This case arises out of some 13,000 law suits that defendant Cataldo filed between 1970 and 1978 against the Penn Central and other railroads. Cataldo is a freight claims agent, not a lawyer. He filed claims on behalf of freight receivers. The claims almost uniformly alleged that fresh vegetables, particularly lettuce, arrived late in Boston, their condition was poor, and the market price had dropped. For many years the railroads simply settled many of these claims for a few hundred dollars (or less) each. Then, in 1978, Penn Central brought this suit against Cataldo and several receivers alleging that this freight claim and litigation activity amounted to fraud, see 7 U.S.C. §§ 499b, 499e (private right of action under Perishable Agricultural Commodities Act for dishonest practices by commodities receivers), abuse of process, unfair trade practices, see Mass.Gen. Laws ch. 93A, §§ 2, 11, unauthorized practice of law, and champerty.
The district court, sitting without a jury, tried the case for sixteen days. The evidence, for the most part, sought to answer such questions as: What was Cataldo’s system for deciding when to file claims or suits? Were the claims and suits filed unlikely to reflect legitimate legal entitlements? Was Cataldo reasonable in believing that the claims were legitimate? Did he take sufficient care to screen likely invalid from likely valid claims? The district court’s opinion reflects a careful weighing of this evidence; it reflects a realization that plaintiff bore the burden of proving its allegations; and it reaches a result that is mostly, but not entirely, in defendant Cataldo’s favor.
The court found that plaintiffs did not prove abuse of process or champerty. They did not prove fraud against Cataldo (with minor possible exceptions as to which they did not prove damages). But they did prove that Cataldo behaved unfairly, in violation of Mass.Gen.Laws ch. 93A, § 2, in one respect, namely in having an employee file directly with the railroads 1,900 claims that he later withdrew without suit. The court assessed Cataldo $76,000 in damages to reimburse Penn Central for administrative costs related to these 1,900 ‘unsuedupon’ claims.
The court also found that plaintiffs proved fraud against Community-Suffolk, one of the vegetable receivers. The fraud consisted of this receiver’s having misled Penn Central about the proceeds it obtained when it disposed of vegetables that allegedly arrived late or damaged. The court assessed Community-Suffolk $55,000 as double damages.
The court further found that Frank Infelise, a lawyer associated with Cataldo, had violated Fed.R.Civ.P. 11, for he had allowed Cataldo’s employees to sign his name to thousands of court-filed complaints that he had not examined to verify whether they stated legitimate claims. The court ordered, as a remedy, that Infelise review the many suits still on the district court’s dockets and file appropriate affidavits.
Finally, the court found for defendant Cataldo on a counterclaim, charging Heinzen, a Penn Central attorney/investigator, with improper interference in Cataldo’s efforts to settle a large batch of similar *872lawsuits against the Santa Fe Railroad. The court awarded Cataldo $17,500 damages.
Penn Central, Cataldo, and Community-Suffolk all appeal from the district court’s judgment.
I.
The major issue on appeal is that raised by Penn Central: Does the record support the district court’s findings that Cataldo’s behavior was hot (for the most part) abusive, fraudulent, unfair, or otherwise improper? We agree with Cataldo’s characterization of what Penn Central has asked us to do, namely, in the Supreme Court’s words, to
try the case de novo on the record, reject nearly all of the findings of the trial court, and substitute contrary findings of our own.
United States v. Yellow Cab Co., 338 U.S. 338, 340, 70 S.Ct. 177, 178, 94 L.Ed. 150 (1949). We also agree that the Supreme Court’s response in Yellow Cab is appropriate here, namely:
The judgment below is supported by an opinion, prepared with obvious care, which analyzes the evidence and shows the reasons for the findings. To us it appears to represent the considered judgment of an able trial judge, after patient hearing, that the [plaintiffs] evidence fell short of its allegations — a not uncommon form of litigation casualty ....
Id. at 340-41, 70 S.Ct. at 179.
We add that we have reached this conclusion after studying the parties’ briefs with care and after reading all of the nearly 4,000 pages of record material presented in the parties’ briefs and appendices. We undertook this effort because familiarity with detail seemed necessary to understand the essence of the case and the likelihood of district court error, and because we wished to understand the details, given the relationship of the case to the administration of justice within the circuit. Having satisfied ourselves that the district court was correct on these basically factual and evidentiary matters, we see no need (given the one thousand or so other cases on this court’s annual docket) to take the time to write a lengthy opinion reiterating all the facts. We shall instead simply mention a few features of the record, as we read it, which may help the parties understand why we believe the district court was correct.
First, the record shows that Penn Central tried to depict Cataldo as having filed thousands of suits, lacking in merit, primarily for their nuisance value, hoping for small settlements. Cataldo sought to depict Penn Central as having resisted thousands of likely legitimate claims, hoping through resistance to discourage receivers from pursuing proper claims for small amounts of damage. Under these circumstances, the likely merits of the claims were highly relevant to the propriety of Cataldo’s methods.
Second, two important Penn Central efforts, designed to show that many Cataldo claims likely lacked merit, ended in stalemate. For one thing, Penn Central pointed to Cataldo’s practice of filing thousands of claims based on late delivery when a carload arrived in Boston on the eighth morning after dispatch from California. Penn Central tried to show that “eighth morning delivery” was not “late.” It pointed to schedules, to transportation methods, to delivery times elsewhere, and so forth. Cataldo, however, contended that case law established a standard for lateness — called “reasonable dispatch” — that required faster delivery from California to Boston than an “eighth morning schedule” provided. See Condakes v. Southern Pacific Co., 295 F.Supp. 121 (D.Mass.1968) (citing New York, P. & N. R.R. Co. v. Peninsula Produce Exchange, 240 U.S. 34, 36 S.Ct. 230, 60 L.Ed. 511 (1916)). We do not know whether or not Penn Central would have won its point about lateness, had it decided to contest and to appeal one or more of the thousands of lateness claims based upon “eighth morning delivery.” At most, the railroad’s schedule establishes a presumptive standard for reasonable dispatch, and here the parties disputed what the operative schedule was, whether Penn Central *873adopted it in good faith, and whether it realistically reflected actual railroad practices or reasonable railroad capabilities. Given these unresolved disputes, and the habitual delays to which Cataldo pointed, we see nothing erroneous in a district court conclusion that Penn Central failed to prove Cataldo unreasonable in believing he had a valid legal claim.
For another thing, Penn Central pointed out at some length that many "damage” suits were based upon lettuces that an “inspection service” stated suffered from 13% bruising. Penn Central sought to use the facts that this number appeared with amazing regularity, that it was just above the "rejection” level for sales to Canada, that vast numbers of lettuce carloads were sold to Canada anyway, and that the service was hired by the receivers, to suggest that something very fishy was afoot (“afin”?). Cataldo, however, introduced evidence that the 13% figure arose from the method of inspection: an inspector would take, say, six lettuce cartons out of a freight car; he would examine sixteen of the twenty-four lettuce heads in each carton, and if, say, two heads out of the sixteen in a carton were bruised, he would note that the carton sample was “13%” bruised. He might well mark his report on the six cartons as, “zero, zero, 13%, 6%, 6%, zero,” denoting both the existence of one sample with 13% bruising and a 4% (four heads out of 96) bruising in the overall sample. It is not surprising that most of the cartons in such a car might still be sold to the Canadians, for most are likely all right. Yet, why is it obviously wrong for the receiver to wish to collect some small amount of damages for the carton with the two bruised lettuces — if the bruising is inordinate and harmful? How is anyone deceived, if everyone understood the system? Why could Cataldo not have reasonably believed in the legal merits of these small claims? Perhaps there are good answers to these questions. But Penn Central did not provide them to the satisfaction of the district court. The court wrote,
Cataldo may have concluded, as did F.W. Woolworth, that a nickel and dime product is worth marketing if one can develop sufficient volume.
The record adequately supports this conclusion.
Third, the other facts to which Penn Central points are insufficient to tip the scales. It notes, for example, that the amounts at stake in each suit were small, the settlements still smaller, the settlements took place out of court, there was virtually no activity (discovery or trial) on docketed suits, Cataldo took a high percentage (50%) of each settlement as a fee, an attorney played little if any role in processing the claims, Cataldo brought 6,297 claims that other claims agents had rejected. Yet, the court could find these features of Cataldo’s activity to be reasonable characteristics of a “Woolworth” style operation, insufficient to prove unlawfulness without proof that the underlying claims themselves likely lacked merit.
The district court did find fraud in Cataldo’s handling of claims and suits concerning shipments that had been consigned to the receivers for sale on a commission basis. Federal regulations require that a receiver bringing a suit involving such a shipment obtain the shipper’s consent and keep him informed of developments. 7 C.F.R. § 46.29(d). Cataldo apparently failed to abide by these regulations. In addition, he excised from the documents he presented to the railroad all indications that the shipments in question were consignment shipments. The district court found that these excisions amounted to fraud on the railroad, since they deprived it of possible defenses to the affected suits. But it declined to award any damages for the fraud or to infer from it any broader pattern of improprieties.
In light of considerable conflicting evidence concerning both Cataldo’s reasons for concealing the consigned status of these shipments and the uses to which the railroad was likely to put such information if it were provided, we agree with the district court that Cataldo’s conduct in this class of cases does not support a conclusion *874that his litigation operation was improper as a whole. And, in light of the railroad’s failure to establish that Cataldo’s violations of the federal regulations in fact would have afforded it a defense to the consigned-shipment suits — that Cataldo, for example, could not have obtained shipper consent to continue with the suits had he been required to do so — we believe its refusal to award damages for Cataldo’s excisions was within its power, particularly since the railroad made no attempt to quantify the specific injuries it suffered from the fraud. See, e.g., Newton v. Rockwood & Co., 261 F.Supp. 485, 488 (D.Mass.1966), aff'd, 378 F.2d 315 (1st Cir.1967); A. DaPrato Co. v. City of Boston, 334 Mass. 186, 134 N.E.2d 438 (1956).
Penn Central also points out that numerous claims were brought without proof in the relevant file that the vegetables were in good condition at the point of origin. And, even though the railroad’s burden in a “defective goods” case is heavy, the plaintiff must at least show good condition at origin. See, e.g., Missouri Pacific Railroad Co. v. Elmore & Stahl, 377 U.S. 134, 84 S.Ct. 1142, 12 L.Ed.2d 194 (1964). Yet, Cataldo introduced testimony by attorneys familiar with freight claims that claimants often lacked such information; they commonly could get it if necessary; and the railroad (noting the size of the claim and the expense of obtaining the information) would simply use the lack of origin inspection reports as a bargaining weapon in settlement discussions.
Penn Central points as well to Infelise’s numerous violations of Fed.R.Civ.P. 11. But these violations do not compel a district court finding of abuse of process or other unlawfulness on the part of Cataldo or the receivers. See, e.g., Jones v. Brockton Public Markets, Inc., 369 Mass. 387, 340 N.E.2d 484 (1975) (abuse of process requires wrongful purpose behind, not merely wrongful conduct of, litigation). The court, in concluding that the appropriate remedy for these violations was to require Infelise to review and certify the complaints still on the docket, acted within its lawful powers.
We have not canvassed all Penn Central’s detailed arguments and factual contentions here. But we have said enough to make our reaction to Penn Central’s factually based arguments understandable. Cataldo sought to show that he had devised a method to collect legitimate small claims from the railroad — claims that administrative costs would otherwise make uncollectible. He did not prove this. But Penn Central, according to the district court, did not prove the contrary. And the record provides the district court’s decision in this respect with adequate support.
II.
There is likewise adequate support for the district court’s finding that Cataldo acted unfairly, in violation of Mass. Gen.Laws ch. 93A, in filing 1,900 claims which he later withdrew without suit. The testimony of Cataldo’s employee Pearson suggests that these and other claims were filed on the basis of very little preliminary checking. There was also testimony suggesting that Cataldo would bring suit on any rejected claim provided his file showed the slightest basis for doing so. The court might have concluded that those rejected claims as to which he did not sue lacked any significant basis in fact, and that he therefore was unfair in submitting them to the railroads without prior checking. At the same time there was adequate support for the court’s conclusion that “the oppressive conduct was more likely slothful than wilful,” so the denial of double or treble damages was appropriate.
Penn Central goes on to argue that, in any event, principles of agency law make the receiver-defendants liable along with Cataldo for their share of the unfounded claims. The district court declined to impose liability on the receivers because Pearson (Cataldo’s employee who filed the improper claims), while “indirectly an agent of the receivers,” was “not their employee.” In our view the district court’s analysis of agency principles was incorrect.
*875Pearson’s relationship to the receivers was that of a “subagent,” i.e. “a person appointed by an agent empowered to do so, to perform functions undertaken by the agent for the principal, but for whose conduct the agent agrees with the principal to be primarily responsible.” Restatement (Second) of Agency § 5(1) (1958). As the Restatement comment observes, “The inference is that the regular employees of an agent are subagents....” Id., comment c, quoted in Rayden Engineering Corp. v. Church, 337 Mass. 652, 660-61, 151 N.E.2d 57, 62-63 (1958). See generally Seavey, Subagents and Subservants, 68 Harv.L. Rev. 658 (1955). A principal’s liability for the acts of its agents’ employees — its sub-agents — is normally the same as its liability for the acts of the agents themselves. See Restatement (Second) of Agency § 255. Hence the receiver-defendants — the principals here — are liable for Cataldo’s (and hence Pearson’s) authorized acts on their behalf.
Of course, one might argue that the principals did not authorize Cataldo to act improperly or to commit torts. Yet, as the Restatement also notes,
If a servant or other agent, authorized to do an act provided certain conditions exist, and to determine whether or not such conditions exist, does the act in the erroneous belief that such conditions do exist and thereby commits a tort, the principal is subject to liability for the act.
Restatement (Second) of Agency § 252; see also id., § 253 (stating similar principle relating specifically to “Tortious Institution or Conduct of Legal Proceedings”). Indeed, “it is common experience for a principal to be held accountable in tort for unauthorized acts of an agent not too far removed from the scope of his authority, even though, strictly, they were not authorized.” Doody v. John Sexton & Co., 411 F.2d 1119, 1122 (1st Cir.1969) (citing Robichaud v. Athol Credit Union, 352 Mass. 351, 225 N.E.2d 347 (1967)). The acts at issue here — the “improper” filing of claims by one given broad powers to decide when and how to file — fall easily within these principles.
In the absence of any indication that Massachusetts would depart from the Restatement approach in a chapter 93A suit, we conclude that the receivers share Cataldo’s liability on this count. Our calculations, based on Plaintiff’s Exhibit 1, which categorizes the thousands of claims involved in the suit, suggest that Community-Suffolk generated 49% and Mutual Produce 12% of the 1,900 unfounded claims for which the district court held Cataldo liable. We presume that these ratios should also determine their respective shares of joint liability for damages on this count, but we leave this question for final determination by the district court on remand.
III.
There is adequate support for the findings of fraud against Community-Suffolk. Cataldo typically calculated damages on late or damaged shipments by simply comparing the money the receiver obtained from resale of the carload on the open wholesale market with the United States Department of Agriculture report of typical “good-condition” resale market prices on the day the carload should have arrived. There is evidence that Community-Suffolk, knowing this, provided Cataldo with sales records that appeared to reflect open-market sales at prices below the Department of Agriculture benchmarks, but which in fact did not represent open-market sales at all. In some cases, Community-Suffolk had transferred the carloads to a different part of its own company further down the distribution chain. The ‘prices’ on these records were thus prices that Community-Suffolk set for internal accounting purposes. In other cases, it had sold entire carloads to another produce distributor at negotiated prices that did not typically reflect open-market prices. In both situations, to interpret the records as representing open-market sales could wrongly suggest “resale” losses on those sales. Evidence supports the conclusion that the records Community-Suffolk presented to Cataldo made such a *876misinterpretation likely and that Community-Suffolk knew this.
Community-Suffolk argues that, in any event, the two-year statute of limitations bars most of the damages awarded. It says that the finding of “fraudulent concealment” used by the district court to toll the statute of limitations is unsupported by the record. Unlike the federal approach, see Janigan v. Taylor, 344 F.2d 781, 784 (1st Cir.) (cause of action for fraud does not accrue until it is, or reasonably should be, discovered), cert. denied, 382 U.S. 879, 86 S.Ct. 163, 15 L.Ed.2d 120 (1965), Massachusetts law does not presume “fraudulent concealment.” See e.g., id.; Lynch v. Signal Finance Co., 367 Mass. 503, 327 N.E.2d 732 (1975); Connelly v. Bartlett, 286 Mass. 311, 190 N.E. 799 (1934); see also Mass.Gen.Laws ch. 260, § 12 (codifying fraudulent concealment rule). Rather, the Massachusetts rule dates the cause of action from the commission of the fraud, unless, after such commission, the defendant “conceal[s] the existence of a cause of action through some affirmative act done with intent to deceive.” White v. Peabody Construction Co., Inc., 386 Mass. 121, 133, 434 N.E.2d 1015, 1022-23 (1982); see Connelly v. Bartlett, supra; Premium Management, Inc. v. Walker, 648 F.2d 778, 782 n. 5 (1st Cir.1981) (dicta).
Here, however, the district court could have found that Community-Suffolk’s behavior satisfied this test. Community-Suffolk continuously submitted misleading “sales” records to Cataldo (and thus to Penn Central) up to the very eve of this lawsuit. Each submission not only involved a new fraud but also helped cover up the old fraud. Had Community-Suffolk accurately disclosed that many of its newer claims involved cars that had not been sold on the open market, the railroad might well have worked out the existence of the earlier fraud. At least it would have been alerted to the problem.
We believe this course of conduct sufficient to toll the state statute of limitations for two reasons. First, Massachusetts courts accepted a similar argument in a 1947 case, Stetson v. French, 321 Mass. 195, 72 N.E.2d 410 (1947). There an employer, over a number of years, falsely assured two of his brothers who worked for him that he was depositing withheld portions of their wages in bank accounts for them. When they ultimately discovered their brother’s deceit, the employee-brothers were allowed to sue for damages stemming from the entire course of conduct, on the ground that their brother’s frequently repeated assurances that currently withheld funds would be deposited in the bank accounts served to fraudulently conceal the previous defaults. The continued course of fraudulent conduct fraudulently concealed the previous fraudulent acts.
Second, even though one might view Stetson as a special “fiduciary obligation” case (a point mentioned but left open by the Stetson court, see 321 Mass. at 200, 72 N.E.2d at 413), more recent Massachusetts cases suggest we should not so limit it. In Lynch v. Signal Finance Co., 367 Mass. at 507-08, 327 N.E.2d at 735, the court spoke of the Massachusetts “fraudulent concealment” rule as requiring “positive acts” only where “the aggrieved party has full means of detecting the fraud.” In Friedman v. Jablonski, 371 Mass. 482, 485 & n. 3, 358 N.E.2d 994, 997 & n. 3 (1976), the court, rather than apply the fraudulent concealment rule to a cause of action that was “inherently unknowable,” held that such a cause is subject to a “discovery rule” under which it does not accrue until the plaintiff learns or should have learned of the injury. See also Woodcock v. American Investment Co., 376 Mass. 169, 174, 380 N.E.2d 624, 627 (1978) (no tolling of statute of limitations where there were “no allegations of positive acts of concealment or of the existence of inherently unknowable causes of action”). In light of this more liberal attitude, it seems unlikely that the Massachusetts courts would read Stetson restrictively. We conclude that the district court could have found sufficient later “affirmative” acts of concealment in Community-Suffolk’s continuous course of fraudu*877lent conduct. See also Federal Insurance Co. v. Summers, 330 F.Supp. 1041 (D.Mass.1970) (keeping supposedly lost stock certificate out of circulation and failing to report its “recovery” suffice to establish fraudulent concealment under Massachusetts law).
IV.
We next examine Cataldo’s counterclaim against the Penn Central. Some time before Penn Central instituted this lawsuit, Cataldo and Infelise met with the Santa Fe Railroad to discuss the receivers’ outstanding claims against the Santa Fe. They were close to a “global settlement.” Heinzen, acting on behalf of the Penn Central, learned of this imminent settlement and called a Santa Fe officer. As a result of Heinzen’s phone call, the Santa Fe stopped negotiating. The district court concluded that Heinzen prevented Cataldo from obtaining a settlement of the claims, that Heinzen acted maliciously, and that his conduct unlawfully interfered with Cataldo’s business relations with the Santa Fe. The court awarded Cataldo $17,500 in compensatory damages.
In our view, the district court erred in finding a violation of law, for Heinzen’s phone call falls within the absolute privilege covering attorneys’ statements made in connection with the conduct of litigation. Under Massachusetts law, an attorney’s statements are absolutely privileged “where such statements are made by an attorney engaged in his function as an attorney whether in the institution or conduct of litigation or in conferences and other communications preliminary to litigation.” Sriberg v. Raymond, 370 Mass. 105, 109, 345 N.E.2d 882, 884 (1976). The statements must be “pertinent to the proceedings” to come within the privilege, but this requirement is to be broadly construed. Sullivan v. Birmingham, 11 Mass.App. 359, 416 N.E.2d 528, 531 (1981) (citing Aborn v. Lipson, 357 Mass. 71, 73, 256 N.E.2d 442, 443 (1970)). As the Supreme Judicial Court explained in Sriberg:
The public policy of permitting attorneys complete freedom of expression and candor in communications in their efforts to secure justice for their clients commends itself to us. The basic elements of such a policy were recognized early in this Commonwealth by Chief Justice Shaw in the following terms: “[I]t is, on the whole, for the public interest, and best calculated to subserve the purposes of justice, to allow counsel full freedom of speech, in conducting the causes, and advocating and sustaining the rights, of their constituents; and this freedom of discussion ought not to be impaired by numerous and refined distinctions.” Hoar v. Wood, 3 Metc. 193, 197-198 (1841) ....
370 Mass. at 108-09, 345 N.E.2d at 884. Thus, the Massachusetts courts have applied the privilege, not only in defamation cases, but as a general bar to civil liability based on the attorney’s statements. Sullivan v. Birmingham, 416 N.E.2d at 533. The Massachusetts district court has previously recognized the applicability of the privilege to suits for interference with advantageous relations. Sriberg v. Raymond, 414 F.Supp. 396, 398 (D.Mass.), aff'd, 544 F.2d 15 (1st Cir.1976). Thus the issue is not Heinzen’s motive but whether Heinzen’s statements were “pertinent” to litigation involving his client.
The record shows that they were pertinent. The discussions between Cataldo and the Santa Fe in part involved shipments carried over both Santa Fe and Penn Central lines. Similarly, this lawsuit involves shipments in which both carriers participated. The Carmack Amendment, 49 U.S.C. § 20(11), recodified at 49 U.S.C. § 11707, makes clear that railroads that share the task of transporting a shipment also share the liability for damages. See, e.g., New York, P. & N. R.R. v. Peninsula Product Exchange, 240 U.S. 34, 39, 36 S.Ct. 230, 232, 60 L.Ed. 511 (1916). The railroads have agreed with each other to apportion damages; and they each have authorized whoever is nominally sued to act as the others’ spokesman. We do not see how Cataldo could deny that Penn Cen*878tral had a legal interest in some of his “global settlement” claims against the Santa Fe. Moreover, at the time of the phone call Heinzen was negotiating with a receiver, Strock, about claims arising out of shipments carried in part by both Penn Central and Santa Fe. It is conceded that Heinzen’s efforts to stop the Cataldo/Santa Fe settlements led to a Penn Central/ Strode settlement that was considerably more favorable to the railroads. The parties have pointed to nothing in the record in respect either to motive or to content that could remove Heinzen’s call from the scope of the absolute Massachusetts privilege for communications pertinent to litigation activities.
V.
We turn, finally, to Penn Central’s claim for its reasonable attorneys’ fees and costs incurred in establishing Cataldo’s illegal practices. Chapter 93A provides,
If the court finds in any action commenced hereunder, that there has been a violation of section two [chapter 93A’s substantive provision], the petitioner shall, in addition to other relief provided for by this section and irrespective of the amount in controversy, be awarded reasonable attorneys’ fees and costs incurred in said action.
Mass.Gen.Laws ch. 93A, § 11. The district court concluded, however, that, since Penn Central prevailed on only “a relatively small part” of its original chapter 93A claims and since the railroad acknowledged its inability to segregate the portions of its fees and costs — totalling nearly $1 million — attributable to the issues on which it prevailed, any attempt to calculate an award of fees “would be wholly arbitrary.” Hence, the court declined to order any recovery of fees or costs.
The district court’s resolution of this question might well have been appropriate had the claim for fees and costs been grounded in federal law. See Hensley v. Eckerhart, 461 U.S. 424, 103 S.Ct. 1933, 76 L.Ed.2d 40 (1983). But, where an award of fees or costs rests on state law, state law also controls the method of calculating the size of the award, McGinty v. Beranger Volkswagen, Inc., 633 F.2d 226, 232 (1st Cir.1980) (citing Alyeska Pipeline Service Co. v. Wilderness Society, 421 U.S. 240, 259 n. 31, 95 S.Ct. 1612, 1622 n. 31, 44 L.Ed.2d 141 (1975)); F.F. Instrument Corp. v. Union de Tronquistas, 558 F.2d 607, 610 n. 3 (1st Cir.1977); cf. Brown v. Joiner International, Inc., 523 F.Supp. 333, 338 (S.D.Ga.1981) (following state rule under which attorneys’ fees are a jury question), and Massachusetts cases interpreting chapter 93A’s fee provision appear to dictate a contrary result.
In cases where substantive violations of chapter 93A are established, an award of fees is mandatory. Raymer v. Bay State National Bank, 384 Mass. 310, 424 N.E.2d 515, 521 (1981); Linthicum v. Archambault, 379 Mass. 381, 388-89, 398 N.E.2d 482, 488 (1979). And some fee is evidently to be awarded even when the plaintiff fails to provide any evidence to establish a reasonable value of the award. Heller v. Silverbranch Construction Corp., 376 Mass. 621, 628-31, 382 N.E.2d 1065, 1071-72 (1978) (affirming award of fees in absence of evidence concerning time expended by attorney or rate or reasonableness of charges); McLoughlin v. Disarro, 7 Mass.App. 853, 385 N.E.2d 524 (1979) (rescript) (remanding for award of fees when trial court had denied award for lack of any supporting evidence). The responsibility for establishing a reasonable fee under Massachusetts law lies primarily with the court, not with the party requesting fees:
A judge presiding over the action for which the plaintiff seeks reasonable attorney’s fees has ample opportunity to acquire firsthand knowledge of [the relevant] factors. He thus can discern from his own experience as a judge and expertise as a lawyer, the amount that the attorney should be paid.... In directing the judge to award reasonable attorney’s fees ..., the Legislature listed no specific facts requiring independent proof. From this, we conclude that the judge is *879to rely on his firsthand knowledge of the services performed before him.
Heller v. Silverbranch Construction Corp., 376 Mass. at 629-31, 382 N.E.2d at 1071-72 (footnote omitted); see Hanner v. Classic Auto Body, Inc., 10 Mass.App. 121, 123-24, 406 N.E.2d 686, 688 (1980) (court roughly allocates fees among issues on which plaintiff was, and was not, successful).
Of course, it may be that the district court here believed that Penn Central’s attorneys were entitled to no more than nominal fees and costs in respect to its chapter 93A claims. We found in the record very little indication of attorney efforts devoted to the limited issues on which Penn Central won. The court would seem to have adequate power to make but a nominal fee award in such circumstances. See, e.g., Raymer v. Bay State National Bank, supra (courts have broad powers to determine the magnitude of fees and costs' in ch. 93A actions); Linthicum v. Archambault, supra (same). Nonetheless, because of the risk here that the court used a “federal” standard and declined to make any award simply because of perceived inadequate documentation, we remand on this point.
In sum, our review of the record in this case convinces us that the conclusions of the district court are amply supported, with three comparatively minor exceptions. First, the receiver-defendants must share Cataldo’s liability for bringing unfounded claims against the railroad. Second, Cataldo cannot prevail on his counterclaim. And third, Penn Central is entitled to the reasonable fees and costs related to its successful chapter 93A claims. Since the first and third of these issues require further findings, we remand the case on these two points.
Affirmed in part, reversed in part, and remanded for further proceedings as specified above.