants] and thus in effect against the many local taxpayers who have derived no pecuniary benefit from the underlying litigation.” (Opposition of Respondents to Petitioners’ Motion for Costs and Attorneys’ Fees, at 10.)

. Furthermore, Brewer v. School Bd., 456 F.2d 943 (4th Cir.), cert. denied, 406 U.S. 933, 92 S.Ct. 1778, 32 L.Ed.2d 136 (1972), upon which the trial court relied, is not persuasive precedent here and must be considered narrowly limited to its unusual facts and extraordinary equities. Brewer concerned litigation to protect a constitutional right and to implement a desegregation plan, the object of which would have been totally frustrated if the beneficiaries had been ordered to contribute to the successful litigant’s attorneys’ fees. In this case, so long as each beneficiary’s contribution to attorneys’ fees is less than their individual tax savings, the objects of the litigation in Green I and II would not be frustrated.

. See cases cited in text supra at p. 10.

. Appellees as well as the District ■ already “know precisely which taxpayers benefited. .” Brief for Appellees at 18. The fact that some of those taxpayers have probably sold their residences and are difficult, if not impossible, to locate — after having been identified — is not relevant here, but is considered in the context of another Alyeska criterion. See discussion infra.

. The projected assessed value was reduced by an estimated $113,713,800. The tax rate was $3.32 per $100 of assessed value. Opinion of Trial Court at 36.

. Brief for Appellees at 18.

. Whether shifting costs with some exactitude means ensuring that all beneficiaries, and only beneficiaries, bear the costs, or whether it means that the costs should be borne only by all those benefitting and borne only in proportion to the individual benefits received is not clear. Even assuming that this latter, stricter interpretation is correct, Green I still appears to fall within the scope of this criterion. See discussion infra.

.“The fact that a court of equity cannot achieve perfect justice does not mean it should forsake pursuit of substantial justice.” NTEU v. Nixon, supra, 172 U.S.App.D.C. at 222 n.13, 521 F.2d at 322 n.13 (discussing the fact that some beneficiaries of that litigation would “escape” bearing any of the costs).

. We leave to the trial court, on remand, the determination of how to deal with amounts due from beneficiaries who can no longer be found to contribute their share of the costs.

. Although the precise method of allocation of costs among the beneficiaries is for the trial court’s determination in the first instance, the following formula may be considered for use: with the total award of attorneys’ fees and related paraprofessional expenses as the numerator and the total taxpayer savings for all beneficiaries as the denominator, a fraction is derived which, when multiplied by each individual taxpayer’s tax savings (as a result of the reduced assessment percentage), should produce the proportionate amount of costs to be borne by each of the 77,485 beneficiaries of Green I. Thus, for example, using the figures set forth in the trial court’s opinion below: $101,005 as the attorneys’ base fee plus a 35% incentive bonus, which equals $35,351.75, plus paraprofessional expenses of $20,285.14, a total (rounded off to the nearest dollar) of $156,641 would be the numerator. The denominator would be $3,775,298 (rounded off). Applying this fraction in the suggested formula with a hypothetical beneficiary who received a $100 tax savings, this person’s share would be (3 775 298 x ?10°) $4'12 aPProximately.

.See note 19 supra. Thus, a fraction can be derived by using the total amount of attorneys’ fees for Green II as the numerator and the total amount of tax savings for all the Green II beneficiaries as the denominator and multiply *586 ing this fraction by the amount of individual tax savings realized by each such beneficiary.

. For example, the lower court awarded $7,000 in attorneys’ fees plus a 35% incentive bonus — which equals $2,450 — for a total of $9,450 (no paraprofessional expenses incurred here) for appellees’ successful efforts in Green II. The total tax savings to the benefitted class equalled $13,468 (rounded off to the nearest dollar). Applying this fraction in the suggested formula to the highest tax overcharge of $72.18 181 tlie ProPor-(13,468 ' tionate share of costs to be borne by the taxpayer overcharged $72.18 in Green II would be $50.65.

. See notes 17 and 18 supra.

. See, e. g., Tenney v. City of Miami Beach, 152 Fla. 126, 11 So.2d 188 (1942); Leader v. Cullerton, 62 Ill.2d 483, 343 N.E.2d 897 (1976); City of East Peoria v. Tazewell County, 60 Ill.2d 263, 327 N.E.2d 331 (1975); Flynn v. Kucharski, 59 Ill.2d 61, 319 N.E.2d 1 (1974); Gamboni v. County of Otoe, 159 Neb. 417, 67 N.W.2d 489 (1954). In each of those cases except Tenney, the fees were to be paid out of existing common funds, which consisted of illegally collected taxes. In Leader, supra, and Gamboni, supra, not all benefitted taxpayers had contributed to the existing common funds and the courts in each case did not make any orders respecting contribution from those other benefitted taxpayers — a result which does not completely square with the common benefit rationale. That result stems from a strict construction and application by those courts of the common fund doctrine — awarding fees and expenses solely out of the common fund already created and preserved. A more equitable result would have been to design some method for collecting all other beneficiaries’ prorata shares. For example, in Tenney, supra, the taxpayer’s suit had successfully voided certain assessment liens placed on numerous properties, and the court ordered that the voided liens would be cancelled individually on application to the court and upon payment by each such applicant of his/her proportionate share of attorneys’ fees and expenses.

. But see, e. g., Von Holt v. Izumo Taisha Kyo Mission of Hawaii, 44 Haw. 147, 260, 365, 355 P.2d 40 (1960); Hamer v. Kirk, 64 Ill.2d 434, 1 Ill.Dec. 336, 356 N.E.2d 524 (1976); Hoffman v. Lehnhausen, 48 Ill.2d 323, 269 N.E.2d 465 (1971); Appeal from Auditor’s Report of Borough of Exeter, 178 Pa.Super. 75, 113 A.2d 349 (1955). In each of those cases, despite conferring definite, monetary benefits on a class of taxpayers, the successful taxpayer-litigants were denied attorneys’ fees under the common benefit doctrine because no literal, segregated fund under the control of the court existed. We do not favor this rigid approach.

. As Justice Marshall stated, in discussing Mills v. Electric Auto-Lite Co., supra, in his dissenting opinion in Alyeska, supra, 421 U.S. at 277, 95 S.Ct. at 1631, the common fund exception “could apply, the Court wrote, where there was no fund at all, . . . [Mills, supra] 396 U.S. at 392, 90 S.Ct., at 625, but simply a benefit of some sort conferred on the class from which contribution is sought. Id. at 393-94, 90 S.Ct. at 626. As long as the court has jurisdiction over an entity through which the contribution can be effected, it is the fairer course to relieve the plaintiff of exclusive responsibility for the burden.”

. But see note 30, infra, concerning de minim-is considerations. See generally NTEU v. Nixon, supra, 172 U.S.App.D.C. at 221-23, 521 F.2d at 321-23, for a discussion of how that court handled a similar problem of dealing with administrative costs.

. The proposed award consists partially of a 35% incentive bonus which should be eliminated altogether in both Green I and II. Though there have been numerous comments at the various stages of this proceeding on the motivation and public purpose of this litigation, as we have stated earlier, counsel for appellees candidly set the record straight in this appeal and made the earthy statement that this “was not a suit to vindicate anyone’s rights.” Brief for Appellees at 17. Their purpose was admittedly only for selfish, monetary gains. See id. at 1, 6. Thus, an incentive bonus appears unwarranted. Furthermore, the incentive bonus, when added to all of the other fees and costs involved in the collection process, may contribute substantially to bring the costs too far out of proportion to the benefits gained. When the initial amount found to constitute reasonable compensation for attorneys’ fees is a large proportion of the total recovery, then the court may conclude that an increased allowance for the contingent nature of the fees should be minimal or that there should be no increase at all. See Lindy Bros. Builders, Inc. v. American Radiator & Sanitary Corp., 487 F.2d 161, 168 (3d Cir. 1973).

. If at an earlier stage there had been created a common fund by means of escrow deposits, or some other way, there might well not be any impediment to an award.

. Nothing in this opinion should be construed as inhibiting a settlement such as occurred on remand in NTEU v. Nixon, supra, on remand, No. 1787-72 (D.D.C. May 28, 1976).

. It strikes us that at the beginning of the exploration a rule should be adopted to exclude from consideration property owners with amounts involved so small as to be regarded as de minimis and too costly administratively for involvement as a practical matter.