POSNER, Circuit Judge.

The plaintiffs in this Sherman Act price-fixing case appeal from the grant of summary judgment to the defendants. The plaintiffs had opted out from a large antitrust litigation, other phases of which are discussed at 123 F.3d 599 (7th Cir. 1997), and 186 F.3d 781 (7th Cir. 1999); despite some overlap with the issues in the other phases, it will simplify analysis to treat this offshoot of the litigation as if it were completely free-standing, like an adversary proceeding in a bankruptcy case. The question for decision is simple, at least to state: did the plaintiffs present enough evidence to create a triable issue? Equivalently, could a reasonable jury, if all it had before it was the evidence presented in the pretrial proceedings, conclude that the defendants had engaged in price fixing?

The plaintiffs are retail sellers of prescription drugs, and the defendants that are the appellees are wholesale sellers of such drugs, that is, the plaintiffs' suppliers. The plaintiffs argue that the wholesalers conspired with the manufacturers of the drugs to deny the plaintiffs discounts they would have received had it not been for the conspiracy. The manufacturers are also defendants but they are not before us. They remain, their liability as yet unresolved, in the district court. But as authorized by Rule 54(b) of the Federal Rules of Civil Procedure, the district judge entered final judgment for the wholesaler defendants in order to enable an immediate appeal.

The plaintiffs' theory is that the manufacturers agreed not to give discounts to pharmacies and other retail sellers and enlisted the wholesalers to police the agreement by means of a "chargeback" system that the wholesalers had adopted early in the 1980s. Then as later, manufacturers of brand-name prescription drugs engaged in price discrimination. That is, a manufacturer would sell the same product, costing the same to make and sell, at different prices to different customers. The lowest price, presumably, covered the manufacturer's cost (for there is no allegation that the manufacturers were engaged in predatory pricing or forced by adverse business conditions to sell at distress prices), implying that the higher prices in the discriminatory price schedule generated revenues in excess of cost. That sounds like monopoly pricing, but we must be careful here to distinguish between fixed and variable costs. Many of the costs of a new drug are incurred before manufacturing for sale begins — costs of research, of development, of obtaining patents, of obtaining FDA approval, and so forth. A price equal to just the cost of manufacturing and selling the drug, the cost that varies with the amount of the drug sold (the marginal cost, in other words, as distinct from the average total cost of the drug), would therefore not cover the product's total costs. The firm could try to cover those costs by charging a uniform markup over marginal cost, but if customers vary in their willingness to pay for a particular drug, the firm may do better to charge different prices to different customers or groups of customers. A customer's willingness to pay will be a function of the customer's options. In the drug industry, as it happens, hospitals and HMOs, because they "control" to a considerable extent the physicians whom they employ or contract with, are in a good position to effect the substitution of generic equivalents for brand-name prescription drugs. They therefore are unwilling to pay as much for the brand-name drugs as the typical drugstore, which simply fills the physician's prescription — and physicians are often indifferent to the price of the drug they are prescribing.

If brand-name drugs were interchangeable not only with generics but with each other, then unless the manufacturers colluded they would be unable to discriminate in price between hospitals and HMOs on the one hand and drugstores on the other hand. The high markup in the price to the disfavored customers would be competed away. Suppose marginal cost is $5 and price $10; at any price above $5, the seller obtains some contribution to his fixed costs, and so a seller who starts out with a price of $10 will be tempted to shade it to attract sales from his rivals and this competitive process will continue until price is bid down to $5. If, however, the drugs are not interchangeable, whether because of chemical differences protected by patents against being duplicated or because of perceived differences having to do with a manufacturer's reputation or his advertising or other promotional activity on behalf of particular brands, then each manufacturer might be able to engage in price discrimination. Price discrimination is in fact quite common in competitive industries; think only of the difference in price between hardback and paperback books, a difference that almost always exceeds the difference in the marginal cost of the two types; or the difference in ticket prices between the same first-run and subsequent-run movie; or discounts for senior citizens. As long as competitive products are not perfect substitutes to all consumers, the fact of their being competitive does not preclude discriminatory pricing. The publishing industry is extremely competitive but, as just noted, price discrimination is the norm in it. Just as copyrights give the publisher a temporary monopoly of each book he publishes, so patents give manufacturers of drugs a temporary monopoly of each drug he manufactures. These monopolies create preconditions for discriminatory pricing.

Yet even in the case of the differentiated product protected from immediate competitive duplication by a patent, copyright, or trade secret, price discrimination would be feasible only if the manufacturer could prevent (or at least limit) arbitrage — the erasure of a price difference not attributable to a cost difference (that is, a discriminatory price difference) by a middleman's buying from the favored customers and reselling to the disfavored. (Or the favored customer might overbuy and resell the surplus directly to the disfavored one.) And that brings us to the chargeback system. Suppose that a manufacturer wanted hospitals to be able to buy its drugs for 10 percent less than pharmacies, and so it granted its wholesalers a 10 percent discount on all sales intended for resale to hospitals. The wholesaler would have an incentive to overstate the number of those sales and divert the excess above those necessary to meet the hospitals' demand to the pharmacies. Suppose the retail price suggested by the manufacturer for some drug was $10 to pharmacies and $9 to hospitals, and the price charged by the manufacturer to wholesalers for sales destined for pharmacies was $5 and the price for sales destined for hospitals was $4.50. Any quantities of the drug that the wholesaler obtained for $4.50 and resold to a pharmacy would yield the wholesaler $5.50 per sale ($10-$4.50) rather than $5, and so the wholesalers might end up buying their entire supply for $4.50 per unit on the representation that they were selling only to hospitals. The result would be a reduction in the manufacturer's revenue from sales destined for pharmacies from $5 to $4.50, its anticipated revenue from sales destined for hospitals.

Enter the chargeback system, whereby the manufacturer contracts directly with the retail level of distribution, the hospitals and the pharmacies in our example. Continuing with the example, the manufacturer promises a hospital a 10 percent discount, the hospital so informs its wholesaler, the wholesaler reduces its price to the hospital accordingly, and then the wholesaler, which had bought the drug for $5, bills the manufacturer 50, thus preserving its own margin. To get the 50 it must present proof that it sold the drug to a customer authorized to buy at a discount. There is no way the wholesaler can pay only $4.50 for a drug that he resells to a pharmacy for $10. Arbitrage by wholesalers is thus prevented.

Price discrimination by a firm that is not a monopolist and is not colluding with its competitors is generally not an antitrust violation at all; as we said, it is common in competitive industries. It is particularly difficult to object to it when, as in the case of drugs or books, a producer has heavy fixed costs, so that a price equal to marginal cost would not be compensatory. The alternative in such a case to price discrimination is as we said a high uniform markup over marginal cost, and there is no reason to think it a superior alternative to a differential pricing scheme. Indeed often, as quite probably in the book case and possibly in the drug case as well, the differential scheme enables the industry to achieve a larger output than with a uniform price. If the average hardback book is priced at $25 and the average paperback at $9, and if publishers were constrained to price them on the basis of the cost difference and as a result set a price of (say) $18 for hardbacks and $16 for paperbacks, they would probably lose more sales at the low end than they gained at the high end, and if so their total output would be less.

Since price discrimination is not (in general) unlawful, neither are efforts to prevent arbitrage. An agreement by distributors to adopt a system for preventing arbitrage, the better to serve their suppliers, would surely not be a per se violation of the antitrust laws, as the principal effect of invalidating it might simply be to induce the manufacturers to take over the wholesale function themselves. In any event, the charge here is not that a "horizontal" agreement to ban arbitrage is unlawful (there is a hint of such a charge in the plaintiffs' brief, but it is insufficiently developed to preserve the issue for our review), but that the defendant wholesalers joined a conspiracy by manufacturers of brand-name prescription drugs to fix prices, the wholesalers' role in the conspiracy being to prevent arbitrage that would undermine the manufacturers' price-fixing scheme.

To make the charge stick, the plaintiffs would have had to prove two things. The first was that the manufacturers conspired to fix prices. The second was that the wholesalers joined the conspiracy by adopting the chargeback system. We said that manufacturers of differentiated products can engage in price discrimination unilaterally, because each manufacturer has a little monopoly power. But in addition manufacturers who have no individual monopoly power may find it feasible and attractive to engage in price discrimination collusively, that is, to agree on a discriminatory schedule of prices. Suppose the defendant drug manufacturers cannot charge a very high price to hospitals because if they do the hospitals will simply substitute generics, but that drugstores have no such option. Then it would make sense for the manufacturers, even if their brand-name drugs were fungible, to agree (if they thought they could get away with it and that the agreement would be effective) to charge a higher price for drugs destined for drugstores and a lower price for drugs destined for hospitals. But if they reach their customers through wholesalers, they need a way of preventing the wholesalers from engaging in arbitrage; and the chargeback system is that way. That is, the chargeback system is equally efficacious whether used to prevent arbitrage against individual price discrimination or arbitrage against collusive price discrimination. In the second case the chargeback system would have the additional benefit to the conspiring manufacturers of preventing any of them from cheating on their coconspirators (the bane of conspiracy) by trying to lure the high-margin customers (the drugstores in our example) with prices slightly below the agreed-upon price to disfavored customers.

It is not easy to distinguish factually between the two forms of discrimination — one legal, the other illegal — in the setting of this case, since, as we said, the defendant manufacturers do sell differentiated products. The plaintiffs (who remember are the disfavored purchasers, the drugstores and other retail sellers) did present considerable evidence that some of them, at least, have competitive options just as good as those of some of the favored purchasers. For example, in a number of states pharmacists have the legal right, unless the prescribing physician expressly forbids, to substitute a chemically identical drug for the one prescribed. The refusal of the defendant manufacturers to grant discounts to pharmacies in such states must mean, the plaintiffs argue, that the manufacturers have agreed to hold the line on discounts. This is not an impressive argument, because if pharmacies have the same competitive options as hospitals, why would the manufacturers' cartel charge different prices to pharmacies and to hospitals? Stated differently, if the manufacturers can by agreement avoid giving discounts to pharmacies, why do they give discounts to hospitals that are identically situated so far as competitive options, such as generics, are concerned?

We need not pursue this issue, because the district court has not yet determined whether the manufacturers conspired among themselves. Assuming they did, the plaintiffs would still have to prove that the wholesalers joined the conspiracy. They would have to show not only that the wholesalers knew that the manufacturers' price discrimination which the chargeback system assists was collusive rather than individual, but also that the wholesalers agreed with the manufacturers to support that price discrimination by means of the chargeback system.

There is authority for prohibiting as a violation of the Sherman Act or of section 5 of the Federal Trade Commission Act an agreement that facilitates collusive activity, see 6 Phillip E. Areeda, Antitrust Law ¶¶ 1407a-b, 1435d-g (1986 2000 supp.); Herbert Hovenkamp, Federal Antitrust Policy: The Law of Competition and Its Practice § 4.6b-d, pp. 178-86 (2d ed. 1999) — for example, a basing-point pricing system, FTC v. Cement Institute, 333 U.S. 683, 696-700, 711-12, 68 S.Ct. 793, 92 L.Ed. 1010 (1948); Clamp-All Corp. v. Cast Iron Soil Pipe Institute, 851 F.2d 478, 484-85 (1st Cir. 1988); Boise Cascade Corp. v. FTC, 637 F.2d 573, 576-77 (9th Cir. 1980), or a system of exchanging price information, United States v. Container Corp. of America, 393 U.S. 333, 337, 89 S.Ct. 510, 21 L.Ed.2d 526 (1969); Todd v. Exxon Corp., 275 F.3d 191, 198-99 (2d Cir. 2001); In re Baby Food Antitrust Litigation, 166 F.3d 112, 118, 137 (3d Cir. 1999); Wallace v. Bank of Bartlett, 55 F.3d 1166, 1169 and n. 5 (6th Cir. 1995), or industry-wide adoption in contracts with customers of most-favored-nation clauses (which make discounting more costly by requiring that a discount to one buyer be granted to all buyers protected by such clauses). George A. Hay, "Faciliting Practices: The Ethyl Case," in The Antitrust Revolution: Economics, Competition, and Policy 182 (John E. Kwoka Jr. Lawrence J. White eds., 3d ed. 1999); 6 Areeda, supra, ¶ 1435e; but see E.I. DuPont de Nemours Co. v. FTC, 729 F.2d 128, 133-34 (2d Cir. 1984). But that is not the theory of this case. The theory is that the wholesalers joined the manufacturers' conspiracy. And of that there is too little evidence to permit a reasonable jury to infer the wholesalers' guilt. There is first of all no evidence that the wholesalers knew that the manufacturers' price discrimination was collusive rather than individual. To argue that because the uniform refusal to grant discounts even to pharmacies that seemed to have competitive options as good as hospitals and HMOs smacks of collusion the wholesalers knew that the manufacturers were colluding is too much of a stretch; it would amount to basing an inference of conspiracy on negligence, a careless failure to tumble to the nature of one's suppliers' business methods. It would mean that any time a seller asked a distributor not to engage in arbitrage the distributor would be in peril of being found to have conspired with the seller to fix prices should it turn out that the price discrimination engaged in by the seller was collusive rather than individual.

An inference of knowledge would be particularly shaky here because, so far as appears, the chargeback system was adopted before the alleged collusion of the manufacturers began and because the system is supported by commercial reasons independent of any desire to prevent arbitrage, let alone to facilitate collusive pricing. Without the chargeback system or some simulacrum, a hospital or other purchaser of brand-name prescription drugs that was entitled to a discount would have to pay the full price to the wholesaler and then ask the manufacturer for a rebate. The delay entailed in that system would give these purchasers something the wholesalers dreaded, namely an incentive to buy directly from the manufacturers. The chargeback system enabled the purchasers to obtain their discount at the moment of purchase, just as they would have done had they bought directly from the manufacturers. It was a wholesaler survival tactic.

At argument, pressed for examples of evidence of the wholesalers' culpability, their lawyer referred us to a number of places in the voluminous appendix. We have looked at them and we cannot find anything to create more than the barest suspicion that the wholesalers knew the manufacturers were colluding (at present only a hypothesis, we remind) and knowing this decided to help them by means of the chargeback system. Most of it is evidence that the manufacturers were engaged in collusive pricing. This evidence was deemed sufficient by the district court to defeat the manufacturers' motion for summary judgment, but it was hardly strong enough to compel or even permit an inference that the wholesalers knew the manufacturers were colluding. (In a trial of the same claim by the plaintiffs who did not opt out of the class action, the manufacturers were exonerated of the only price-fixing charge in which the wholesalers may have been implicated. 186 F.3d at 784-88.) There is evidence that the chargeback system was intended to discourage arbitrage; undoubtedly it was, but that is consistent with the wholesalers' believing, whether correctly or not, that the manufacturers' price discrimination was individual rather than collusive.

The plaintiffs' best evidence is testimony concerning meetings of the wholesalers' trade association at which manufacturers were present (not that there is anything suspicious in general about suppliers talking to distributors). All the testimony concerns retail buying groups. At one meeting the wholesalers asked the manufacturers whether the latter would be selling directly to groups of buyers, bypassing the wholesalers, and were relieved to be assured that they would not be. This could be a form of compensation for the wholesalers' preventing arbitrage that would thwart collusive price discrimination by the manufacturers, but that is sheer conjecture. A similar statement — "that neither wholesalers nor manufacturers should deal with retail buying groups" — was made at another meeting and embellished with the more intriguing suggestion that the retail buying groups should be denied "access through the chargeback system to the same discounts on BNPDs [brand-name prescription drugs] given to favored purchasers." But actually that is just another way of saying don't deal with retail buying groups, which if they don't receive a discount are not going to take over the wholesale function from the wholesalers — indeed, they are not going to have any reason for being. At worst, the statements we have quoted might be evidence of a boycott by the manufacturers of retail buying groups, cheered on by the wholesalers; but the plaintiffs have not argued boycott.

The plaintiffs' evidence is consistent with the existence of an overarching conspiracy of which the wholesalers were members. But it is equally consistent with there being either no conspiracy at all or one of which the wholesalers were unaware, let alone of which they were members. And we haven't even reached the question whether, if the wholesalers did know there was a manufacturers' conspiracy, they joined it. To infer membership from knowledge would erase the distinction between conspiring on the one hand, which means joining an agreement, and aiding and abetting on the other, which means materially assisting a known-to-be-illegal activity in the hope that it will flourish to the benefit, pecuniary or otherwise, of the aider. E.g., United States v. Pino-Perez, 870 F.2d 1230, 1235 (7th Cir. 1989) (en banc); United States v. Peoni, 100 F.2d 401, 402 (2d Cir. 1938) (L.Hand, J.). We can assume without having to decide that if the wholesalers knew there was a manufacturers' conspiracy and wanted it to succeed, presumably in order to kill the buying groups (mere knowledge that some of the customers for one's goods or services plan to use them for an illegal purpose is not enough to make one an aider and abettor, United States v. Pino-Perez, supra, 870 F.2d at 1235), then, because preventing arbitrage was important to the conspiracy and the chargeback system was an effective way of thwarting it, the wholesalers were guilty of aiding and abetting (which can be a civil as well as a criminal violation of antitrust law, e.g., Allen Bradley Co. v. Local Union No. 3, 325 U.S. 797, 801, 808-10, 65 S.Ct. 1533, 89 L.Ed. 1939 (1945); see generally Electronic Laboratory Supply Co. v. Cullen, 977 F.2d 798, 805 (3d Cir. 1992)) even though as we have said a chargeback system has innocent commercial virtues as well. But that is not argued, and anyway essential premises are missing.

On this record, no reasonable jury could find for the plaintiffs, and so the judgment of the district court must be

Affirmed.