OPINION

JOHNSON, J.

In this case we issued a published opinion and remanded for the trial court to consider whether attorney fees should be awarded to the prevailing party under the "private attorney general" statute (Code Civ. Proc., § 1021.5). The trial court denied the request for fees and the prevailing party in the earlier appellate proceeding appeals that denial. (1a) We conclude an appellate court owes only limited deference to a trial court determination on this issue when the successful legal action resulted in a published appellate opinion and, in any event, find this trial court had "no reasonable basis" for denying a fee award to the prevailing party for its efforts in the appellate court and thus abused its discretion. Consequently, we reverse that portion of the judgment which denies attorney fees to appellant for its work on the first appeal.

FACTS AND PROCEEDINGS BELOW

This is the second time this case has been before our court. In 1985 we reversed the trial court for denying the appellant Los Angeles Police Protective League (League) a writ of mandate against the City of Los Angeles (City). ( Los Angeles Police Protective League v. City of Los Angeles (1985) 166 Cal.App.3d 55 [ 212 Cal.Rptr. 251].) The underlying case arose when the Los Angeles City Council proposed a new $5 a month parking charge on city employees in a city-leased parking lot. They did this without complying with the "meet and confer" provisions of the Meyers-Milias-Brown Act (Gov. Code, § 3500 et seq.). Some of the affected employees were members of the League which filed an "unfair employee relations practice charge" with the City employee relations board (ERB). The ERB agreed with the League and ordered the City to cease and desist from charging the fee, to make whole the employees who had been paying the fee, and to meet and confer with the League before reimposing the fee.

Instead of complying with the ERB order the city council elected to continue imposing the parking fee and refused to "meet and confer" with the League and other employee groups. The League sought a writ of mandate to compel compliance with the ERB order and also sought an attorney fee award. The trial court denied the writ. As a consequence, it failed to reach the attorney fee issue.

On appeal the City took the position it need not comply with orders of the ERB and other City boards and commissions and furthermore that it could not seek a writ to challenge a decision of its own board because the City and its boards are all part of a single entity. We reversed denial of the writ of mandate. We held the City had no power to refuse to obey the ERB order. If it disagreed with the order the City's only recourse was to challenge the ERB order through a section 1094.5 mandate proceeding. Since it had failed to do so the order was binding.

Because the trial court had not reached the attorney fee issue we remanded the case to that court to determine whether the League was entitled to its attorney fees for the trial and appeal of its writ of mandate petition. On July 30, 1985, the League moved for proceedings on remand including a request for attorney fees — $28,120.50 — for the cost of the first trial and appeal of the petition. The League offered to waive the attorney fees required to litigate the attorney fee issue if extensive briefing and discovery were not necessary. (However, since it has been forced to expend considerable sums on the instant appeal this waiver offer is no longer in effect.) On September 4, 1985, the court denied the motion for an attorney fee award, ruling the League's legal action, although successful, had not produced a "substantial benefit" nor "vindicated an important public right" as required by section 1021.5

DISCUSSION

The City argues the appellate court should defer completely to the trial court resolution of all four of these elements of the section 1021.5 test. It quotes Supreme Court language in Serrano v. Priest (1977) 20 Cal.3d 25, 49 [ 141 Cal.Rptr. 315, 569 P.2d 1303]: "The `experienced trial judge is the best judge of the value of professional services rendered in his court, . . . [and] will not be disturbed unless the appellate court is convinced that it is clearly wrong.'" (Italics added.) However, when deciding whether or not some attorney fee award is appropriate in the instant case the trial court was not appraising the "value of professional services" but whether any award is warranted at all. Moreover, for the most part the trial court was looking at the legal services rendered in this court, that is the appellate court, not "his court."

We first consider what is the proper scope of review of a trial court determination on the attorney fee question where the legal work produced a published appellate court opinion and then examine whether the trial court's denial of attorney fees in this case should be reversed in whole or in part.

I. PROPER SCOPE OF REVIEW WHERE LEGAL ACTION RESULTED IN APPELLATE DECISION

In this case, we remanded to the trial court for the purpose of considering, in the first instance, the request for an attorney fee award under Code of Civil Procedure section 1021.5 This was necessary because at least some of the criteria outlined in section 1021.5 entail factual determinations an appellate court is in no position to undertake. However, this does not mean we are bound by the trial court's decision about the appropriateness of an attorney fee award nor its findings of fact. The California Supreme Court's most complete statement on the degree of deference the appellate court owes to the trial court's decision and findings is found in Baggett v. Gates (1982) 32 Cal.3d 128, 142-143 [ 185 Cal.Rptr. 232, 649 P.2d 874]. (1b) "Where, as here, a trial court has discretionary power to decide an issue, its decision will be reversed only if there has been a prejudicial abuse of discretion. `"To be entitled to relief on appeal . . . it must clearly appear that the injury resulting from such a wrong is sufficiently grave to amount to a manifest miscarriage of justice. . . .'" (6 Witkin, Cal. Procedure (2d ed. 1971) Appeal, § 242, p. 4234, citations omitted.) However, `discretion may not be exercised whimsically and, accordingly, reversal is appropriate "where no reasonable basis for the action is shown." [Citation.]' ( Marini v. Municipal Court (1979) 99 Cal.App.3d 829, 835-837 [ 160 Cal.Rptr. 465]; . . . . [¶] Analysis of plaintiffs' action leads to the conclusion that there was no reasonable basis for the trial court's denial of their motion for attorney fees." (Italics added.)

What the California Supreme Court actually did in Baggett may be even more instructive than the verbal test it set out. It overturned the trial court's denial of an attorney fee award after independently making its own findings as to each of the section 1021.5 criteria and then concluding there was "no reasonable basis" for the trial court's denial of an award.

We regard it crucial to distinguish between two situations — when the trial court is considering a court-awarded fee where the litigation began and ended at the trial level as contrasted with one which resulted in an appellate decision. In the former situation, the trial court may well be in a better position than the appellate court to assess whether what happened in that court enforced an "important right affecting the public interest" and "conferred a significant benefit on the general public or a large class of persons." Accordingly, the appellate court owes the trial court a full measure of deference in deciding whether the trial court abused its discretion. But the second situation is something quite different. An appellate court is in at least as good a position as the trial court to judge whether the legal right enforced through its own opinion is "important" and "protects the public interest" and whether the existence of that opinion confers a "significant benefit on the general public or a large class of persons."

Although we reach the same conclusion under the "abuse of discretion" standard, this case highlights why the appellate court indeed appears to be in a better position to determine whether its opinion vindicated an "important right" which "affects the public interest" and yielded a "significant benefit." We have only the highest regard for the particular trial judge who examined these questions in the court below. The fact a jurist of this caliber can err suggests these particular questions are often best decided initially in the appellate court when it is the appellate court which has rendered the decision which vindicates the right and confers the benefits. (As is explored later, this court decided these two questions itself in Compton Community College etc. Teachers v. Compton Community College Dist. (1985) 165 Cal.App.3d 82 [ 211 Cal.Rptr. 231]. [See pp. 15-17, post.] Perhaps it should have done so in the instant case as well.)

It is true the prevailing standard of review speaks in the language of "abuse of discretion." (3) However, at least two of the four elements of the section 1021.5 test are issues more easily evaluated by appellate courts than by trial courts. Normally the appellate court will be in a better position to assess whether a given legal action has had a significant impact on the law. This is particularly true when the legal action produces an appellate decision favorable to the party seeking an attorney fee award. (And still more true when the Court of Appeal reversed the trial court in an appellate decision favorable to the party seeking fees.) Except in rare situations this does not require either court to make factual findings based on conflicting testimony from live witnesses of varying credibility. It is in the true sense a question of law. It makes little sense to defer to the discretion of a single trial judge who may have had to make this decision in a matter of moments on the basis of a rather cursory review of the legal field involved when the deferring body would be three judges who have already researched the legal aspects of the case in depth in order to produce a full-fledged appellate opinion on the subject. If there ever was a question an appellate court is better equipped to decide than a trial court, it is whether the appellate decision it issued "enforced an important right."

The second element of the section 1021.5 criteria likewise is often decided as well — if not better — by an appellate court as opposed to a trial court. How many people will receive what kind of benefit, and how much, as a result of a given legal action is usually more of a value judgment than an issue of fact. And most often it is a value judgment about legal effects and the like which appellate courts are well situated to make. True, it sometimes may be useful to have a trial court take evidence about the numbers of people in the population group likely to be affected by the legal action or to take testimony about the predicted quantitative impact especially where money is somehow involved. But once the figures and estimates are in, it is a legal question not a factual question whether what they portray represents a significant public benefit to a substantial segment of the citizenry. There is no good reason for an appellate court to defer to a trial court's value judgment on that question and, since numbers seldom tell the full story anyway, very good reason not to defer to the trial judge on this question.

The degree of deference we owe to the trial court as to the third element of the section 1021.5 test presents a more complex issue. Here the trial court is asked to ascertain a number of more or less quantifiable variables and then compare them.

The trial court must first fix — or at least estimate — the monetary value of the benefits obtained by the successful litigants themselves. (This ordinarily is a far different figure than the benefits conferred on the general public or similarly situated individuals and groups.) Once the court is able to put some kind of number on the gains actually attained it must discount these total benefits by some estimate of the probability of success at the time the vital litigation decisions were made which eventually produced the successful outcome. This discounting is essential because the goal is to encourage the filing and prosecution and, if necessary, appeal of cases which advance "the public interest." At the time each filing, litigation and appeal decision is being made, the litigants must discount any monetary benefits they hope to achieve by the probability of success. Thus, if success would yield them, the litigant group, an aggregate of $10,000 but there is only a one-third chance of ultimate victory they won't proceed — as a rational matter — unless their litigation costs are substantially less than $3,000.

After approximating the estimated value of the case at the time the vital litigation decisions were being made, the court must then turn to the costs of the litigation — the legal fees, deposition costs, expert witness fees, etc., which may have been required to bring the case to fruition. Ordinarily it will be easier to assign a definite figure for these costs since the litigant should be able to produce bills and payments for these sort of expenditures.

The final step is to place the estimated value of the case beside the actual cost and make the value judgment whether it is desirable to offer the bounty of a court-awarded fee in order to encourage litigation of the sort involved in this case. Once again, as pointed out earlier, a bounty will be appropriate except where the expected value of the litigant's own monetary award exceeds by a substantial margin the actual litigation costs.

(2b) All these factors under section 1021.5 are interrelated, however. Where the benefits achieved for others are very high it will be more important to encourage litigation which achieves those results. Accordingly, it will be more important to offer the bounty of a court-awarded fee than where the public benefits are less significant. Thus, the courts should be willing to authorize fees on a lesser showing of need than they might where the public benefits are less dramatic. This means the court sometimes should award fees even in situations where the litigant's own expected benefits exceed its actual costs by a substantial margin.

In contrast, where the public benefits are rather modest the courts should award fees only where the litigant's own expected benefits do not exceed its costs by very much (or possibly are even less than the costs of the litigation). To put it another way, when the ratio between public benefits and the litigant's expected benefits is high the court should award fees even though the ratio between expected litigant benefits and litigant costs is high. On the other hand, where the ratio between public benefits and expected litigant benefits is relatively low so must be the ratio between expected litigant benefits and litigant costs in order to justify a fee award.

(4a) This is not to imply these estimates and calculations must produce an all-or-nothing decision. Even though the circumstances may not justify an award of all the litigation costs the litigant incurred, they may justify some partial subsidy in order to encourage similar litigation. As an example, a group may have filed a lawsuit which at the trial level only involved the member's own rather narrow interests. Had the case been decided in their favor at that level it would not have had any significant impact beyond the litigant group itself. But having lost at the trial level the litigant group may have been forced to pursue an appeal which raised and decided an issue of considerable general importance and thus conferred benefits on a substantial portion of the public. In that instance, it might be appropriate to award the litigant group its costs and legal fees on appeal but not the expenses or legal fees involved in planning and prosecuting the trial phase of the litigation.

(5) Where a trial court acquires the necessary available evidence and uses valid methodology to arrive at the required estimates, its conclusions about the third and fourth elements of the 1021.5 test will merit deference by the appellate courts. On the other hand, except for issues such as witness credibility, the Court of Appeal ordinarily is in just as good a position as the trial judge to estimate benefits, probabilities of success, and the like. Accordingly, when an appellate court spots a questionable estimate or a faulty calculation it need not shrink from the task of correcting this element of the equation and reassessing whether that change alters the ultimate result.

In expressing the view that appellate courts often should not feel compelled to defer to trial court determinations, especially of the first two criteria in the section 1021.5 test, we are only elaborating the sentiment expressed in the rationale of Wilkerson v. City of Placentia (1981) 118 Cal.App.3d 435 [ 173 Cal.Rptr. 294]. In that case, the Fourth District reversed a lower court and itself decided the prevailing plaintiffs were entitled to an award of attorney fees under section 1021.5 (but with the amount, of course, to be determined by the trial court after remand). In justifying its decision to determine whether the section 1021.5 criteria had been met instead of deferring to the trial court, the Fourth District held: "It appears to us that where the entire basis for the decision [on the attorney fee issue] is what we have done in this opinion, we should make the decision that it qualifies as a case for granting of fees." (118 Cal.App.3d at 445, italics added.)

II. THE LEAGUE'S LEGAL ACTION SATISFIED THE CRITERIA ENTITLING IT TO AN ATTORNEY FEE AWARD FOR THE APPELLATE PHASE OF THIS LITIGATION

In appraising both the significance and importance of the rights vindicated — or more accurately recognized — by this court's appellate opinion the trial court trivialized the issues involved. The judge admitted: "It strikes me as sort of a nothing issue . . ., I didn't quite understand what the Judge was saying. . . . [It] is another instance of appellate courts . . . creating new causes of action so that we get a little more work here, because we don't have enough."

Narrowly conceived, all this court decided in its appellate opinion was that the City of Los Angeles cannot impose parking fees on police officers without "meeting and conferring" with the officers' representatives. And viewed from the trial court's perspective this may have appeared to be the major issue involved.

The trial judge did not appear to fully appreciate what was really at stake in this appeal. The most important issue was created when the City took the action of refusing to comply with the employee relations board's decision in favor of the League and by its position during oral argument that it was entitled to refuse to comply without seeking permission from the court because the City and its agency were the same entity. This principle is much broader than disputes over "meet and confer" requirements or disputes between police officer organizations and City governments. Instead it implicates the enforceability of the decisions made by a wide variety of municipal agencies, commissions and boards. It was this principle which was announced and this right which was vindicated in our decision in the first Los Angeles Police Protective League v. City of Los Angeles opinion.

Had we only decided cities must meet and confer before they impose $5 a month parking fees it is unlikely we would have published our decision. (6) The fact we or some other appellate court decides to publish an opinion does not conclusively establish the underlying action "vindicated an important right." On the other hand, it is strong evidence on that question. To meet the criteria for publication under rule 976 et seq. of the California Rules of Court, the appellate court must find that a decision announces a new rule or creates a conflict with another appellate decision or at least criticizes an existing rule. True, merely criticizing a rule or creating a conflict might not "vindicate an important right." But where as here the reason for publication of the opinion is to announce a rule not found in previously published opinions the decision clearly vindicates a right and one deemed important enough to warrant publication. Admittedly, the fact of publication does not reach the level of a "prima facie showing" the right was important. Nonetheless, it goes some distance in that direction.

(7) In any event, we find the trial court had "no reasonable basis" for finding the appellate opinion obtained through appellant's legal action failed to vindicate an "important public right." The right enforced in this case falls within the range of significance found sufficient to warrant attorney fee awards in earlier cases. (See, e.g., Baggett v. Gates (1982) 32 Cal.3d 128 [ 185 Cal.Rptr. 232, 649 P.2d 874] [Supreme Court reversed denial of attorney fee award where right enforced was that police officers cannot be demoted by City of Los Angeles without an administrative appeal]; People ex rel. Seal Beach Police Officers Assn. v. City of Seal Beach (1984) 36 Cal.3d 591 [ 205 Cal.Rptr. 794, 685 P.2d 1145] [without remand to trial court, Supreme Court reversed the lower court on its decision on the substantive issue and ruled court-awarded fees must be given to several public employee unions where appellate opinion enforced right that city must "meet and confer" before proposing city charter amendments which would have punished city employees who participated in a strike]; American Federation of Labor v. Employment Dev. Dept. (1979) 88 Cal.App.3d 811 [ 152 Cal.Rptr. 193] [court award of fees to large nationwide unions ruled appropriate where courts held unemployment insurance payments must be continued while recipient appeals unfavorable ruling].) Thus, under the standard the Supreme Court announced in Baggett v. Gates, supra, 32 Cal.3d 128, 143, we find the trial court abused its discretion with respect to this first element of the section 1021.5 test.

For similar reasons, we find the trial court abused its discretion in finding the right involved did not confer a "significant benefit . . . on . . . a large class of persons." What made the right "important" in this particular instance also meant it conferred a significant benefit on a large class of persons.

The trial court once again focused primarily on the narrow issue brought to that court when assessing whether the legal action produced a "substantial benefit." As the court conceived it, "I don't see a significant benefit has been conferred on a large class. You have apparently 76 police officers who didn't have to pay $5 a month for parking, but I don't know that that rises to the level of significance sufficient to allow the order for attorney fees under the statute."

Five dollars a month for 76 police officers may have been the main reason the League took this case to the trial court initially. And indeed if that $60 a year for 76 individuals were the only benefits involved it would be difficult to characterize what the League gained as a "substantial benefit" warranting a fee award. But for reasons suggested above, the trial court's description scarcely captures the full range of benefits and beneficiaries of the appellate opinion in this case.

The class of persons benefited is not merely 76 police officers or, for that matter, the several hundred or several thousand other City employees who have been affected by this particular parking fee issue. Rather, it extends to all employees and all citizens whose claims and rights are decided in the first instance by employment relations boards or similar boards, commissions and agencies. The primary benefit is that when they win decisions before these bodies the decision will be binding on the City unless the City takes the affirmative step — and makes the financial investment — to challenge the body's action in the courts. As a result of the decision reached in this legal action, the City cannot ignore a decision of a board or commission on a whim or for political expediency. The city council will know it must be prepared to prove in court that the board or commission erred. These are benefits which may be difficult to quantify. But they are wide-spread and "substantial."

Since the trial court determined the League's legal action did not vindicate an important public right or confer a substantial benefit on a large public or class of persons, it did not fully inquire into the remaining elements of the section 1021.5 test. However, it did have before it sufficient facts to judge whether these criteria have been met.

(8) The salient financial facts are that the League had expended $28,120.50 for the first trial and first appeal and the League's members had recovered at the most $22,800. (To the League's costs must be added the cost of trying and appealing its attorney fee request which presumably will add several thousand dollars to the total.) This disparity between financial burden and financial benefit is enough in itself to satisfy the fourth element of the section 1021.5 test. Clearly it would be inappropriate to require the attorney fees to be paid out of the proceeds of the litigation since that would completely wipe out the members' monetary recovery.

The third element requires a more complex set of estimates and calculations. We must determine whether the financial burden on the League was disproportionate to its members' stake in the case. The evaluation is further complicated by the fact the League is a union which paid for the lawyer as opposed to this merely being a class composed of 76 police officers represented by their own attorney. The League had a greater stake in this litigation than the few thousand dollars in back parking fees. The League and its other members had something to gain in its future disputes with the City by enforcing the "meet and confer" requirement. Moreover, the League and its other members gained considerably from the actual grounds the appellate court used, that is, the requirement the City comply with ERB rulings unless successful in its own court challenge to those decisions. Unfortunately, it is difficult to quantify these additional benefits to the League, especially in monetary terms.

Whatever the future benefits may be for the League and its membership they are but a small share of the similar benefits which the League's legal action conferred on similarly situated public unions, public employees, and others who deal with the municipal government. Moreover, as pointed out earlier, whatever those benefits might be for the League and its membership they must be discounted by the probability they would be attained at the time the League was making the relevant litigation decisions. In this case, these larger benefits only entered the picture when the League was deciding whether to appeal the trial court's denial of its writ petition. At this point, the chances of success were problematical and the nature and extent of the benefits the appeal actually produced not readily perceived. Accordingly, we find the costs of the appeal were disproportionate to the discounted benefits the League could reasonably perceive to be at stake at the time it was deciding whether to bring this case to the appellate level. Unless the League could look forward to an attorney fee award should it succeed it would not have made sense for it to have pursued the appeal. Thus, without the prospect of a fee award this court would not have had the opportunity to address the substantial legal issues affecting the public interest which this case posed.

The trial court itself conceded this "financial burden" element probably would have been satisfied had the 76 individual employees rather than the union been the appealing party in this case: "Well, clearly, if the petitioners — or if the prevailing parties here had been the individual parkers, that would be a good point. The prospects of getting back $5 a month or $60 a year per person at an expense of thousands of dollars wouldn't make the flame worth the candle." Actually the trial court did not reach this element because it decided the case flunked the first two elements of the section 1021.5 test. Nonetheless, the court did reveal how it would have ruled when it went on to point out this appeal was not brought by 76 individual officers: "But, having been done on this group basis by an organization which does represent them all, I don't know that you have much . . . to show here."

In its brief, the City relies heavily on similar language in this division's opinion in Compton Community College etc. Teachers v. Compton Community College Dist., supra, 165 Cal.App.3d 82, where we observed in remanding the attorney fee issue to the trial court: "(I)n this case the appellant is a union. One of the functions of unions is to provide legal counsel to enforce the terms of contracts they sign on behalf of their members. This essentially is what was involved in the instant case." (165 Cal.App.3d at p. 98.) From this language, the City argues the League as a union is not entitled to an attorney fee award in this case even though the 76 officers might have if they were proceeding as an ad hoc group of individuals rather than as members of a labor organization.

Quite often, appellate courts blame trial judges and lawyers for their failure to distill the true meaning of an ambiguous holding or rationale. However, we reviewed the above quoted language from in our Compton Community College opinion. Looking at those words in isolation, we can understand how the City and the trial court may have been misled into a belief we were holding labor unions are not entitled to fee awards when representing their members' interests in the courts. In retrospect, we should have drafted this portion of the Compton Community College opinion more carefully.

Nonetheless, read in context this language does not support the City's position in the instant case. In the preceding two paragraphs of the Compton Community College opinion we had just explained why we could without remand to the trial court determine that that case satisfied the first two criteria under section 1021.5 — it vindicated an important right affecting the public interest and conferred substantial benefits on a large class of persons. This brought us to the third criterion and we were explaining why we could not without remand decide this issue. We were contrasting a union which does have some means to employ legal counsel with a lonely individual litigant or small neighborhood organization which ordinarily would not be in a position to bring any significant piece of litigation unless an attorney fee award awaited them if they succeeded. But we did not hold that merely because a union was involved it would be foreclosed from receiving an attorney fee award. Rather we remanded the case to the trial court to take evidence and make findings on this question and to determine "whether appellant union is entitled to attorney fees and if so, how much." (165 Cal.App.3d at p. 98.)

(9) Indeed in the paragraph immediately after the language on which the City relies we listed the sort of evidence not available in the record in the Compton Community College case which we felt the trial court should examine and which, if present, would justify an attorney fee award for the union. Among the factors we catalogued are whether the legal costs were extraordinarily large or the union small, and whether the benefits to nonlitigants were large in relation to the benefits received by the union membership thus justifying the attorney fee award as a means of encouraging similar suits, and whether it was a situation where the legal costs were so high and litigants' benefits so modest there would be no net recovery or a very small one unless the union were to receive an attorney fee award. (165 Cal.App.3d at p. 98.)

It should be noted these are precisely the kinds of facts which are present in the record from below in the instant case. For instance, while in Compton Community College our appellate decision resulted in some $350,000 in backpay for members of the union and there was no evidence in the record as to how much the litigation had cost the union, in this case we know the membership's immediate economic benefits are limited to some $22,800 and the costs for the first trial and appeal (to say nothing of this second hearing and appeal attempting to obtain attorney fees) amounted to $28,120.50. So here we indeed have a situation where, by our own language from our own Compton Community College case, there is evidence in the record at this time demonstrating this to be a situation where the legal costs are so high and the litigants' benefits so modest there will be no net recovery — or a very small one — unless the other party is compelled to pick up the winner's attorney fees. Moreover, the benefits to nonlitigants are high compared to the net benefits obtained by union members. Thus our Compton Community College opinion actually supports rather than precludes an attorney fee award under the facts of the instant case.

(4b) On the other hand, we conclude the League's initial action in the trial court was designed only to serve the interests of its own membership. And more significantly, that action had it been successful at the trial level would not have conferred substantial benefits on nonmembers or vindicated an important public right. Moreover, the union had an ample economic incentive and the legal resources to take the case through the trial stage. It is at the appellate level where benefits to nonmembers were added to the equation and where the legal costs mounted to $28,120.50. Thus, it is at the appellate level the case implicated the public interest and where it might not have made economic sense to proceed without the incentive of an attorney fee award should the appeal prove successful. Accordingly, we are limiting the League's attorney fee award to the portion allocable to the appellate phase of the underlying action. The League also is entitled, of course, to an award of attorney fees for both the trial and appeal of the attorney fee issue itself.

DISPOSITION

The judgment denying attorney fees to appellant is reversed and the cause remanded to the trial court for the purpose of determining how much of the League's attorney fees are allocable to the appeal of the denial of the writ of mandate. The trial court is ordered to award appellant those legal fees on appeal of the first action and furthermore to award appellant its costs and attorney fees on the instant appeal.

Lillie, P.J., and Thompson, J., concurred.

A petition for a rehearing was denied January 16, 1987, and respondents' petition for review by the Supreme Court was denied March 11, 1987.