This unwieldy commodities-fraud case, rich in parties and issues, comes to us after a five-week trial that produced a record of several thousand pages. However, the essential facts are simple (and we shall make them even simpler, to the extent this can be done without distorting our analysis), and many of the issues require little or even no discussion.

At the heart of the fraud was Robert Serhant, who between 1980 and 1982 offered investors a "Hedge-Spread Program" that he said would work as follows. For every $100,000 invested, Serhant would use $97,000 to buy a 90-day U.S. Treasury bill having a value at maturity of approximately $100,000 (this was a period of high interest rates) and would invest the remaining $3,000 in Treasury bill futures traded on the Chicago Mercantile Exchange, of which he was a member. He told investors that their risk would essentially be limited to the interest on the Treasury bill, because, at worst, at the end of 90 days they would have $100,000 (more or less)--the $97,000 principal of the Treasury bill plus interest thereon for 90 days. They would have lost only the interest they would have earned if the full $100,000 rather than $97,000, had been invested in Treasury bills. The "Program" was a gimmick, of course; the investors would be no better off than if they gave Serhant just $3,000 each. And it was misleading, as we shall see, to suggest that an investor could lose no more than the amount of the investment used to buy futures.

To do the actual trading of the futures Serhant needed the services of both a clearing member of the Exchange, whose function is to guarantee that each party to a trade will make good on his commitment, and a futures commission merchant, who acts as custodian of the investors' funds. (On the mechanics of commodities trading, and especially the role of the clearing member, see United States v. Dial, 757 F.2d 163, 164-66 (7th Cir.1985); Bernstein v. Lind-Waldock & Co., 738 F.2d 179, 181 (7th Cir.1984); Leist v. Simplot, 638 F.2d 283, 286-88 (2d Cir.1980), aff'd under the name of Merrill Lynch, Pierce, Fenner & Smith v. Curran, 456 U.S. 353, 102 S. Ct. 1825, 72 L. Ed. 2d 182 (1982); Fishman, Commodities Futures: An Introduction for Lawyers, 65 Chi.Bar Record 306, 309 (1984); Chicago Board of Trade, Commodity Trading Manual (1982).) For most of the period of the fraud, Serhant used K & S Commodities, Inc. (owned by Messrs. Schiller and Krumhorn) as both his clearing agent and futures commission merchant.

Serhant's scheme was fraudulent in three respects. First, he invested not 3 percent but almost 100 percent of the investors' money in Treasury bill futures, thus making the investments far riskier than the investors had originally supposed they would be. Erroneously predicting trends in short-term interest rates, Serhant lost $21 million of the $51 million invested in the Hedge-Spread Program. He didn't pocket the money; he just lost it in trading.

Second, he tried to conceal the losses from the investors by various dishonest tricks such as "allocating" profits to investors who had suffered losses. Suppose Serhant made a "block" sale (that is, a consolidated sale of futures contracts held in different accounts, see United States v. Dial, supra, 757 F.2d at 165, in this case accounts of different customers) of two futures contracts, at a price of $2 a contract. And suppose that he had paid $3 for one of the contracts and $1 for the other and that the first contract had been bought for the account of an investor who had later suffered large losses and the second contract for the account of a new investor, who had suffered no losses; Serhant might switch the accounts so that the account of the first investor would show a profit (of $1) that offset some of that investor's losses. As in a Ponzi scheme, Serhant was using newly invested money to make old investors think they were earning profits rather than losing their shirts.

Third, Serhant exaggerated the degree to which the Hedge-Spread Program, had it been implemented as represented, was secure. In commodities trading you can lose more money than you invest, even if you ignore margin calls. A futures contract is a contract to buy or sell a commodity or a financial instrument at a specified price on a specified date, and should an investor be unsuccessful in unloading the contract as the price shifts against him (he might be unsuccessful because the price changed precipitately or simply because his broker failed to offset the contract), he may end up having to ante up more than he invested.

For these various frauds Serhant is now serving an 11-year prison sentence.

The investors whom he fleeced brought this civil suit (actually suits, but we'll suppress that irrelevant detail to simplify the opinion) against a variety of individuals and institutions. Many of the defendants settled before trial, to the tune of more than $8 million. Some did not settle--Serhant, and companies owned by him which he used as vehicles for the fraudulent scheme; K & S, and one of its co-owners, Krumhorn; the Chicago Mercantile Exchange; and the First Bank of Schaumburg. The district court granted summary judgment for the exchange and for the bank but allowed the case to go to the jury against the other defendants, which is to say against the Serhant and K & S groupings. The jury, asked to assess compensatory damages separately against each defendant, returned a verdict that awarded total damages--after some trebling under RICO, see 18 U.S.C. Sec. 1964(c), which Serhant and his companies were held to have violated--of about $3.3 million. Most of the assessment is against the Serhant group of defendants. Only $120,001 was assessed against K & S and $60,001 against Krumhorn. Schiller, the co-owner of K & S, had settled before trial for $350,000.

The plaintiffs appeal from the judgment in favor of the exchange and the bank, and from the district court's refusal to set aside the damage judgment against the other defendants as being too low. Those defendants cross-appeal, contending that they should have had judgment in their favor. We discuss the liability of the Mercantile Exchange first, then the adequacy of the damages judgment, then the liability of K & S and Krumhorn and the bearing of the settlements generally, and last the liability of the First Bank of Schaumburg. There are a few other issues worthy of some discussion and we'll tuck them in at convenient places.

The plaintiffs seek to rope in the Exchange under either of two sections of the Commodity Exchange Act. Section 5a(8), 7 U.S.C. Sec. 7a(8), as it read in the period relevant to this case, required each exchange to "enforce all by-laws, rules, regulations, and resolutions, made or issued by it ..., which relate to terms and conditions in contracts of sale ..., and which have been approved by the [Commodity Futures Trading] Commission...." Section 13(a), 7 U.S.C. Sec. 13c(a), provided that "any person who ... willfully aids, abets, counsels, ... [etc.] a violation of any of the provisions of this [Act] ... may be held responsible in administrative proceedings under this [Act] for such violation as a principal." The Act nowhere expressly authorized private damages suits against violators of either section. The Futures Trading Act of 1982 changed this, see 7 U.S.C. Sec. 25(b)(1)(A); 7 U.S.C. Sec. 13c(a), but the parties agree that the changes are inapplicable to this case. Proceeding under the unamended statute, the district court held that a private right of action is implicit in section 13c(a) (aiding and abetting) but not in section 5a(8) (failure to enforce rules), and so dismissed the section 5a(8) claim. Then he granted summary judgment for the defendants on the aiding and abetting claim, on the ground that there was insufficient evidence to create a triable issue.

The whole question of "implied" rights of action is deeply vexed. It lies at the crux of a series of debates over statutory interpretation. Those judges who believe that most statutes are compromises between rival interest groups hesitate to create implied rights of action no matter how defective a statute's remedial scheme is without them, for they believe that in all likelihood the absence of effective remedies was a part of the compromise that enabled the statute to be passed, and they rightly do not want to undo the compromise. Those who believe that a regulatory statute should be viewed not as the point of balance between conflicting interest groups but as a straightforward effort to eliminate abuses do not hesitate to enforce a statute by whatever remedies are expedient.

Some take a different view; they believe that the question is not whether the statute's ostensible purposes would be served by adding a private right of action to the remedies expressly provided, but whether Congress consciously intended (without bothering to say) that there should be a private right of action. The Supreme Court followed this approach in Merrill Lynch, Pierce, Fenner & Smith v. Curran, 456 U.S. 353, 102 S. Ct. 1825, 72 L. Ed. 2d 182 (1982), and held that there is an implied right of action under the provisions of the Commodity Exchange Act that are designed to prevent fraud. Just as in this case, one of the provisions was section 5a(8). Curran had alleged that a rule of a commodity exchange designed to prevent price manipulations had not been enforced. The Court reasoned that because in 1974, when Congress expanded the Act's remedial provisions (although without creating a private right of action under section 5a(8)), courts were freely implying rights of action and had even done so in a number of commodities cases, and because Congress must be presumed to know the decisional background against which it amends its statutes, it must have wanted the courts to keep on implying private rights of action in the same class of cases; otherwise it would have told them to stop. See 456 U.S. at 377-82, 102 S.Ct. at 1838-41. There is more to the majority opinion than this, but this is the heart.

As an original matter we might question the last two links in this chain. They might be thought to impute to Congress greater knowledge of case law than is realistic to suppose it has, to underestimate the degree to which Congress is content to avoid deciding questions (such as the question whether courts should have continued implying private rights of action in the Commodity Exchange Act notwithstanding the strengthening of the public remedies by the 1974 amendments), and to equate legislative inaction to action.

More pertinent than these reservations is the fact that some courts, including our own, have, while following Curran 's general approach, declined to imply a right of action against violators of provisions of the Commodity Exchange Act relating to disputes over membership in exchanges and disputes over exchanges' failing to alter the terms of their futures contracts in ways that might benefit investors. See, e.g., Cardoza v. CFTC, 768 F.2d 1542, 1554-57 (7th Cir.1985); Sam Wong & Son, Inc. v. New York Mercantile Exchange, 735 F.2d 653, 666-70 (2d Cir.1984) (Friendly, J.). Central to these decisions is the proposition that because such disputes had not been the subject of private damages actions before 1974, Congress could not be presumed to have wanted to authorize such actions as a means of resolving them.

This may be slicing the cake too thin. The point emphasized in Curran was not that courts had entertained private suits to enforce specific provisions of the Commodity Exchange Act before 1974, though this was mentioned, but that courts in that era were liberal in implying private rights of action in remedial statutes generally. A stronger defense of Cardoza and Sam Wong is that in the circumstances of those cases a private right of action would not have been an expedient adjunct to the statute's remedial system. Membership disputes, for example, are traditionally and on the whole satisfactorily handled by arbitration, while as Judge Friendly pointed out in Sam Wong to allow a trader to bring a suit just because he could imagine a change in the terms of a futures contract that might have averted a loss could flood the courts with litigation unmanageable in its speculativeness. See 735 F.2d at 669.

In the case of a massive multi-million dollar fraud facilitated by an exchange's failure to enforce rules designed to prevent fraud, it can be argued that the remedies provided by the Act are inadequate and that that is why Congress amended the Act in 1982 to make the right to sue an exchange guilty of such failures explicit. Since we are proceeding under the Act as it stood before 1982, Congress's intentions in 1982 are relevant only insofar as they cast light on the understanding of the earlier Congresses that created the Act in the form in which it existed at that time. They cast this much light: Congress did not believe in 1982 (or at least did not say) that it was changing course and revising an earlier compromise that had ruled out effective remedies for violations of the Act.

The present case is too much like Curran to be easily distinguished. The fact that the cases decided before 1974 did not involve any violations of exchange rules designed to prevent fraud as such, but only violations of rules designed to prevent price manipulation, is immaterial, for some of the cases involved violation of the statutory prohibition against fraud, and anyway price manipulation is a species of fraud. Although Curran jostles uneasily with other, and apparently still authoritative, statements of the principles governing implied rights of actions, see, e.g., Cort v. Ash, 422 U.S. 66, 78, 95 S. Ct. 2080, 2088, 45 L. Ed. 2d 26 (1975), the Mercantile Exchange concedes that if the plaintiffs can fit themselves within the principles of Curran they have an implied right of action.

Even if the district judge erred in dismissing the plaintiffs' section 5a(8) claim at the threshold, however, the error would be harmless if as the Exchange argues there was not enough evidence to create a triable issue with regard to its alleged failure to enforce its rules.

The express private right, created in 1982, to sue an exchange for failing to enforce rules that it is required to enforce is limited to situations where the exchange's failure is in "bad faith." 7 U.S.C. Sec. 25(b)(4). In so limiting the right Congress was merely codifying what had become the settled understanding of the scope of the implied right of action recognized in Curran and other cases. Brawer v. Options Clearing Corp., 807 F.2d 297, 302 (2d Cir.1986); Sam Wong & Son, Inc. v. New York Mercantile Exchange, supra, 735 F.2d at 676 n. 30. By "bad faith," however, the cases did not mean actual participation in fraud (or other misconduct), for such conduct would make the exchange liable directly under section 4b, 7 U.S.C. Sec. 6b, the anti-fraud section. Yet, read naturally, the words seem to imply that the exchange's failure to enforce must be more than merely negligent. How much more? In ordinary English "bad faith" implies a deliberate wrong rather than just failing to come up to an objective standard of care, which is what negligence is. Therefore, as an original matter (an important qualification, as will appear) we might conclude that the plaintiffs in this case would have to prove that the Mercantile Exchange knew that a rule it was required to enforce was being violated, or--what is in fact a form of knowledge--deliberately closed its eyes so that it would not discover what it strongly suspected was going on. This is the form of reckless disregard that we discussed recently in the context of prison officials' liability for failing to protect prisoners from dangers to their health or safety. See Duckworth v. Franzen, 780 F.2d 645, 651-56 (7th Cir.1985); cf. United States v. Ramsey, 785 F.2d 184, 189 (7th Cir.1986); United States v. Josefik, 753 F.2d 585, 589 (7th Cir.1985).

The Mercantile Exchange's Rule 536 requires that "at the time of execution [of an order], every order received from a customer ... must be in writing, with the customer's designation indicated." One purpose of the language we have italicized is to prevent what we earlier called "allocation"--an integral part of Serhant's fraud. When a member of the Exchange wants to make a trade, he must transmit his order to the floor. This is usually done by phone to save time. The floor broker writes down the order dictated over the phone, then steps away from the phone booth and, by means of hand signals, transmits the order to the "pit," where it is executed. He then makes a note of the transaction, which is later posted to the customer's account with the futures commission merchant. Often Serhant would be placing a block order on behalf of a number of investors in his Hedge-Spread Program. Each investor had a customer designation (a number), which Serhant could have dictated over the phone to the floor broker; but he refused to do this, on the ground that it was too time-consuming and that recording the customer designations when the memo of the transaction came back from the floor satisfied the rule. An officer of K & S disagreed and complained to the Exchange. At a conference where she and Serhant argued their opposing points of view, the Exchange's representative agreed that it would be too much of a burden on Serhant to require him to dictate the customer designations to the floor broker. Later the Exchange conducted an audit of Serhant's records to make sure that the customer designations appeared on the office copies of the floor brokers' memos of transactions; they did.

So the Exchange knew that Serhant was not dictating customer designations to the floor broker, but the anterior question, on which liability under section 5a(8) depends, is whether in failing to do this Serhant was violating Rule 536. Although read literally the rule requires that the customer designation be on the order at the moment of execution, the Exchange interpreted the rule as requiring merely that the designation appear on it within a reasonable time after execution. Is this a permissible interpretation? There is a danger that if the order is executed before the customer designation is placed on it, the customer's representative (Serhant) will bide his time and not place the designation on the order when it comes back to his office until he knows how each customer is doing and can shuffle designations around as may be necessary to conceal losses. In fact this is what happened. But the issue presented to the Exchange was not whether Serhant was likely to play any such games but whether the customer designation could be affixed to the floor order immediately after rather than immediately before execution. Concerned with the delay that would be created if the designations had to be on the floor order before it was executed, the Exchange decided that Serhant could wait till right after the order was executed. Even if this interpretation of Rule 536 was incorrect or even unreasonable, it would not be in bad faith under the demanding standard suggested. Cf. Trans World Airlines, Inc. v. Thurston, 469 U.S. 111, 126-27, 105 S. Ct. 613, 624-25, 83 L. Ed. 2d 523 (1985).

Now in fact Serhant often waited much longer than this to put on the customer designations, and therefore he was violating the rule even as interpreted by the Exchange. The Exchange argues that there is no evidence either that it knew this or that it went out of its way to avoid knowing because it was fearful of discovering something it would rather not know. It argues, in short, that it acted in good faith. But this argument works only if "bad faith" in the 1982 amendments, and (what is more germane to this case, which arises under the earlier statute) in the case law that the amendments codified, has the same meaning it bears in ordinary English. The law does not always use the words "bad faith" in their ordinary-language sense, see, e.g., Harlow v. Fitzgerald, 457 U.S. 800, 818, 102 S. Ct. 2727, 2738, 73 L. Ed. 2d 396 (1982); 1 Anderson on the Uniform Commercial Code Sec. 1-201:86 (3d ed. 1981), and, in the present setting, involving an exchange's failure to enforce its rule in accordance with the exchange's own interpretation, the courts, including our own, have treated the term "bad faith" as if it read "negligence." See Hochfelder v. Midwest Stock Exchange, 503 F.2d 364, 368 n. 2 (7th Cir.1974), a case involving the duty of self-regulation under the Securities Exchange Act rather than the Commodity Exchange Act but not distinguishable on any principled ground; also Rich v. New York Stock Exchange, 522 F.2d 153, 155 n. 4 (2d Cir.1975).

Thus, if an exchange's regulation imposes a duty that the exchange should know is being flouted, the exchange is acting wrongfully--in an attenuated sense, perhaps, but one sufficient under the cases we have cited to demonstrate bad faith. The test is different if the exchange injures a trader through the exercise of a discretionary power, such as the power to take emergency actions to prevent substantial losses; then more must be shown to constitute bad faith--either that the exchange acted unreasonably or that it had an improper motivation. See Sam Wong & Son, Inc. v. New York Mercantile Exchange, supra, 735 F.2d at 670-78; Daniel v. Board of Trade of City of Chicago, 164 F.2d 815, 819 (7th Cir.1947). Discretion implies latitude for judgment, and commodity exchanges must not be deterred from exercising judgment by the prospect of heavy liability if they make a mistake.

Thus in the present case, if the alleged bad faith were just the Mercantile Exchange's interpreting its rule to allow Serhant to place customer designations on orders shortly after rather than before or at the moment of execution, the higher standard for proving bad faith would be applicable. Interpretation is often, and was here, discretionary, cf. Homemakers North Shore, Inc. v. Bowen, 832 F.2d 408, 411-12 (7th Cir.1987); and the Exchange's interpretation, while somewhat free, was not unreasonable, and there is no evidence of improper motivation. But with regard to Serhant's violations of Rule 536 (as interpreted by the Exchange), there was no scope for the exercise of discretion and hence no excuse (under the cases) for the Exchange's carelessness in ignoring evidence that such violations were being committed. Furthermore, despite the audit there is some evidence that the Exchange knew that Serhant was delaying up to six hours in placing customer designations on the order form--a delay too long for the Exchange to deem him in compliance with its regulation, even when the regulation is interpreted to allow reasonable delay.

So the plaintiffs appear to have a strong case under section 5a(8)--though not one on which they can recover any damages from the Mercantile Exchange in this case, as we shall see later on.

The plaintiffs' case for aiding and abetting is weak. To begin with, the language of section 13c(a) appears to exclude a private right of action for aiding and abetting, by stating that "any person who ... aids, abets, ... [etc.] ... may be held responsible in administrative proceedings." (Emphasis added.) The statute seems not merely to create a right and specify a remedy for it but to define the right in terms of a remedy--to create a right to an administrative remedy rather than a free-standing right to which courts may attach such remedies as seem necessary and proper to assure reasonable compliance with the statute. Moreover, the legislative history of the 1982 amendment to section 13c(a), though scanty, does indicate that Congress thought it was changing the law rather than confirming the judicial interpretation of the existing statute. See H.Rep. No. 565, 97th Cong., 2d Sess. 104 (1982) U.S.Code Cong. & Admin.News 1982, pp. 3871, 3953 ("At present, the provision [section 13c(a) ] is applicable only to administrative proceedings under the Act"). This is some evidence regarding the scope of the prior law. Although, surprisingly, the question whether section 13c(a) was enforceable by an implied right of action before the 1982 amendment has never before risen to the appellate level, the district courts that have considered the question have unanimously answered "no"--until this case. See Strax v. Commodity Exchange, Inc., 524 F. Supp. 936, 944 (S.D.N.Y.1981); Johnson v. Chilcott, 590 F. Supp. 204, 206 (D.Colo.1984); Bennett v. E.F. Hutton Co., 597 F. Supp. 1547, 1552-54 (N.D.Ohio 1984); Evanston Bank v. Conticommodity Services, Inc., 623 F. Supp. 1014, 1022-23 (N.D.Ill.1985).

We are puzzled why Judge Leighton thought there was a private right of action under section 13c(a) but not under section 5a(8). But even if there is such a right under section 13c(a), as we greatly doubt, the evidence that was before him when he ruled on the defendants' motions for summary judgment did not create a triable issue; summary judgment should have been granted. Aiding and abetting in the criminal law requires not only knowledge of the principal's objective but a desire to help him attain it. See, e.g., United States v. Ambrose, 740 F.2d 505, 509 (7th Cir.1984); United States v. Peoni, 100 F.2d 401, 402 (2d Cir.1938) (L. Hand, J.). There is no evidence of such knowledge and desire here; indeed it would be most unlikely for the Exchange to want to help a Serhant defraud his customers, for scandals such as the Serhant frauds can only hurt the Exchange. The plaintiffs argue that civil aiding and abetting requires less proof of participation in the principal's scheme than criminal aiding and abetting, but they present no authority to support this argument and they overlook a ruling by the CFTC interpreting section 13c(a) to adopt the criminal standard of aiding and abetting. In re Richardson Securities, Inc., CCH Commodity Futures L.Rep. p 21,145, at pp. 24,642-24,646 (CFTC Jan. 27, 1981). As the Commission explained there, the legislative history of section 13c(a) shows that the aiding and abetting provision was modeled on, and was intended to be interpreted consistently with, the federal statute that makes aiding and abetting a crime, 18 U.S.C. Sec. 2.

We move to the question whether the damage award was so inadequate (perhaps because of errors in instructions or trial rulings) that the plaintiffs are entitled to a new trial. The plaintiffs point out that they indisputably lost $7 million during the period when K & S (a principal defendant in the part of the case that went to the jury) was handling Serhant's trades, and they say the jury was irrational in awarding so much less. They wax especially wroth over the award to Judge Bosco (one of the plaintiffs) of a mere $1 in damages, though the evidence that he lost many thousands of dollars is undisputed. And they complain that the district court, in remarking to the jury that the confirmations that the plaintiffs received from K & S were "in plain English," made the jury think that the plaintiffs were the authors of their own losses.

It is not hard to reconstruct the jury's thinking. In arguing to the jury, the defendants' able lawyers harped on the theme that the plaintiffs are affluent, knowledgeable persons--judges, businessmen, lawyers, even a CPA and a broker on the Chicago Mercantile Exchange--who if they had read the extensive mailings from K & S would have realized, and probably did realize despite their denials, that Serhant was putting most of their investments into Treasury bill futures. When a confirmation states that the recipient has just bought "T-bill Sept." for $88, it is plain that he has bought a futures contract rather than a Treasury bill, which is not sold in such small denominations. Granted, contributory negligence is not a defense to fraud, Greycas, Inc. v. Proud, 826 F.2d 1560, 1562 (7th Cir.1987), the main charge on which the case went to the jury. But the defendants' argument is not that the plaintiffs were guilty of contributory negligence but that they knew Serhant was deviating from the terms of the Hedge-Spread Program and ratified his deviation. If so, the plaintiffs were the authors of their own losses.

These were trading losses. Serhant did not pocket any of the plaintiffs' money; on the contrary, he lost hundreds of thousands of his own dollars. Apparently the jury believed that the plaintiffs, trusting in Serhant's investment genius (as they thought it), consented to his using as much of their money to buy Treasury bill futures as he saw fit. This is not the whole story, because by using the dishonest device of "allocation" to conceal losses Serhant may have caused some of the plaintiffs to stick by him longer than they would have done if they had known the truth, and, in sticking by him longer, to lose more money. Maybe this is why the jury did award some damages, though much less than the plaintiffs had sought.

The jury heard the plaintiffs testify and read the confirmation notices and it was therefore in a good position to determine whether, and approximately how much of, the losses were due to the plaintiffs' willingness to assume the greater risk that Serhant imposed on them when he put more of their money into Treasury bills than he had said he would. The verdict is harsh, and may well be wrong in imputing to the plaintiffs an intelligent interest in their financial affairs, but as it is not irrational it must stand. The alleged trial errors are trivial, consisting as they mainly do of a few comments by the judge in the course of a long trial, and of unimportant details in the instructions. It is most unlikely that the alleged errors affected the verdict. Jurors, especially in a long trial, acquire confidence about their knowledge and ability to decide the case and are not easily swayed by the judge's infrequent comments on particular bits of evidence, or by the minutiae of jury instructions rarely drafted with an eye to maximum intelligibility.

The next issue is the liability of K & S and Krumhorn, challenged by these parties' cross-appeal. The jury exonerated them of the charge of fraud but found them liable for having breached a fiduciary duty to their customers, the investors in the Hedge-Spread Program. The first finding moots the argument that the judge should not have dismissed the charge that K & S and Krumhorn violated Rule 10b-5 of the Securities Exchange Commission. Rule 10b-5 is applicable to this case, even though it is a commodities case, because limited partnerships are securities, Rodeo v. Gillman, 787 F.2d 1175, 1177-78 (7th Cir.1986); but it requires proof of fraud.

We greatly doubt whether K & S--and its co-owner, Krumhorn, who had no personal dealings with any of the plaintiffs--could be liable under either the Commodity Exchange Act or state law (if not preempted, another issue we need not decide) to customers of K & S for failing to communicate its suspicions of Serhant to them, merely in virtue of K & S's status as a clearing member. The clearing member's job is to make sure that trades go through, rather than to prevent misrepresentations by the member whose trades it is guaranteeing. But K & S was also Serhant's futures commission merchant and in this capacity designated one of Serhant's companies as an agent to assist it in trading for the accounts of the investors in the Hedge-Spread Program. Section 2(a)(1) of the Commodity Exchange Act, 7 U.S.C. Sec. 4, discussed in Rosenthal & Co. v. CFTC, 802 F.2d 963 (7th Cir.1986), imputes liability for an agent's violations of the Act to the agent's principal (i.e., to K & S), consistently with usual notions of respondeat superior. Another provision of the Act, section 13(c), 7 U.S.C. Sec. 13c(c), also discussed in Rosenthal, see 802 F.2d at 966-67, imputes liability to "controlling persons" (that is, persons controlling the violator) as well, i.e., to Krumhorn and Schiller.

However, the $350,000 settlement that Schiller made with the plaintiffs wipes out any liability of K & S and Krumhorn and, as we shall see, most of the liability of the other defendants as well.

A tort victim can obtain only one recovery for his harm, no matter how many tortfeasors inflicted it. E.g., Dundee Cement Co. v. Howard Pipe & Concrete Products, Inc., 722 F.2d 1319, 1324 (7th Cir.1983); Prosser and Keeton on the Law of Torts Sec. 48 (5th ed. 1984). This principle is seen most clearly in cases where a group of tortfeasors inflicts an indivisible harm, as for example where one tortfeasor places a bucket under the plaintiff's chair, another fills it with gasoline, and a third drops a match into it, causing it to explode and injure the plaintiff. Because the injury is indivisible each tortfeasor is liable for the full harm. But the "plaintiff's total recovery, from all tortfeasors together, is not allowed to exceed his total damages." Donovan v. Robbins, 752 F.2d 1170, 1179 (7th Cir.1984); see also Zenith Radio Corp. v. Hazeltine Research, Inc., 401 U.S. 321, 348, 91 S. Ct. 795, 811, 28 L. Ed. 2d 77 (1971). So "once the plaintiff has been fully compensated for his injuries by one or more of the tortfeasors, he may not thereafter recover any additional compensation from any of the remaining tortfeasors." Watts v. Laurent, 774 F.2d 168, 179 (7th Cir.1985). We are speaking only of compensatory damages; punitive damages do not measure the plaintiff's loss, so piling them on top of compensatory damages is permissible.

The present case is one of indivisible injury, most clearly so with regard to K & S and its co-owners. K & S failed to warn the plaintiffs of its suspicions; the owners of K & S caused K & S to fail to warn; and as a result--an indivisible result of the wrongs committed by these three tortfeasors--the plaintiffs were hurt. Thus, in splitting damages inflicted by the group of assessing $120,001 against K & S and $60,001 against Krumhorn, the jury erred, for when tortfeasors inflict an indivisible injury each is liable for the full amount (although the plaintiff cannot collect from all together more than that amount). The jury was led to commit this error by the judge's action--to which the defendants did not object, however, and of which they therefore cannot now complain--in asking it to assess compensatory damages separately against separate defendants. In the case of a single injury the jury should be asked to determine the amount of the plaintiff's loss, which then becomes a joint and several liability of the defendants who inflicted it. Watts v. Laurent, supra, 774 F.2d at 175-81; Douglass v. Hustler Magazine, Inc., 769 F.2d 1128, 1146 (7th Cir.1985); McKinnon v. City of Berwyn, 750 F.2d 1383, 1387 (7th Cir.1984). Only punitive damages should be assessed separately.

We could decide that the highest assessment of compensatory damages against any of the jointly liable defendants placed a ceiling on the recovery of compensatory damages, and hence that the plaintiffs could recover only $120,001 from the K & S group of defendants. Authority for proceeding so is discussed in Prosser and Keeton on the Law of Torts, supra, Sec. 47, at p. 329; and Watts v. Laurent, supra, 774 F.2d at 180, looks in the same direction. But even if, rather than do this, we suppose that the jury found that the K & S group inflicted $180,002 in harm on the plaintiffs ($120,001 + $60,001), this cannot save the judgment. The plaintiffs had already collected $350,000 from Schiller on account of the identical harm. They do not argue that any part of the settlement with him represented punitive rather than compensatory damages. The plaintiffs were fortunate enough to have been able to obtain from one of the joint tortfeasors an amount greater than their actual harm as authoritatively fixed by a valid jury verdict, and they are therefore barred from collecting the verdict itself. "[T]he verdict rendered must be considered as the full amount of damages suffered by the plaintiff. Furthermore, it is the court's duty ... to credit the judgment by any settlements." In re Jones, 804 F.2d 1133, 1143 (10th Cir.1986) (citation omitted); see also Royster Co. v. Union Carbide Corp., 737 F.2d 941, 950 (11th Cir.1984); Sumitomo Bank of California v. Product Promotions, Inc., 717 F.2d 215, 219 (9th Cir.1983).

This discussion flags another question, the correct answer to which will require us to set aside the entire damages award except the award of punitive damages against Serhant.

Serhant was the primary wrongdoer and his entire course of wrongful conduct was placed before the jury, which decided that the total injury inflicted by that conduct on the plaintiffs justified a recovery substantially smaller than what they had already collected from the settling defendants. Indeed, after the punitive damages assessed against Serhant are subtracted, the total damage award was only $3.1 million, compared to more than $8 million awarded against the settling defendants, no part of which is plausibly viewed as an award of punitive damages since no defendant except Serhant was guilty either of an actual fraud or of some other form of intentional or otherwise aggravated wrongdoing that might justify punitive damages. Having been more than compensated by the settlements, the plaintiffs were entitled to no more at trial, except the portion of the verdict that represents punishment rather than compensation. Realistically, since the jury assessed compensatory damages separately against each defendant, it is possible that the more defendants there had been the more damages would have been assessed. But a party has no right to an improper method of submitting damage issues to a jury. Nor is it very likely, in view of the jury's skepticism about the plaintiffs' damage claims, that adding the Chicago Mercantile Exchange as a defendant would have led the jury to award aggregate damages greater than the settlements.

The last issue is the liability of the First Bank of Schaumburg. Its then president, Weaver, urged a man named Baranski to borrow money from the bank to invest in Serhant's scheme, as Weaver had done (losing, we were told at argument, $1 million of his own money when the scheme collapsed). Weaver also helped put the loans through. Baranski lost the principal of the loans when the Hedge-Spread Program foundered, but he still had to pay the loans back, and he seeks to recover the principal from the bank on the theory that the bank empowered Weaver to suck Baranski into Serhant's whirlpool.

The parties agree that the bank is liable, if at all, only if Weaver knew that Serhant was engaged in fraud. If so, it can be argued that by holding Weaver out to the world as authorized to make loans the bank facilitated his fraud, however unwittingly, and is liable for the consequences. But there is no evidence that Weaver knew of the fraud. So far as appears he was one more sucker, who, dazzled by Serhant's apparent investment genius, saw an opportunity both to build goodwill among the bank's customers and to increase the bank's loan portfolio by encouraging those customers to invest in the Hedge-Spread Program. Although Weaver received a commission from Serhant, this is no evidence that he knew Serhant was a crook; his own loss of $1 million is some evidence he did not. Baranski conjectures that Weaver promoted the Hedge-Spread Program in an effort to obtain funds that would be used to cover Weaver's own trading losses, but there is no evidence, direct or circumstantial, of any such motive.

Finally, Baranski argues that he was prevented from completing discovery before the judge acted on the bank's motion for summary judgment. In fact Baranski conducted extensive, though wholly fruitless, discovery. All he failed to do was depose Weaver. There was a freeze on discovery for a time but there were many months during which Baranski could have deposed Weaver but made no effort do so--made no effort even to obtain a copy of the deposition that Weaver had given in another case arising out of Serhant's fraud. The district court was entitled to conclude that Baranski was stalling around with no serious intention of obtaining evidence that Weaver knew of Serhant's fraud.

This completes our discussion. There are some issues we have not discussed but they are either irrelevant on the view we take of the case or too insubstantial to have even colorable merit. To summarize, we affirm the judgment except insofar as it awards compensatory damages to the plaintiffs (that is, the punitive component of the RICO judgment against the Serhant group of defendants stands). We award costs on appeal to the defendants.

AFFIRMED IN PART AND REVERSED IN PART.