This appeal, limited to the narrow question of counsel fees, has its roots in events long ago. In 1953, four Massachusetts *575towns in the Southern Berkshire area of the state formed a regional school district, and in 1954, a fifth town joined. The five towns are Sheffield, Monterey, Egremont, New Marlborough and Alford. The district, empowered to govern public schools within the region, was formed pursuant to state law, see 1951 Mass Acts c. 638, and was approved by the voters in each town.
At all times, the district has been governed by a ten member committee with a fixed apportionment of seats among the towns. Sheffield has four seats, Egremont and New Marlborough have two each, and Alford and Monterey have one each. Every committee member’s vote is equally weighted, so Sheffield has 40 percent of the voting power, even though its population has at times represented close to half of the district’s total population. From 1954 until very recently, each town’s members on the committee were chosen by popular vote of the residents of that town.
In 1986, this court upheld a challenge to another, similarly structured Massachusetts school district based on the Supreme Court’s one person, one vote jurisprudence. See Kelleher v. Southeastern Reg’l Vocational Technical High Sch. Dist., 806 F.2d 9, 10-13 (1st Cir.1986).1 The court held that the substantial discrepancy in voter power as among the respective towns was unlawful but could be cured in various ways, including resort to district-wide, at-large elections; weighting committee member votes according to the population that they represent; keying the number of locally elected members from each town to population; or eliminating direct popular elections and having local officials from each town appoint district members. See id. at 13 n. 17.
A few months later, state officials advised all districts of the Kelleher decision and urged them to examine their own situations; even if popularly elected, a district’s governance structure might or might not conform, depending on the deviation between each town’s population and the voting power of its own members on the district committee. For whatever reason, the Southern Berkshire district involved in our case did nothing to alter its structure. Then, in 1993, it solicited an opinion from state officials and was told that its own district failed to comply with Kelleher. A subcommittee, consisting of representatives from all five towns, was set up to consider reforms.
For the next three years, the subcommittee considered various schemes but failed to reach an agreement. Sheffield wanted a plan that would retain local election of committee members but that would increase Sheffield’s voting power on the district committee to conform to Sheffield’s population share. The other towns preferred either district wide, at-large voting (which would increase Sheffield’s power) or the appointment of the district committee members by elected bodies in each town (which would moot the equal protection issue) — both solutions that Sheffield opposed.
Dissatisfied with the status quo, the Town of Sheffield and a group of its residents (including members of the town’s board of selectmen) brought the present suit in 1996 against the district and the other four towns. The suit, brought in federal district court under 42 U.S.C. § 1983, sought a declaration that the present scheme was unconstitutional and in-junctive relief to prevent new elections under the present regime and to require a new court-approved structure for elections. The complaint also sought attorney’s fees for the plaintiffs under 42 U.S.C. § 1988.
*576Sheffield quickly moved for and received partial summary judgment from the district court declaring the existing scheme unconstitutional. The defendants did not dispute that the scheme was unconstitutional but questioned the need for a declaration since the parties had been working on that assumption in seeking to negotiate a solution. Just before this ruling, Egre-mont proposed a compromise' — that the present committee structure be retained but that the legal infirmity be remedied by having a local elected entity in each town select the town’s committee members in place of popular elections. Sheffield rejected the proposal.
In early 1998, the parties reached agreement without further litigation, and the district court approved the settlement. The final scheme was that earlier urged by Egremont — retaining the present committee structure and allocation of voting power but eliminating direct popular election of the committee — with one variation: Egremont had initially proposed that local popularly elected school boards select the committee members, while the final scheme let the local board of selectmen choose the committee members for the town. Thus Sheffield retained its original four seats and 40 percent of the voting power on the committee, but its members were now chosen without popular election.
Some might think this an odd triumph for equal protection, possibly raising a doubt or two about the Supreme Court jurisprudence underpinning our decision in Kelleker, cf. id. at 13 (Campbell, C.J., concurring), but that is not our concern. Rather, the present appeal grows out of the plaintiffs’ post-settlement request for attorney’s fees under section 1988. The district court denied that request on the ground that Sheffield was not a “prevailing party,” an explicit condition of such an award under the statute. See 42 U.S.C. § 1988(b) (“[T]he court, in its discretion, may allow the prevailing party ... a reasonable attorney’s fee as part of the costs.... ”). It is this ruling that Sheffield now seeks to overturn.
The usual “American rule” is that parties bear their own counsel fees, see Alyeska Pipeline Serv. Co. v. Wilderness Soc’y, 421 U.S. 240, 247, 95 S.Ct. 1612, 44 L.Ed.2d 141 (1975), but Congress may provide otherwise. Congress has sometimes directed that a prevailing plaintiff “shall” recover a reasonable attorney’s fee, e.g., 15 U.S.C. § 15(a) (antitrust cases); by contrast, in section 1983 actions, Congress has said that the court “may” make such an award to a prevailing party “in its discretion.” But the Supreme Court has read “may” to mean “must,” save for rare cases in which “special circumstances” would render an award unjust. Blanchard v. Bergeron, 489. U.S. 87, 89 n. 1, 109 S.Ct. 939, 103 L.Ed.2d 67 (1989); Hensley v. Eckerhart, 461 U.S. 424, 429, 103 S.Ct. 1933, 76 L.Ed.2d 40 (1983).
Perhaps in compensation, the Supreme Court has construed the term “prevailing party” in a more flexible fashion. For example, it has said that “a technical victory may be so insignificant ... as to be insufficient to support prevailing party status.” Texas State Teacher’s Ass’n v. Garland Indep. Sch. Disk, 489 U.S. 782, 792, 109 S.Ct. 1486, 103 L.Ed.2d 866 (1989). More recently, it ruled that a plaintiff who wins only nominal damages can “prevail” yet be entitled to no fee at all. Farrar v. Hobby, 506 U.S. 103, 113-15, 113 S.Ct. 566, 121 L.Ed.2d 494 (1992). And it has described the concept of “prevailing party” in phrases so general as to leave a good deal of latitude. E.g., id. at 111-12, 113 S.Ct. 566 (“[A] plaintiff ‘prevails’ when actual relief on the merits of his claim materially alters the legal relationship between the parties by modifying the defendant’s behavior in a way that directly benefits the plaintiff.”).
We have taken the Supreme Court’s cases to mean that, absent unusual circumstances, the plaintiff is to be awarded attorney’s fees in cases under section 1983 where it wins significant practical relief *577favorable to its position. E.g., Pearson v. Fair, 980 F.2d 37, 43-44 (1st Cir.1992). Our cases have also recognized that a plaintiff may achieve such ends through litigation, and deserve attorney’s fees, even without a formal victory; for example, the so-called “catalyst” theory might justify an award where the defendant abandoned an unlawful practice after the case was brought, as a direct result of the lawsuit, but without the need for a decree. See id. at 43-45.2 Again, the inquiry is a practical one.
Because the test of “prevailing party” is both general and practical, we have relied heavily on the informed judgment of the district courts in applying it to the widely varying facts that are presented. So long as the legal rules are properly understood, a district court’s application of them to individual circumstances is normally upheld, if there is no abuse of discretion. See Paris v. HUD, 988 F.2d 236, 238-39 (1st Cir.1993); McDonald v. HHS, 884 F.2d 1468, 1473-74 (1st Cir.1989). This is one of many instances in which reasonableness is the touchstone, reflecting the trial court’s greater familiarity with the details of the case and the absence of any special need for de novo review.
Against this background, we readily sustain the district court’s denial of attorney’s fees. Far from obtaining their basic objective in this case — increased power on the committee — the Sheffield plaintiffs got exactly the opposite solution: their votes on the committee remain four out of ten or 40 percent of the whole. True, popular election of members is gone; but this can hardly be described as a victory “for” Sheffield since its own earlier demand had been to retain popular election (while increasing Sheffield’s share of the voting power on the committee).
Sheffield describes its law suit as a “catalyst” for the change that occurred. Cf. Pearson, 980 F.2d at 43. Whether this is so in fact could fairly be debated since Sheffield was offered pretty much the same outcome soon after the lawsuit was commenced, namely, the present committee structure and apportionment of votes but a switch from popular election to appointment of committee members. Sheffield says that the earlier proposal would have had elected school boards appoint the committee members and it wanted at least to ensure the power of its selectmen to choose them; but this choice has nothing to do with protecting the one person, one vote principle for voters.
In all events, the decisive question is not whether Sheffield’s lawsuit was a catalyst but whether it was a catalyst for an outcome favorable to Sheffield. Under the catalyst theory — itself simply a practical test for discerning victory — a plaintiff may achieve its victory in a lawsuit even if there is no award or injunction, so long as the lawsuit effectively achieved a favorable result sought by the plaintiff. See Pearson, 980 F.2d at 43-45. The problem here is that, by any realistic measure, the result achieved was not that sought by Sheffield.
Nor does it help Sheffield’s claim to point to the interim declaratory ruling by the district court that the preexisting scheme violated the one person, one vote principle. No one in the litigation appears seriously to have doubted that the existing regime had to be changed; indeed, the parties had been negotiating on this premise well before the lawsuit began. In a number of voting rights cases, the courts have declined to award attorney’s fees *578where efforts to amend the unlawful scheme began prior to the filing of suit by the plaintiff.3 A party is not entitled to attorney’s fees when he obtains relief through litigation that he would have obtained anyway in the absence of bringing a lawsuit. See New Hampshire v. Adams, 159 F.3d 680, 687 (1st Cir.1998).
In any event, even if the declaratory ruling sped up the negotiations, which may or may not be so, the question remains whether the Sheffield plaintiffs ultimately prevailed in obtaining relief in their favor, which they did not. As for town voters throughout the district, the main outcome of the case is that they no longer have any direct vote for committee members. This is hardly a compelling case for awarding attorney’s fees to a plaintiff.
Affirmed.