Richard H. Stern,* Post-Sale Patent Restrictions After Mallinckrodt – An Idea in Search of Definition, 5 Alb. L.J. Sci. & Tech. 1 (1994)

Albany Law Journal of Science & Technology



For over a century it was a fixture of United States patent law that a patent owner could not use the patent laws to restrain how its customers used patented products that they purchased from the patent owner or its authorized distributors and licensees. A customer did not commit patent infringement when it used the products--wherever, however, and for whatever purpose the customer pleased. n1 Nor could a patent owner prevent a customer from reselling the products wherever and however it pleased. n2

A. The License-Sale Distinction

To be sure, a patentee could, in many circumstances, give a manufacturer-licensee a limited license to exploit the same patent only in a designated field of use. n3 Although patent law did not recognize the legitimacy or legal efficacy of restrictions imposed on customers' use of patented products that a patentee sold them, patent law did recognize and enforce what appeared to be functionally equivalent limitations on the scope of licenses to manufacture and sell patented products. n4

This part of United States patent law was so clearly established, that by 1926 the Supreme Court observed:

It is well settled . . . that where a patentee makes the patented article and sells it, he can exercise no future control over what the purchaser may wish to do with the article after his purchase. It has passed beyond the scope of the patentee's rights. But the question is [*4] a different one . . . when we consider what a patentee who grants a license to one to make and vend the patented article may do in limiting the licensee in the exercise of the right to sell. n5

When a patentee granted such a "make and vend" license, the patentee could rightfully limit the right of use and disposition of the licensee-made articles, "provided the conditions of sale are normally and reasonably adapted to secure pecuniary reward for the patentee's monopoly." n6 Stated differently, the patentee was entitled to "grant a license . . . upon any condition the performance of which is reasonably within the reward which the patentee [as a patentee] is entitled to secure." n7

The phrases "reasonably adapted," n8 "reasonably within the reward," n9 and "entitled to secure" n10 are not self-defining. The outer limits of how far a patentee might reasonably go in limiting licensees' conduct have not been clear, but courts have thus far allowed [*5] virtually any limitation on the scope of a license to manufacture a patented product unless the limitation was shown to be a device to form an industry cartel. n11 On the other hand, it seemed clear for most of this century that a patentee could go virtually nowhere in limiting its customers' conduct, except for restrictions shown necessary to protect health, safety, and product reliability. n12

A rationale for the distinction between the two kinds of patent-related limitation – those accompanying sales and those accompanying licenses to manufacture – can be based on the type of property right conceived to be paramount in the given transaction. The courts attempted to accommodate a perceived need to defer to customers' property rights in goods that they bought with a desire to defer to perceived rights of patentees to relinquish only so much of their patent monopoly rights as the patentees wished to bargain away. When a patentee sold its patented goods to a customer, customers' personal property rights governed over intellectual property rights of the patentee after the point of the patentee's sale of the patented goods. n13 At that point, courts considered that the patentee relinquished the patent rights; such rights were said to have become "exhausted" or "extinguished" (hence the name "exhaustion doctrine"). n14 On the other hand, when a patentee licensed someone to manufacture goods under the patent, courts considered that the intellectual property rights of the patentee were never extinguished in respect of unlicensed subject matter. n15 [*6]

B. The Mallinckrodt Decision

In 1992, a century of law under the exhaustion doctrine was abruptly swept away in Mallinckrodt, Inc. v. Medipart, Inc. n16 Mallinckrodt owned patents on a device for trapping radioactive aerosol mist, used by hospitals in diagnostic imaging of patients' lungs. n17 The device was a plastic box, containing a filter to trap exhaled mist, placed within lead shielding which controlled radiation from the exhaled mist to prevent it from contaminating the environment. n18 Mallinckrodt labeled the unit "single use only," n19 sold it to hospitals for approximately $50, and instructed the hospitals to dispose of the units after a single use by sending them to a hazardous waste disposal center n20 (which would presumably destroy them). Medipart offered to recycle units n21 for hospitals for approximately $20. Medipart cleaned the units, replaced the filters, and exposed the refurbished units to gamma radiation to kill bacteria. n22 Mallinckrodt filed a patent infringement suit against Medipart. n23 Both parties filed motions for summary judgment, n24 and the district court ruled for Medipart on the basis of the exhaustion doctrine. n25

Mallinckrodt appealed the case to the Federal Circuit, the single court of appeals for patent cases. n26 The Federal Circuit reversed, holding that it was patent infringement for a customer to disobey the restrictive notice. n27 An injunction under the patent law would therefore compel customer obedience to the restriction, n28 and infringers would be liable to the patentee for damages that the disobedience caused. n29 The court held the prior case law — including [*7] many decisions of the Supreme Court — to the contrary to be mere obiter dictum and without legal force. n30 According to the Federal Circuit’s new Mallinckrodt doctrine, the two older legal doctrines concerning patent-based limitations and restrictions merged. n31 Or rather, the doctrine permitting limited licensees to manufacture goods under a patent wholly swallowed up the exhaustion doctrine. n32 With the possible exception of price-fixing restrictions and some tie-in restrictions, n33 one doctrine effectively governs all patent-related restrictions and limitations, whether the patentee imposes them on customers who buy the goods or on licensees who make them. n34 That single doctrine — the Mallinckrodt doctrine — is that a patentee may restrict use and disposition of patented articles and that violation of the restriction is patent infringement that unless the restriction violates some provision of positive law, such as the antitrust laws. n35

C. Business Impact of Mallinckrodt

This Article does not address whether the Mallinckrodt decision’s legal analysis is consistent with precedent, principled, doctrinally sound, or even legally supportable. The short answer is that it is not, for it is simply fiat, judicial legislation. The author has addressed that elsewhere, n36 and that is not the focus of this Article.

[*8] For purposes of this Article, we may consider the institution of the Mallinckrodt doctrine a fait accompli. (Conceive, if you will, that metaphorically the exhaustion doctrine was Custer and that it is now shortly after the Battle of Little Big Horn. Now what, after the exhaustion doctrine's last stand?)

Instead, this Article is about possible business consequences of the Mallinckrodt doctrine in areas that the exhaustion doctrine formerly governed: What can a patentee now accomplish by invoking the Mallinckrodt doctrine? Are there alternative expedients that permit a patentee to accomplish much the same results? Are alternative expedients superior or inferior? What are the likely public policy impacts? Is the Mallinckrodt doctrine beneficial or pernicious? Or in which circumstances is it one or the other? Would some compromise, intermediate between the Mallinckrodt and exhaustion doctrines, provide a wiser rule?


Taken at face value, the Mallinckrodt decision is a remarkable change in the law which may have remarkable consequences. If followed as a precedent, Mallinckrodt would permit patentees to accomplish many things that previously were infeasible. Patenteesellers may now be able to limit their customers to a significant extent in repairing, refurbishing, recycling, and modifying equipment. The Mallinckrodt decision may make it easier for patentee-manufacturers to maximize their revenue from niche or differentiated markets, in which the same technology may have different values. These are markets in which effective post-sale use restrictions would permit the patentees to charge appropriately different sales prices for the same product, on the basis of the different values. This part of this Article will address such possible business impacts of Mallinckrodt in various industries.

A. Why Aren't Patentees Pushing the Edge of the Envelope?

A preliminary word of caution may be in order. Assessing the possible effects of a new, post-Mallinckrodt dispensation is problematical. Taking Mallinckrodt at face value may require a certain amount of willing suspension of disbelief. The author has asked a number of corporate patent counsel what advantage they are tak [*9] ing, or plan to take, of the ruling. To what extent are they restructuring their business arrangements? Generally, their reaction was not aggressive. Indeed, it was timid. Many seemed concerned that it would be risky to restructure their company's business arrangements extensively on the basis of Mallinckrodt.

Their fear was that Mallinckrodt might not endure as a new rule of law because its support as precedent seems fragile or shaky. Courts may distinguish it and confine it narrowly to its facts. The Federal Circuit may ignore its ruling in subsequent cases. n37 Hence, counsel who recommended aggressive business strategies based on Mallinckrodt might later be embarrassed by a judicial retreat from it and a return to the exhaustion doctrine.

To that extent, the discussion that follows may be premature. The report of the demise of the exhaustion doctrine may be exaggerated. With that warning, this Article hereafter operates on the theory that the Mallinckrodt decision has supplanted the exhaustion doctrine in federal patent law and that, barring Supreme Court review of a subsequent Mallinckrodt-based decision, n38 the [*10] traditional exhaustion doctrine is unlikely to be revived.

That patentees thus far are not pressing the Mallinckrodt decision to its limits, whatever they are, may therefore be attributed to extreme, perhaps excessive, caution and conservatism of advisers. That may change with time, however, as some firms now on the edge of the bell-shaped curve of cautiousness about use restrictions set examples that those now closer to the center-line of the curve decide to follow. We turn therefore to a survey of things that patentees may be able to accomplish by resorting to the Mallinckrodt doctrine.

B. Wealth Transfer Effects

The thrust of the new rule is to increase the economic power of patentees and, presumably, their ability to devise ways to extract revenue from their products and the technology that the products embody. By the same token, the thrust of the new rule is to take power and/or revenue away from customers. n39

The likely consequences of the Mallinckrodt judgment on the hospital-customers of the patentee, Mallinckrodt, illustrate the wealth transfer effect. The hospitals must now obey the "single use only" notice that Mallinckrodt placed on the units that it sold them, or they can be held patent infringers. The hospitals will now pay an additional $ 20 or $ 30 for each unit that they use, above what they paid the infringer, Medipart, who previously recycled the products. Presumably this will affect what they charge patients, medical insurers, the government, or whoever ultimately pays for the products. On the other hand, the patentee, Mallinckrodt, will now make an additional $30 to $40 for each unit that the recycler would previously have handled. Those effects are illustrative of the kind of wealth transfers that the Mallinckrodt decision can cause. n40

Probably, most of the time the result of application of the Mal [*11] linckrodt doctrine will be a wealth transfer from customers to the patentee. That may be harmful, beneficial, or neutral, depending on the facts of the case and one's premises. n41 At this point in the discussion, one may appropriately note the phenomenon of wealth transfer, but it is not useful to try to draw any further conclusions except in the context of specific types of transactions. Again, the facts of the Mallinckrodt case illustrate this.

The single-use restriction of the Mallinckrodt case is a fairly unusual patent-related restriction. n42 The Mallinckrodt opinion does not disclose why Mallinckrodt was unwilling to recycle the devices. Possibly, it was uneconomical for Mallinckrodt to do so, given its business strategy. For example, the cost of manufacture of the plastic boxes and associated equipment may have been less than $10. Mallinckrodt was able to charge hospitals approximately $50 for the units because that was what they were worth to the hospitals, given the utility of the product, the stringency of regulatory requirements, and the hospitals' available alternatives. In other words, the technology was probably worth approximately $40 per unit. Medipart was able to clean and refurbish the units for probably about the same cost as, or maybe a little less than, Mallinckrodt's $10 cost of manufacture. It could then supply the refurbished units to hospitals for $15 to $20 and still make a profit.

Was that because Medipart was a more economical producer than Mallinckrodt? Was recycling more efficient and cheaper than selling only new units? Maybe, but surely not by much, given the overall price structure. A simpler explanation is that hospitals would prefer to pay Medipart $15 or $20 dollars than pay Mallinckrodt $40 or $50. That option was a possible transaction because Medipart was not trying to extract the full per-unit or per-use value of the technology, as Mallinckrodt was. One way of looking at this, then, although not the only way, is that Medipart was able to do business at a profit to itself and the hospital-customers by giving away Mallinckrodt's technology for free.

Mallinckrodt was unsympathetic to this sort of Robin Hood-ism, which was contrary to Mallinckrodt's business strategy. Hence the patent infringement suit. Obviously, Mallinckrodt would assert that it could not afford to develop and bring to market inventions of this type if it could not do so profitably. Without getting bogged [*12] down in the rights and wrongs of each side's possible arguments, one may conclude that presence of a wealth transfer effect need not be outcome-determinative. Here, the hospitals' protest against transferring their wealth to Mallinckrodt can be counterbalanced, at least in the first instance, by Mallinckrodt's protest against Medipart's transfer of Mallinckrodt's wealth (that is, the value of its technology or intellectual property) away from Mallinckrodt.

C. Controlling Supplies, Repairs, and Parts Replacement

The Mallinckrodt case is not unique in involving technology whose value is proportionate to how much or how many times it is used. That is probably more the rule than the exception. n43 Other examples exist; medical device companies are not the only patentees vexed by recyclers and comparable intervenors in the patentee's selected chain of distribution. For example, many laser printer and photocopy machine manufacturers are unhappy with toner refillers. Toner refillers take an empty, used toner cartridge, drill a hole in it, fill it with new toner, close the hole, and charge half the price of a new, officially filled toner cartridge. n44
What if laser printer manufacturers take to selling new, patented toner cartridges with a label saying "For One Use Only"? While the effect of the restriction may be the same as that of a tiein restriction, n45 the form of the restriction is not. n46 The form is identical to that of the restriction in the Mallinckrodt case, which [*13] the Federal Circuit considered not to be the same as a tie-in. n47 Some courts, perhaps most, will be impressed by the argument that there is no explicit agreement between the machine manufacturer and the customer that the customer must buy toner only from the machine manufacturer. All that the machine manufacturer does is refrain from granting any license under its patent. n48 (Well, it does a little more than that. It places a label on the cartridge advising everybody that it is granting its customers no license under its patent except a single-use license.)
Still other refurbishers may be swept within the scope of Mallinckrodt. Consider the doctrine of repair and reconstruction, obliquely involved in the Mallinckrodt case. Under the patent law, the owner of a product has an unlimited right to repair the product by replacing one or more worn or otherwise unsatisfactory parts (elements) of the product. n49 Sometimes, however, purchasers of patented devices exasperate patentees by refurbishing the devices instead of buying new ones--even where the patentee makes it clear to the customers by the construction of the device that the seller expects them to replace worn-out element A of the combination A + B + C, but not to replace B when it wears out. n50 Nevertheless, this conduct is not patent infringement, because the Supreme Court has held that the "mere replacement of individual unpatented parts, one at a time, whether of the same part repeatedly or different parts successively, is no more than the lawful right of the owner to repair his property." n51
Hypothesize a case, however, in which the seller labels the device: "Licensed for use only until element B wears out." Under the Mallinckrodt decision, that label would provide a detour around the Supreme Court's rulings on repair. n52

D. Modifications and Enhancements

The same principle seen in the repair cases may be extended to modification or enhancement of patented devices. Under federal [*15] patent law, the owner of a patented machine has the right to modify it to improve its speed, rate of production, and manner of operation. n53 For example, the Supreme Court has held that the exhaustion doctrine entitled the owner of a patented sardine-canning machine to modify it to pack a different number of sardines into a different size of can. n54

Hypothesize a case, however, in which the patentee-machine manufacturer places a notice on the machine stating "machine licensed for use only as is, without modification." The sardine canner of the preceding case would now be an infringer if it modified the machine to put more or fewer sardines into bigger or smaller cans, or to operate more quickly or slowly. The enhancement of computer programs subject to patent protection would seem to be in the same category. n55

E. Charging Different Prices in Different Market Niches

More broadly, the Mallinckrodt decision may affect how manu- [*16] facturers address niche or differentiated markets. The same technology may have different values in different market niches. A Motorola 68030 or Intel 80386 microprocessor chip, for example, may have a different comparative advantage, and thus a different value, in the following different possible end uses: personal computer, workstation, arcade video-game machine, home video-game console, microwave oven, or automobile. In some of these niches, a much cheaper Z80 microprocessor chip may be just as good; in others, nothing else is as good; in still others, another more expensive microprocessor chip may function much better. Accordingly, the price that a 68030 or other microprocessor chip can command should vary from niche to niche.

Yet, under the pre-Mallinckrodt exhaustion doctrine, once Motorola (or any other semiconductor chip manufacturer) sells a chip, the customer is free to use the chip as it pleases or to resell the chip across markets. The chip manufacturer cannot succeed in maximizing revenue by charging each group of customers a price for the chip commensurate with the value to the consumers of the technology. In theory, at least, a microwave-oven manufacturer customer who paid a low price for a 68030 chip could resell at a slight markup to a workstation manufacturer customer who would otherwise pay the chip manufacturer a much higher price. The transaction would occur for the same reason that hospitals dealt with the recycler in the Mallinckrodt case. The hospitals preferred paying a lower price for the product.

Suppose, however, that a chip manufacturer could sell a 68030 chip "for microwave oven use only" or "for home video use only" and collect patent infringement damages from those product-manufacturing customers who disregarded the notice restriction. This would permit the chip manufacturer to extract the full value of the technology to the given user and to charge an appropriate price for each niche market: all that the traffic would bear.

An analogous scenario already occurs when some microprocessor chips are rated and labeled as capable of running properly at one speed, such as "25 MHz," when they have a faster speed capacity, for example 33 MHz. n56 Their manufacturers sell these "under [*17] rated" chips at the lower price that their lower-labeled speed ratings will bring, rather than at the higher price that their actual speed capabilities could command. n57 Currently, customers do not get sued for patent infringement if they run a microprocessor chip labeled "25 MHz" at 33 MHz. Accordingly, some computer manufacturers attempt to use under-rated chips at higher speeds in order to reduce their parts-procurement costs. n58 Microprocessor manufacturers frown on this practice but do not actively suppress it, presumably for want of any effective mechanism of suppression. n59 The Mallinckrodt doctrine could change all of that by providing an effective mechanism of suppression.

Similar niche-market considerations apply to the pricing of patented drugs, n60 herbicides, and insecticides. n61 The possibility of ar- [*18] bitrage may defeat efforts by manufacturers of these products to maximize their revenue by charging prices geared to the full value of the technology under patent (that is, basing the price on the demand in the niche market). n62 Under the Mallinckrodt doctrine, however, a purchaser who arbitraged patented products in violation of a label restricting use to a particular market would be a patent infringer. The threat of a patent infringement suit could deter arbitrage.

Moreover, the Mallinckrodt doctrine does not limit patentees merely to using field-of-use restrictions, such as "run only at 25 MHz" or "to be used only in car radios." More generally, patentees may totally prohibit any resale of a patented product or may prohibit its resale or other transfer except as a component of a larger system. For example, in the microprocessor chip hypothetical case just discussed, the chip manufacturer might sell its chips to a customer in a low-price market (for example, an automobile manufacturer or microwave oven manufacturer) with a restrictive notice stating "for use only as automobile component, no resale as [*19] such." n63 That would make any purchaser who resold the chip (presumably, to customers in a high-price market, such as workstation manufacturers) a patent infringer and would also make a patent infringer out of the customer to whom the chip was resold (the notional workstation manufacturer who would use the chip other than as a component of an automobile).

Further, a patentee may control resale of a product via certain distribution channels. n64 Such a patentee might therefore utilize a restrictive notice such as "for resale only to dental market." n65 Under Mallinckrodt, customer disobedience of the notice — for example, by resale to stores, jobbers, or hospitals--would be patent infringement.

In short, Mallinckrodt suggests that when patentees sell adequately labeled patented products they may do all of the things that in the past were held unlawful or, at least, in excess of the patentees' rights under the patent laws. n66 The only exceptions left would appear to be price-fixing and tie-in restrictions. n67


A. Who Needs the Mallinckrodt Doctrine?

At this point one may ask whether manufacturers wanting to effect the marketing arrangements described above really need the Mallinckrodt doctrine and resale restrictions to do so. Could not patentees accomplish the very same thing by granting limited licenses of the "make and vend" type approved by the Supreme Court since 1926 or earlier? n68 If it is important to a patentee to restrict what a customer does with a patented product, shouldn't the patentee license the manufacturing rights of the product instead of selling the product to the customer? Before Mallinckrodt, the same restriction that the exhaustion doctrine forbade when the restriction was part of a sale would probably pass legal muster when rewritten as a limitation on the scope of a license to manufacture the same product. If the licensee exceeded the limitation, the licensee's conduct would be patent infringement, unless the licensee could prove a patent misuse or an antitrust violation. n69 Ordinarily, in such a case, this would require proof of a substantial anti-competitive effect in the marketplace, n70 which is no easy task for the licensee. It seems that the Mallinckrodt doctrine is not needed since a patentee can just as easily make licensing limitations as post-sale restrictions.

However, the facts of the Mallinckrodt case illustrate why it is often infeasible to use a "make and vend" license instead of selling with a restriction. It would not have been practical for Mallinckrodt to have licensed the hospitals to manufacture the patented units. First of all, hospitals are not going to go into the plastic injection molding business (or even find a subcontractor to do it) just to make their own plastic boxes for trapping radioactive mist from imaging agents. n71 Nor were the components of the device, such as [*21] its plastic box, tubes, and valves, items that hospitals could readily purchase off the shelf and then combine without further machining or processing, even if hospitals for some reason wanted to engage in do-it-yourself purchase and assembly of the parts. In short, it is one thing for a General Electric to license a Westinghouse to make and vend light bulbs, n72 but quite another for Mallinckrodt to license hospitals to make their own mist-trapping units.

More generally, it is not always feasible to market technology by licensing it. n73 Economies of scale may favor having a single manufacturer, rather than a series of manufacturing licensees. n74 To become a manufacturer, one may need technology transfer. This transfer of technology may involve the patentee's loss (or risk of loss) of valued trade secrets, which may not be acceptable to the patentee. n75 Whether it is more efficient or economical to make or buy a technology-embodying component of one's end product is highly fact and situation specific. At times, the make-or-buy [*22] calculus favors buying, in which case customer restrictions, rather than license limitations, necessarily come to the fore, if the patentee's business strategy calls for limiting or restraining loss (or risk of loss) of its valued trade secrets.

The point is that patentees sometimes employ customer restrictions, rather than comparable licensing limitations, for business or industrial reasons. It is not necessarily a case of the patentees' obstinacy. Until Mallinckrodt, however, one route was lawful and the other was not.

B. Alternative Expedients

Limitations on the scope of licenses to manufacture a patented product are not always feasible because it may not be practical for customers to become manufacturers of the patented product. Sometimes, of course, it is practical. Moreover, limitations in manufacturing licenses and post-sale use restrictions are not the only possible ways to restrict customers' use of patented products.
A combination of two other expedients deserves consideration. First, patentees may rely on different claims-drafting techniques. n76 Then, second, they can structure transactions on the basis of limited licenses of such claims. Appropriate claims-drafting techniques coupled with careful structuring of transactions may permit counsel to avoid many of the legal problems that occur because of the exhaustion doctrine. To be sure, doing that may raise questions of form versus substance, but intellectual property law is a field in which the form of things tends to be very significant, often overshadowing so-called substance. n77 Substance is always elusive but form is readily discernible.

The exhaustion doctrine applies to a patentee’s sale of a product with a restriction on how the customer could use the product. n78 [*23] When this pattern of conduct does not occur, the exhaustion doctrine, at least as such, does not apply. Various ways to break out of the exhaustion doctrine's transactional paradigm may be devised.

One way may be illustrated by reshaping the facts of the Mallinckrodt case. Mallinckrodt had five apparatus patents on various aspects of the device: one on the overall combination of parts and the others on elements or sub-combinations within the device. n79 The dispute that occurred in the Mallinckrodt case occurred because Mallinckrodt sold the patented apparatus to hospitals, and the hospitals then decided to exercise their rights as property owners in disregard of Mallinckrodt’s notice that the hospitals should use their property only one time. n80 Mallinckrodt sold the hospitals exactly what the patents covered, nothing less. n81 (There was nothing less to sell than what the claims covered.)

Suppose, however, contrary to the facts of this case, that Mallinckrodt also had a separate method patent or separate method claim — which can usually be obtained n82 — directed to the use of the apparatus. n83 Now, suppose that Mallinckrodt restructured the [*24] transaction to avoid restrictions on the use of what it sells. Probably, there are many ways to do this. One way would be to market the devices and the method as separate items, or at least as separately priced items. For example, the device and the license under the various product patents might be considered to be worth $20, while a license to use the hypothetical method patent (which Mallinckrodt did not have) once might be worth $30. It would therefore be appropriate to sell a device and a one-time method license for $20 + $30 = $ 50. Mallinckrodt would advise customers that they may obtain as many additional one-time method licenses as they want for $30 each. Alternatively, Mallinckrodt could market the devices with a label or other notice advising that no license under the method patent was conveyed by the sale of the device. The notice would add that method licenses were readily available, but must be contracted for separately.

Licenses comparable to these have been upheld as legal in litigation in the textile machinery industry where machines for making certain kinds of yarn are sold for a price that does not include a free license to make yarn with the machines (even though the machines have no other substantial use). n84 Yarn manufacturers must obtain a separate use license to use the machines to make yarn. n85 The royalty for the use licenses is based on so much per foot (or other unit) of yarn made by the machine. n86

[*25] The rationale for the arrangement is that the value of the machine as such (as iron, so to say) is based on approximately how much it costs to make the machine, but the value of the machine as a yarn maker is based on how much yarn it makes. Hence, the only rational way to set the royalty on the machine properly is to take both factors into account by summing them. n87

Probably, a court would uphold such an arrangement as legitimate under the prior exhaustion doctrine case law, n88 and also as “reasonable.” The patentee imposes no explicit restriction on customers’ use of their property. Rather, the patentee merely “fails to grant a license” under the patentee’s separate method patent. n89 There is no sale of the subject matter to which the unlicensed patent is directed. n90 The case law in the main indicates that, in such circumstances, the exhaustion doctrine is inapplicable. n91

Unauthorized use is therefore patent infringement, unless the presence of antitrust violation, misuse, or another defense prevents patent enforcement in the circumstances. There is no antitrust violation if the conduct is reasonable, which is to say, unless a showing of unreasonableness can be made. n92 That means the arrangement passes muster unless an adversary makes a convinc- [*26] ing showing that the arrangement is wrongful and probably generally harmful to competition in the marketplace. n93 Whether that can be done depends on the facts of the case, but the patentee has the advantage (i.e., superior tactical position). There is no misuse unless an adversary (infringer, aggrieved customer, etc.) can show that for the court to enforce the patent, under the circumstances, would help effectuate a wrongful or anti-competitive scheme. n94 While showing this is a lesser burden than showing an actual antitrust violation, nevertheless, the advantage clearly remains with the patentee. n95

Moreover, the principle of addressing use and the product separately can be extended to other contexts. A product patent on a microprocessor structure can be parallelled with a patent on a method of processing signals in such and such a way. n96 One might then separately license use of the method at 25 MHz speed and 33 MHz speed at different royalties. n97 By the same token, if an insecticide is worth one price for killing moths and a different price for [*27] killing mosquitos, one might obtain one or more method patents (or perhaps different claims of one method patent), respectively covering methods of killing moths and mosquitos. One would then license the different fields of use at different prices. n98

It is thus seen that in many situations the functional equivalent of a post-sale use restriction can be realized by imposing a license limitation on the scope of licensees’ use of a patented method. By the same token, such a license limitation would escape censure, under pre-Mallinckrodt antitrust and misuse law, to the same ex- [*28] tent as a comparable manufacturing-license limitation would. (It would be unrealistic, of course, as well as unwise, to suggest that one can make a horizontal cartel appear legitimate by this means.)

In many instances, therefore, in which a patentee might want to impose post-sale restrictions on customers, the patentee can accomplish the same result by having appropriately ramified patent coverage and licensing the patents carefully. Are such alternative expedients superior or inferior, from the patentee’s vantage point, to proceeding under the Mallinckrodt doctrine? The alternatives are superior if one considers reliance on the Mallinckrodt doctrine risky because it may be overturned and the old rule restored. But the Mallinckrodt doctrine is superior to the alternatives on several other counts. First, it is more work to devise appropriately ramified patent coverage, and it causes at least some increase in patent procurement costs. Second, at the time one procures the patents the nature of the market may be uncertain, so that it is not possible to foresee how to devise the kind of coverage that facilitates the kind of field-of-use licensing described above. n99 Third, the person structuring licensing transactions is not always the person responsible for patent procurement. The patent structure may already be in place, and it may be too late to acquire appropriately ramified coverage by the time the person devising a licensing structure appears on the scene. Finally, it may be easier — in terms of negotiating posture and selling on a “take it or leave it” basis — simply to place a restrictive label on a product than to bargain with a prospective licensee over the scope of a license limitation. n100

In sum, to a very significant extent it is possible to avoid (indeed, circumvent) the exhaustion doctrine by careful patent and license drafting. But it is much easier to use the Mallinckrodt doctrine than to carry out the necessary “fancy footwork” that the [*29] alternative expedients require. From a patent owner's standpoint, therefore, the Mallinckrodt doctrine is probably superior to the alternative expedients discussed above.


This section addresses the Mallinckrodt doctrine’s wisdom, without regard to precedential or doctrinal considerations. The wisdom of the doctrine is evaluated, ultimately, in terms of the public interest in promoting the progress of useful arts. n101 This public interest, rather than simply private benefit or gain for patentees as an end in itself, is the paramount interest that the patent system serves. n102

A. Significance of Wealth Transfers

Ordinarily, there must be winners and losers. Peter is robbed to pay Paul. In theory, the result of an improved economic order is a larger pie for everyone, considered as a total of all slices, n103 that is, aggregate national wealth. n104 In practice, and certainly in the short run, wealth is simply transferred from the losers to the winners. In the Mallinckrodt case, for example, Mallinckrodt profited at the expense of its hospital-customers and those to whom the hospitals pass on their costs of doing business. Is that legally neutral? A matter of indifference? Surely, not to the losers. n105 It is reasonable [*30] to expect that they will protest. They may, to the extent feasible, seek to develop political opposition in order to restore a status quo favorable to them. Accordingly, the transfer of funds from customers to patentees that results from replacement of the exhaustion doctrine by the Mallinckrodt doctrine probably needs justification, if at all, in terms of some kind of calculus of social benefit. Ordinarily, this would mean that the social benefit of the change in law to favor the new winners would outweigh the pain or cost to the losers sufficiently to stifle their groans of protest. Is that true here?

Why should wealth be transferred to patentees from their customers? What is the social benefit? Presumably, a patentee in the position of Mallinckrodt will say that the added revenue it gains will encourage it (and others in its position) to engage in more inventive/innovative activity and that this added inventive incentive results in faster technological progress, from which all of society greatly benefits. Perhaps this is true, although speculative. Perhaps, it is also true (an unarticulated, but necessary, premise of this argument) that the resulting increased rate of technological progress, relative to that which would otherwise be obtained, is socially needed to an extent which justifies the required social payment (the cost to customers of paying for the patentees’ incentives). Perhaps all of this is true, but those who assert this sort of proposition ordinarily provide no hard evidence to support the assertion. This is, therefore, probably the weakest point in the argument for the Mallinckrodt doctrine.

The same argument may be pressed further: because of the existence of the alternative expedients discussed earlier, n106 the same transfer of wealth will occur anyway. The only difference is that forcing patentees to use the other expedients will raise their costs without equivalent benefit to customers. The choice is between having patentees mulct customers cheaply and efficiently under the Mallinckrodt doctrine, or else having them do so inefficiently and at greater cost to themselves under an exhaustion doctrine that they evade by adroit use of expedients. Consequently, resistance is futile, and customers should submit to the new order.

If the premise is valid, so is the conclusion. Doubtless, some customers will challenge the premise and believe that patentees will, at least sometimes, not succeed by means of alternative expedients [*31] or will find them too costly to use. Customers may conceive that patentees’ appetites for restrictions are subject to an elastic demand curve, so that increasing the cost to patentees of engineering these wealth transfers would decrease the amount of wealth transferred from customers to patentees. Therefore, this probably still remains the weakest point in the argument.

B. Considerations of Efficiency and the Relevance of Alternative Expedients

When economies of scale favor large scale manufacture of a product embodying patented technology, rather than smaller scale manufacture of the same product by several licensees, considerations of industrial efficiency call for marketing the technology by the sale of products, rather than by licensing. n107 If all other things are equal, for the law to compel licensing rather than sale is wasteful and irrational. (To the extent that all other things are not equal, however, the question becomes more difficult.)

Economies of scale, or a similar phenomenon, may also occur over passage of time. Manufacturing processes often involve a “learning curve” which results in a decrease in unit costs as the cumulative total historical production volume increases. That is, the more production, the lower the price. n108 Manufacturing costs for semiconductor chips are halved each time that cumulative total historical production volume quadruples. n109 Accordingly, permitting a patentee-chip manufacturer, whose position is like that of Motorola in the hypothetical cases discussed earlier, n110 to opt for making and selling chips to customers whose use of the chips is controlled by restrictions, rather than requiring that the manufacturer impose corresponding limitations only by way of manufacturing licenses, may allow the patentee-manufacturer to descend the [*32] learning curve faster and thus lower its unit costs of manufacture. In theory, such economies in manufacture will lead to better allocation of economic resources. That, in turn, according to the theory, will eventually trickle down to the public in various ways.

Similar considerations may apply to the alternative transactional expedients discussed earlier. If the same effect can be accomplished by alternative expedients, and all other things are equal under either regime, the law achieves no useful purpose by forcing parties to use one, rather than the other, expedient. n111 Moreover, if one form of transaction is more costly to administer than the other, it is wasteful to force its use. Drafting claims and licensing them, as suggested earlier, n112 is probably more costly than simply selling products with restrictive notices affixed to them. To the extent that these suppositions are facts, policy considerations would favor abandonment of the exhaustion doctrine as a rule of law in favor of the Mallinckrodt doctrine.

The problem with this analysis is discussed in the immediately preceding section; the claimed economies that come from allowing use restrictions are not costless for everybody concerned. Giving use restrictions legal force results in a wealth transfer to patentees from customers over and above those savings to patentees generated by efficiency. Thus far, no one has devised a sharing arrangement that will buy off, or at least substantially mute, customer protests over the wealth transfer from their pockets to patentees’ pockets. Nor has a scheme been devised to determine when the customer-patentee wealth transfers are deserved and when they are not, so that the courts can sort out, in terms of enforcing postsale use restrictions, when it is or is not legitimate to rob Peter to pay Paul.

[*33] C. The Issues of Undue State Interference With Private Volition and Oppressiveness

The Mallinckrodt doctrine makes it patent infringement for customers to disobey restrictive notices accompanying patented goods. Implementation of the doctrine will increase the amount of state action, in this instance court interference with defendants. It will subject more firms, or will subject firms more often, to the legal impact of the patent system. This will decrease the extent to which the previous beneficiaries of the exhaustion doctrine will be able to lead quiet lives without worrying about patent infringement threats and without the significant transaction costs (for example, legal fees) of having to deal with them. It is difficult to quantify the negative social value of this possible added degree of governmental (i.e., judicial) harassment of business. But the argument, however speculative, resonates the tuning fork of the Zeitgeist.

Oppressiveness, or appearance of oppressiveness, may be a related policy factor. It may appear oppressive for patentees to impose their will on customers, i.e. to force customers to use their own property only in some ways and not in others. That is what the exhaustion doctrine was about in its inception. n113 But the oppression may be more apparent than real. If we view the customer as an individual person, it may appear that an overreaching corporate patent owner is relentlessly determined to crush the individual (a sturdy yeoman) and deprive him of his freedom as a property owner. n114 In practice, however, the customer is as likely to be a vast corporation as the patentee is. The property, moreover, is probably not someone's automobile, chair, or plough. More likely, the property is a large quantity of agricultural chemicals or microprocessor chips, destined for some industrial use that will never be recognized by members of the general public. If there re- [*34] ally were oppression, that would be a strong argument against supplanting the exhaustion doctrine with the Mallinckrodt doctrine. But fear and loathing of oppressiveness — once a foundation of the exhaustion doctrine, indeed its spiritual fuel — may rest on a figment of the imagination at this end of the industrial era.

D. The Net Balance

The foregoing discussion has tacitly assumed a binary choice of rules — either the traditional exhaustion doctrine (any sale whatsoever liberates the sold product from the patent monopoly) or else the Mallinckrodt doctrine (that is, no exhaustion of rights by sale, ever). Indeed, the Mallinckrodt rule, as the Federal Circuit stated it, is essentially the negative or direct opposite of the exhaustion doctrine. Mallinckrodt holds that post-sale restrictions on customer use will be enforceable by suits for patent infringement, unless the disobedient customer can establish two things. First, the restriction must not be “reasonably within the patent grant.” n115 Second, the restriction must be unjustifiable because it is an unreasonable restriction. n116

The Mallinckrodt decision does not indicate which restrictions on customer use, if any, would not constitute patent infringement, although it seems to suggest that violation of price-fixing and tie-in restrictions is not patent infringement, because the restrictions are unreasonable per se. n117 But that too may well be questioned given the present state of the law. n118 In any event, the Federal Circuit [*35] would probably hold any non-price-fixing restriction or non-tie-in restriction lawful (and, therefore, disobedience a patent infringement) unless it were shown to be an indisputable, cartel-type, antitrust violation. (This is a fair characterization of how the General Electric/General Talking Pictures doctrine, permitting limitedscope licenses to make and sell a patented product, has operated in actual practice. The Mallinckrodt decision has extended the same rule to customer restrictions following a sale of a patented product.)

To convince the Federal Circuit that a patent-related restriction is an antitrust violation would probably be a formidable task. The Federal Circuit would likely conclude that sales of patented products on any restrictive terms whatsoever would be much less restrictive, on balance, than the patentee's total refusal to market the patented technology. n119 Under that reasoning, there would be no antitrust violation. For these reasons, it is fair to say that the Mallinckrodt doctrine, as the Federal Circuit has announced it, is the complete opposite of the exhaustion doctrine. The Mallinckrodt doctrine is a rule under which virtually all customer restrictions qualify for patent infringement enforcement, just as virtually all license limitations do under the General Electric/General Talking Pictures doctrine.

The arguments for and against supplanting the exhaustion doctrine with the Mallinckrodt doctrine, described above, call for a balancing of, and trade-off among, a group of fairly speculative interests. Moreover, some of the interests are wholly incommensurate. How does one balance the alleged transactional efficiency of post-sale restrictions (relative to alternative expedients) against their alleged oppressiveness, let alone against the social cost of some unquantified greater amount of state intrusion into private volition?

Although it may be impossible (as the author believes it is) to balance the pure, polar opposite formulations of the Mallinckrodt doctrine and exhaustion doctrine against one another, it may nonetheless be possible to compare both with an intermediate rule and conclude that the latter would be wiser than either polar doctrine. The Article now turns to that question.


If a more sensible intermediate rule could be formulated, it might appear wiser and more conducive to the public good than either the exhaustion doctrine or the Mallinckrodt doctrine. However, a preliminary point deserves mention. In discussing the relative wisdom of the Mallinckrodt and exhaustion doctrines, nothing has been said so far of who should gauge the wisdom. The patent statute represents a carefully negotiated bargain between the public and inventors. The public trades a limited amount of monopoly power and economic reward to inventors (and those who subsidize invention) as the quid pro quo for increased inventive disclosures. The bargained-for consideration reflects a century of case law under the exhaustion doctrine, which the legislature has ratified by not changing the consideration by statute. If the consideration is inadequate and should be increased, ordinarily we may expect that the legislature should undertake that determination and provide corrective action.

Even assuming (as we may for the moment) that the Mallinckrodt doctrine is wiser policy than the exhaustion doctrine, it is troublesome that the determination and corrective action here were judicial rather than legislative. There are many reasons why legislation by the legislature is preferable to legislation by the judiciary, and they hardly require exposition. Accordingly, it is preferable that the intermediate solution, considered and proposed here, be legislative rather than a further evolution of case law. In principle, however, the courts are as free to create another fiat to replace that of the Mallinckrodt doctrine as they were in 1992 when the Federal Circuit decided Mallinckrodt. Therefore, it would also be possible to implement the proposed solution judicially.

A. The Broad Concept

What would be a sensible intermediate position between the traditional exhaustion doctrine and its polar opposite? In principle, it would have to be something like this: Instead of having all customer restrictions enforceable (the Mallinckrodt doctrine) or all of them unenforceable (the exhaustion doctrine), an intermediate rule would make customers’ disobedience of some restrictions patent infringement, while disobedience of others not. n120 What [*37] would be the distinguishing factors? Unless the criteria were very clear, the costs of administration and uncertainty would likely wipe out any social gains. Some criteria for justifiability are readily definable, and present law probably would already defer to them. For example, restrictions shown necessary because of health and safety considerations, or protection of product reliability and thus reputation, are probably good candidates for status as justifiable and thus reasonable. n121

What of preventing arbitrage, so that the price for technology-embodying goods can be geared to their value in different market niches? That does not seem unreasonable, but it is unclear how far the principle should extend. What of regulating customers’ repair, modification, enhancement, and supply replenishment? That kind of restriction would seem excessive and unjustified in terms of compensating public benefit. Yet, it may be difficult to show a difference in principle between regulating arbitrage and regulating repair, modification, and replenishment, without a persuasive rationale for distinguishing them. Presumably, we require more than an ad hoc checklist.

The Mallinckrodt test, borrowed from the General Electric decision, n122 is remarkably short on substantive content. As suggested above, it has been tantamount to a blank check for licensors when applied to manufacturing licenses. n123 Therefore, at the very least, a test of this type must be rewritten to provide more clarity and more intrinsic substantive content. Moreover, the substantive content of the legal test would need to correspond to something more intuitively fair and even-handed, and less question-begging, than "within the financial reward to which patentees are entitled" n124 if a persuasive case is to be made for the wisdom of the test.

Needed here is a proper answer to the question, "Why should society agree to have the law enforce restrictions and limitations that a patentee wants others to obey?" Perhaps there are other [*38] answers, but the best one of which the author is aware is that society should agree only if it stands to realize a sufficient net benefit from doing so. n125 A similar question, leading to a similar answer, is "Why should society enforce a competition-restrictive covenant, such as one associated with the sale of a bake shop?" n126 Society’s reason for enforcing the restriction is that it (a) makes some possibly desirable transactions feasible that otherwise would not occur because they would be infeasible — for they would be unprofitable or even harmful to one party without the restriction; and (b) does not greatly injure other interests of society. This is, of course, the rule of reason doctrine of Mitchel v. Reynolds, n127 which permits and makes enforceable “necessary” and “ancillary” restraints. n128

Therefore, the compromise for which we are groping may likely be framed as some variation on the doctrine of Mitchel v. Reynolds. For example, under this approach, disobedience to a customer restriction might be patent infringement only if the patentee showed that the restriction was necessary to make it feasible to market the technology, and make available to the public (the marketplace) products embodying the patented technology. This principle might furnish one element of a compromise test.

The concept of “within the scope of the patent grant,” mentioned earlier, has a kernel of sensible policy. But the concept cries out for further refinement and specificity. This could be supplied by reference to the claims of the patent. The claims of a patent are [*39] the proper measure of the monopoly grant that government awards to encourage creation and disclosure of an invention. n129 Hence, the only restrictions or limitations that may be considered within the scope of the patent grant are those that lessen competition, raise prices, and decrease output only of those products or services covered by the relevant patent.

But, by the same token, the complementary proposition is true. Restrictions and limitations should not lessen competition, raise prices, or decrease output of things outside the scope of the patent’s claims, and thus outside the patent grant. This is not to say that competition must be restrained to the point of an antitrust violation for the state to decline to hold disobedient customers to be patent infringers. That would miss the point. The question is not whether the state should punish the patentee for restraining trade. The proper question is whether the state should punish the patentee’s disobedient customers by making them liable for patent infringement under the patent laws. This is a question about the function of patent law, not antitrust law. Patent law states that patents cover what their claims describe, and the claims describe the patentee’s contribution to society for which the patentee deserves a monopoly. Hence, patent infringement should be defined in relation to markets substantially coextensive with patent claims; it should not be patent infringement to disobey restrictions that are directed to, or which adversely affect competition in, products or services outside the scope of the claims.

B. Proposed Language

A possible addition to the patent code to carry out the concepts described above is now set out, followed by a brief explanation of the language.

35 U.S.C. § 273. Infringement by violation of use restrictions [as proposed]

(a) REASONABLE CONSENSUAL RESTRICTIONS PERMITTED. A patentee may, subject to the provisions of this section, agree in writing with another person that the person’s manufacture, use, or sale of the patented subject matter shall be subject to a reasonable limitation, condition, or restriction.

[*40] (b) REASONABLE NONCONSENSUAL RESTRICTIONS PERMITTED. A patentee may, subject to the provisions of this section, impose on its direct or indirect customer for a patented product a reasonable restriction on the customer’s use or sale of the patented product, by giving the customer adequate notice of the restriction.


(d) REASONABLENESS. A limitation, condition, or restriction described in subsection (a) or (b) is reasonable, as that term is used in this section, if and only if each of the following conditions is satisfied:

(e) ADEQUACY OF NOTICE. Notice is adequate for purposes of subsection (b) if the patentee gives the customer full, unambiguous, written notice of the restriction prior to the distribution of the product to the customer and gives such notice to the customer’s predecessors, if any, [*41] in title to the product prior to the distribution of the product to them.

1. What is Authorized

Subsections (a) and (b) authorize patentees to impose “reasonable” limitations, conditions, and restrictions on licensees and customers. (Reasonableness is defined in subsection (d) of the proposed new section.) Patentees may do so by entering into written agreements with licensees and customers, or by giving customers adequate notice. The proposed section refers throughout in parallel terms to limitations, conditions, and restrictions, since a major aspect of the proposal is to treat them all in the same manner. That is, the proposal intends to address in like terms all of the various expedients by which a patentee can, by contract, curtail the scope of another's use of the subject matter of a patent, and apply the same reasonableness test to each of them.

Subsection (a) states that a patentee may enter into an “agreement” with another person. Ordinarily, the agreement would be a license agreement with a licensee, providing that the other person’s manufacture, use, and sale of a product or other thing covered by the patentee’s patent shall be subject to a given limitation, condition, or restriction. However, subsection (a) is not limited to license agreements and, by its terms, applies to any agreement, such as a bailment or lease. Agreements with agents would also fall within the scope of subsection (a), except to the extent that the law may consider an agent and principal a single entity. In that case, the law would presumably follow the legal fiction that a principal does not agree with itself or limit or restrict itself, and therefore, no limitation, condition, or restriction exists. n130

[*42] Subsection (b) states that a patentee may impose comparable restrictions on “its direct or indirect customer.” Thus, a patentee’s reasonable restriction would follow the goods down the distribution channel, so long as those bound each have adequate notice. Furthermore, subsection (a) is broad enough to permit a patentee to require its licensee to impose, as provided by subsection (b), a reasonable use restriction on the licensee’s customers, based on licensee-to-customer notice. Adequate notice is defined in subsection (e).

2. Legal Effects

Subsection (c) describes the legal effect of limitations, conditions, and restrictions. Paragraph (1) provides that intentional disobedience of a limitation, condition, or restriction that satisfies the requirements of new § 273 is an act of patent infringement. That is, if a licensee or customer intentionally disobeys a limitation, condition, or restriction that is reasonable (and adequate notice has been given, if a restriction imposed on customers by notice is involved), then the licensee or customer is liable for the sanctions that the patent law provides for patent infringement. So far, this is essentially the doctrine of the Mallinckrodt case, n131 except that in the proposed new section there is a requirement that the patentee be “reasonable.”

Infringement liability is subject to any affirmative defense that the patent laws recognize, as is any other form of liability for patent infringement. n132 For example, invalidity of the patent would be an absolute defense. n133 Laches and estoppel remain defenses. n134 Further, the sale of a patented product that has only one use may create an implied license to use the product for that purpose, n135 [*43] despite any label notice to the contrary. In such circumstances, the facts surrounding the sale may create an ambiguity in the notice. On the other hand, in some circumstances a similar sales transaction may not imply such a license, or at least not imply a license without a further royalty obligation. n136 In any event, regardless of the rule applicable to particular fact situations, the proposed new section does not abrogate any affirmative defense recognized under patent law, except to the extent that it overturns the pre-Mallinckrodt rule against treating post-sale use restrictions as patent infringement and subjects them to a rule of reason.

Subsection (c) makes the patent law’s remedies for patent infringement exclusive with regard to violation of limitations, conditions, and restrictions. Under this proposal, federal patent law fully occupies the field and provides uniform national regulation of this aspect of commercialization of patented technology. There should be no room for conflicting, cumulative, or alternative liability under other federal or state laws. Cumulative state regulation — for example, under contract or tort law — could undermine the proposed compromise by striking a different balance of interests, n137 and in any case it would interfere with national uniformity of regulation.

Paragraph (2) of subsection (c) provides that a licensee’s or customer’s disobedience of a limitation, condition, or restriction that is not reasonable (or, if a restriction imposed on customers by notice is involved, adequate notice has not been given) is not actionable. Again, under this proposal federal patent law fully occupies and uniformly regulates the field, so that disobeying an impermissible or unlawful limitation, condition, or restriction cannot be made subject to sanction under any other state or federal law. n138

Paragraph (3) provides that patentees, licensees, and customers are immunized from liability under state and federal law for agree- [*44] ing or yielding to, n139 or imposing, a limitation, condition, or restriction permitted by this section. Section 273 thus provides patentees a complete safe haven for reasonable limitations, conditions, and restrictions. That is necessary to effectuate uniform federal regulation of this aspect of commercialization of patented technology. Otherwise, state antitrust laws, for example, might prohibit restrictions that the compromise proposal authorizes. Furthermore, this section would supersede federal antitrust law to any extent that it prohibited reasonable restrictions. After the decision in Continental T.V., Inc. v. GTE Sylvania, Inc., n140 however, it is doubtful that any reasonable limitation, condition, or restriction would be held to violate the antitrust laws. n141

Paragraph (4) is a savings clause that provides that unreasonable limitations, conditions, and restrictions are not immunized from otherwise applicable laws, such as state and federal antitrust laws. Only limitations, conditions, and restrictions qualifying for paragraph (3)’s safe haven are immunized. The issues of supersession and preemption of other federal and state laws do not arise for limitations, conditions, and restrictions that do not qualify for enforcement and immunization under the new section. Since patent [*45] law does not authorize them, it does not regulate them and there is no need or rationale for uniformity. The new section does not prohibit unreasonable limitations, conditions, and restrictions. Other state or federal laws may or may not prohibit them; that is something that new § 273 leaves to the other laws. The only effect of the unreasonableness of a limitation, condition, or restriction, for purposes of § 273, is that § 273 does not make it patent infringement for a licensee or customer to disregard the limitation, condition, or restriction.

3. Definition of Reasonableness

Subsection (d) defines the reasonableness of a limitation, condition, or restriction in terms of the doctrine of necessary and ancillary restraints’ formulation of the rule of reason. n142 A limitation, condition, or restriction is reasonable if three conditions are satisfied: (1) necessity and ancillarity, (2) absence of anticompetitive purpose or effect outside the scope of the claims, and (3) nonviolation of other positive law. The three conditions correspond to the three paragraphs of subsection (d).

Paragraph (1) requires that the limitation, condition, or restriction must be designed to help overcome an obstacle to commercializing the patented technology — to placing the patented subject matter “on sale” n143 or otherwise on the market. In a license, that would make the restrictive provision ancillary to the lawful main purpose of the license and necessary to effectuate that purpose. This requirement restates, for patent licensing purposes, one of the requirements of the rule of reason described in the leading United States case, United States v. Addyston Pipe Co. n144

Overcoming an obstacle to commercializing the patented subject matter may mean different things in different factual contexts. For example, exclusive field licenses are frequently granted to give a licensee a sufficient incentive to invest the effort and cost of developing a particular market. n145 Without field exclusivity, at least for [*46] a period sufficient to assure recoupment of the initial investment, it may be that no prospective licensee would be willing to develop a new market. It may therefore be reasonable for a patentee to agree to, or impose on other licensees or customers, a limitation or restriction that will provide one licensee, or the patentee itself where it decides itself to commercialize a new market or niche, the exclusivity necessary to overcome this obstacle to introducing the technology or its product into the marketplace.

In other cases, the nature of an invention or the market for the subject matter may require that it be licensed subject to limitations or restrictions. For example, an invention of a machine for making a new kind of yarn may need to be commercialized in two markets that must be licensed separately: the machinery-manufacturer market and the yarn-manufacturer market. The value of the invention in these different markets must be measured in quite different ways, based respectively on machine sales and yarn sales, which may have no determinate relationship to one another. Those facts may require the patentee to license machine manufacture without allowing the machine manufacturer's customers (yarn manufacturers) to acquire an unlimited right of use of the machine by virtue of their purchase of it. n146

A somewhat broader formulation of the first part of paragraph (1) might use the phrase “designed and intended to facilitate one or more legitimate business objectives of the owner of patent rights.” That formulation could possibly include reasonable purposes going beyond overcoming obstacles to placing the patented subject matter on sale, although the author is not aware of any specific deserving examples that the first formulation omits and the second includes.

There are sound objections to such a broader statement of the rule. First, it would entail some loss of clarity (what are “legitimate business objectives”?), and therefore greater administrative costs. Second, it is not specifically limited to the scope of the patent, that is, “the patented subject matter.” It is not justifiable to make some form of conduct into patent infringement when the conduct to be proscribed merely interferes with some (undefined) legitimate objective of the patentee. If no other law safeguards the given interest of the patentee, why should the patent law do so? To do that would at best be unrelated to the statutory and constitutional purpose of promoting progress of science and useful arts, and there- [*47] fore insupportable. It would also beg the question and be a move in the direction of writing blank checks for patentees, which has been a basis of criticism of the Mallinckrodt doctrine in this Article. n147

A further requirement under paragraph (1) is that the scope of the limitation, condition, or restriction not be substantially greater than the necessity giving rise to it. This requirement is also taken from the Addyston Pipe decision’s formulation of the rule of reason. n148

Paragraph (2) of subsection (d) states a requirement that the persons responsible for the limitation, condition, or restriction — ordinarily, the patentee and licensee — must not intend to use the limitation, condition, or restriction to lessen competition substantially in unpatented product or service markets. Moreover, that result must not be the actual or probable effect of their adopting the limitation, condition, or restriction. This is an amplification of the concept that a limitation, condition, or restriction should not go beyond the scope of the given patent, and an attempt to give the concept more specific content. Definition of an unreasonable restraint in terms of whether its either intended or actual effect is substantially anticompetitive is well established in antitrust jurisprudence. n149

Paragraph (3) requires that the limitation, condition, or restriction must not violate any specific prohibition of law — for example, the per se rules against price fixing, group boycotts, customer or product allocations among competitors, and certain tie-ins. n150 In general, those limitations, conditions, and restrictions that violate positive law will probably fail to satisfy the requirements of paragraphs (1) or (2) of subsection (d) as well.

[*48] 4. Adequacy of Notice

Subsection (e) provides that notice of a limitation, condition, or restriction imposed on customers will be adequate if the patentee gives the customer full, unambiguous, written notice of the restriction prior to the distribution of the product to the customer and also gives such notice to the customer’s predecessors, if any, in title to the product prior to the distribution of the product to them.
For example, if a patentee desires to restrict customers from using a product in a particular field — such as the end-user consumer market — the patentee must make sure that customers get proper notice of the restriction before distributing the product to them. One way to do this is to place a label or tag on the package or product, clearly explaining the details of the intended restriction. A customer who receives a unit of the product with its label or tag on it would get notice. However, if the label or tag is removed from a particular unit before a customer gets it, there is no notice unless the patentee can show actual notice despite removal of the label or tag.

For products marketed through distribution channels such as wholesalers and retailers, a customer charged with obeying a restriction must receive notice of the restriction before buying the product. Also, the wholesaler, retailer, etc. must have received prior notice. Under this subsection, a patentee would not be entitled to sell a product to wholesalers, without restrictions, and then impose restrictions on the wholesaler’s customers. Any unrestricted sale of a particular unit of the patented product should liberate the sold unit from the patentee’s control or power to impose restrictions. For any restriction to be valid, it should be imposed before a given product unit enters distribution channels or the general stream of commerce.

C. Comparison of Proposal with Mallinckrodt Doctrine

Is the proposed legal standard better, clearer, and generally wiser than the Mallinckrodt opinion’s standard of "reasonably within the patent grant" n151 and "otherwise justified?" n152 Surely, the proposed compromise is more definite and clear. Moreover, it is sufficiently fair and evenhanded to lead to an acceptable com- [*49] promise, which Mallinckrodt does not.

While the proposed test will not administer itself, it makes it easier to determine what result is appropriate for a given restriction than Mallinckrodt does. Moreover, the proposed judgmental process corresponds more to a sensible evaluation of what intuition indicates are the material considerations that ought to be weighed. To be sure, the proposed test does not indicate, without further information, whether a restriction needed to prevent a serious health and safety problem should be enforceable by way of a patent infringement suit. The same can be said of a restriction needed to keep customers who buy low price microwave oven microprocessor chips from reselling them to workstation manufacturer-customers to whom the patentee wants to sell the same chips at a higher price.

Nevertheless, resolution of these questions is surely better facilitated by trying to decide whether the restrictions (a) overcome a marketing obstacle, (b) go well beyond that objective, and (c) substantially lessen competition in the microwave oven business (or whatever non-microprocessor chip, i.e., non-patent, market is relevant) than by trying to decide whether the restrictions (a) give the patentee revenue to which it is “entitled” n153 as a reward, (b) are “reasonably within the patent grant,” n154 or (c) are “otherwise justified.” n155 There is simply no serious comparison between the two approaches in terms of internal clarity, and objectivity versus subjectivity. For these reasons, the proposed compromise test clearly surpasses the Mallinckrodt test.

Whether the proposed compromise test also surpasses the traditional exhaustion doctrine's substantially absolute rule against treating customers’ violations of restrictions as patent infringement is a more difficult question. The author believes that the proposed compromise is superior to the exhaustion doctrine, but there are considerable difficulties in persuasively establishing the point. First, one must make some debatable assumptions about the relative importance of the incommensurate policy factors discussed earlier. n156 Second, one must place substantial weight on the importance of giving symmetrical legal treatment to patent-related limitations, conditions, and restrictions.

[*50] Of the three legal options under consideration, the traditional exhaustion doctrine provides the least amount of state interference with exercise of private volition. The proposed compromise test provides for somewhat more interference, and the Mallinckrodt doctrine for the most. This leaves the question to be decided as whether the loss resulting from an increase in state interference with private volition is outweighed by other gains.

Sometimes, limitations make it possible, or at least easier, n157 to market a patented product or a product embodying patented technology. To the extent that limitations do that, so too can restrictions. Restrictions are often functional alternatives to limitations in marketing a patented product. But in some factual contexts restrictions work more effectively — for example, where the technology is such that it is much more economical for the patentee or another single source to make and sell products embodying the patented technology than it is for several licensees to do so on a smaller scale. n158 Other things being approximately equal, it is counterproductive to force patentees to resort to limitations in manufacturing licenses when economies of scale favor use of equivalent restrictions on the patentee’s customers. In such cases, industry should be free to use the transactional structure that industrial, rather than legal, considerations dictate.

From this it is fair to conclude that, when a restriction is reasonable, it should be tolerated to the same extent a comparable limitation would, since each causes the same kind of socially desirable business and industrial results. Each can facilitate commercialization of technology in the same way. The compromise proposal is superior to the traditional exhaustion doctrine because it opens the door to accomplishing these desirable results without broadly opening the door to mere abuse of the opportunity without a compensating benefit to society, as the Mallinckrodt doctrine does. That is why, in the author’s view, the gains in legal symmetry and rationality from adopting the compromise substantially outweigh the loss from some increase in state interference with private volition.


The probable impact of the Mallinckrodt decision, if it prevails in U.S. patent law, will be to strengthen the position of patentees in a number of ways. By using Mallinckrodt, patentees can control and suppress undesired kinds of parts repair and replacement, supply replenishment, and equipment modifications. Mallinckrodt paves the way for patentees to market products embodying technology to various market niches at prices that differ in respect to the value of the technology in that particular niche. This was a marketing strategy that the prospect of arbitrage, protected by the exhaustion doctrine, hitherto defeated.

For these reasons, the Mallinckrodt doctrine will result in increased revenue for patent owners. The increase will have two components: one results from greater efficiency in marketing products embodying technology, because patentees will be freer to exploit economies of scale and use sales transactions where they are industrially preferable to licensing others. The second component is simply a transfer of wealth from customers to patentees, which the exhaustion doctrine did not permit, and which has no specific policy rationale in its favor.

There are alternative legal expedients that at times permit patentees to accomplish much the same results as the Mallinckrodt doctrine permits, even when the arrangements are judged under the traditional exhaustion doctrine. Nevertheless, the alternative expedients are not always feasible. Moreover, simply imposing post-sale restrictions on use is often much more convenient or economical. From a patentee’s standpoint, therefore, the Mallinckrodt doctrine is far preferable to the traditional exhaustion doctrine.

On the other hand, the very customer-patentee wealth transfers that make the Mallinckrodt doctrine attractive to patentees generate protest. At least in some hypothetical circumstances — for example, use restrictions against refurbishment, repairs, parts replacement, enhancements, and the like — protests over the wealth transfer effect would appear to be legitimate. In other circumstances, however, such as post-sale restrictions necessary to protect health, safety, and product quality, even the case law under the exhaustion doctrine recognized that imposition, and therefore legal enforcement, of the restrictions can be socially beneficial.

Probably, one cannot decide with confidence whether the Mallinckrodt doctrine or traditional exhaustion doctrine is more socially beneficial, on an overall balance. The dispute over the wealth [*52] transfer effect alone would probably suffice to make the question undecidable on objective, rather than political, grounds. Further complicating factors — the Mallinckrodt doctrine’s increased state interference with private volition, and its possible oppressiveness — make balancing the respective merits of the two doctrines difficult.

Although it probably cannot be decided whether the Mallinckrodt doctrine or exhaustion doctrine is socially more beneficial, an intermediate compromise rule may be superior to each of the polar alternative legal rules. The problem is one of providing a legal rule that is reasonably easy to administer, and distinguishes on an intuitively fair and evenhanded basis between those restrictions that should be legally enforced and those that should not. Further, the rule should accord comparable treatment to restrictions and limitations on the scope of a license, because legal asymmetry between them is irrational. The same things that make either one of them good or bad for society apply with the same force to the other one. Moreover, transactions should not be pushed into one mold or another on the basis of noncommercial, nonindustrial considerations flowing from an asymmetry in the law.

The Article proposes a compromise, preferably by legislation, under which reasonable licensing restrictions and limitations should be enforceable and disobedience to them should be held patent infringement. They should be considered reasonable when they are reasonably adapted to overcome obstacles in marketing the technology. They should not be imposed to lessen competition substantially in product and service markets not covered by the claims of the patent, nor where that is their effect. Nor should they violate specific prohibitions of law.

The proposed compromise is superior to the Mallinckrodt doctrine in terms of having a clearer, more objective, and more evenhanded substantive content. It provides a rational, public interest basis for permitting patentees to impose some restrictions and limitations (but not others) on use of technology-embodying products, rather than simply deferring, as the Mallinckrodt doctrine does, to patentee fiat. The substantive criteria of the proposed compromise rule look to the relation of the restriction to the commercial setting, rather than the Mallinckrodt doctrine’s "reward" and "entitlement" concepts. To the extent that the proposed compromise results in some increased state interference with the exercise of business volition and some transfers of wealth from customers to patentee-sellers, it does so only under a rationale justifying the ac- [*53] tion. Moreover, customer-patentee wealth transfer, when it occurs under the compromise, occurs incidentally to accomplishing some other socially desirable purpose, rather than as an end in itself.

While it is more difficult to support the conclusion that the proposed compromise is superior to the traditional exhaustion doctrine, certainly the compromise is superior in terms of providing greater legal symmetry and rationality. Probably, it is fair to conclude that these gains outweigh any attendant loss in increased state interference with private volition.


* Member, District of Columbia Bar. Professorial Lecturer in Law, National Law Center, The George Washington University, Washington, D.C.

n1 E.g., Motion Picture Patents Co. v. Universal Film Mfg. Co., 243 U.S. 502, 518 (1917) (not patent infringement for purchaser to disobey restriction against use of machine with unauthorized other products); Straus v. Victor Talking Mach. Co., 243 U.S. 490 (1917) (patent owners cannot control patented articles after sale); Adams v. Burke, 84 U.S. (17 Wall.) 453 (1873) (not patent infringement for purchaser to disobey geographical use restriction).

There was a contrary ruling in Henry v. A.B. Dick Co., 224 U.S. 1 (1912), but Henry was overruled four years later by Motion Picture Patents.

n2 E.g., Bauer & Cie. v. O'Donnell, 229 U.S. 1 (1913) (violating resale price restriction not patent infringement); Keeler v. Standard Folding-Bed Co., 157 U.S. 659 (1895) (resale by customer in unauthorized territory not patent infringement); Hobbie v. Jennison, 149 U.S. 355 (1893) (selling to customer for use in unauthorized territory not patent infringement).

n3 General Talking Pictures Corp. v. Western Elec. Co., 304 U.S. 175, 182 (1938) (holding valid field-of-use limitations on scope of patent license to make and sell amplifiers), aff'd on reh’g, 305 U.S. 124 (1938).

n4 id. (patent infringement where licensee with limited rights to manufacture and sell knowingly exceeds scope of license).

n5 United States v. General Elec. Co., 272 U.S. 476, 489-90 (1926) (citations omitted).

In the remainder of this Article, the term “customer” will be used to refer to a person to whom a patentee or its licensee sells a patented product. The term “licensee” will refer to a person to whom a patentee has granted a “make and vend” license of the type referred to in the accompanying text. In the language of the present patent act, such a license is one to “make and sell” the patented product. 35 U.S.C. § 154, § 271(a) (1988). A license to make and sell product X, carries by implication of law a license to use product X. Aro Mfg. Co. v. Convertible Top Replacement Co., 377 U.S. 476, 483-84, 141 U.S.P.Q. (BNA) 681, 685 (1964) (Aro II). Patentees have at times attempted to blur the distinction by designating as use licenses transactions otherwise appearing to be sales. See, e.g., Straus v. Victor Talking Mach. Co., 243 U.S. 490 (1917).

Until 1992, however, courts did not consider a transfer of possession of a patented product, accompanied by what purported to be a license to use the product subject to specified limitations, to be a license. Rather, they considered it to be simply a sale. Id. at 500-01 (“Notwithstanding its apparently studied avoidance of the use of the word ‘sale’ and its frequent reference to the word ‘use,’ . . . courts would be perversely blind if they failed to look through such an attempt as this ‘License Notice’ thus plainly is to sell property . . . and yet to place restraints upon its further alienation . . . .”); Bauer, 229 U.S. at 16 (“It is a perversion of terms to call the transaction in any sense a license to use the invention. . . . To call the sale a license to use is a mere play upon words.”); see also United States v. Univis Lens Co., 316 U.S. 241, 249-50 (1942) (holding sale of semifinished, i.e., partially manufactured, patented product with license to finish to be legal equivalent of sale of patented product); Ethyl Gasoline Corp. v. United States, 309 U.S. 436, 457-59 (1940) (similar arrangement, different patents on intermediate and final stages of product).

n6 General Elec., 272 U.S. at 490.

n7 Id. at 489.

n8 Id. at 490.

n9 Id. at 489.

n10 Id; see Ethyl Gasoline Corp. v. United States, 309 U.S. 436 (1940).

n11 For an illustration of such an unlawful industry cartel, see Hartford-Empire Co. v. United States, 323 U.S. 386 (1945). See also Ethyl Gasoline Corp. v. United States, 309 U.S. 436 (1940), in which an unlawful cartel arrangement was based on a mixture of customer restrictions and manufacturing-license limitations.

For an example of a case in which the court held it lawful for a patentee to place in a manufacturing license a use limitation identical to a use restriction that the court found unlawful when imposed on customers, see United States v. Ciba-Geigy Corp., 508 F. Supp. 1118 (D.N.J. 1976).

n12 See, e.g., National Soc’y of Professional Eng’rs v. United States, 435 U.S. 679, 696 & n.22 (1978) (holding that Society’s ethical canon prohibiting its members from submitting competitive bids for engineering services unlawfully suppressed competition and rejecting Society’s claim that competitive bidding would encourage submission of deceptively low bids and thereby endanger public safety).

n13 See Adams v. Burke, 84 U.S. (17 Wall.) 453, 456 (1873) (receiving royalty for invention deprives patentee of further rights); Keeler v. Standard Folding-Bed Co., 157 U.S. 659, 666 (1895) (purchasing patented articles from an authorized seller creates absolute rights in the purchaser).

n14 Unidisco, Inc. v. Schattner, 824 F.2d 965, 968, 3 U.S.P.Q.2d (BNA) 1439, 1441 (Fed. Cir. 1987) (reselling of product cannot constitute infringement of a patent if purchased from an authorized seller).

n15 United States v. General Elec. Co., 272 U.S. 476, 490 (1926).

n16 976 F.2d 700, 24 U.S.P.Q.2d (BNA) 1173 (Fed. Cir. 1992), rev'g 15 U.S.P.Q.2d (BNA) 1113 (N.D. Ill. 1990).

n17 Id. at 701-02, 24 U.S.P.Q.2d (BNA) at 1174.

n18 Id. at 702, 24 U.S.P.Q.2d (BNA) at 1174. For a more detailed description of the patented device, see infra note 79.

n19 Id., 24 U.S.P.Q.2d (BNA) at 1174.

n20 Id., 24 U.S.P.Q.2d (BNA) at 1174.

n21 Id., 24 U.S.P.Q.2d (BNA) at 1174-75.

n22 Id., 24 U.S.P.Q.2d (BNA) at 1174.

n23 Mallinckrodt, Inc. v. Medipart, Inc., 15 U.S.P.Q.2d (BNA) 1113, 1115 (N.D. Ill. 1990).

n24 Id.

n25 Id. at 1121.

n26 28 U.S.C. § 1295(a)(1) (1988); 35 U.S.C. § 141 (1988).

n27 Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 709, 24 U.S.P.Q.2d (BNA) 1173, 1180 (Fed. Cir. 1992).

n28 35 U.S.C. § 283 (1988).

n29 35 U.S.C. § 284 (1988).

n30 Mallinckrodt, 976 F.2d at 704, 24 U.S.P.Q.2d (BNA) at 1176 (“These cases did not hold, and it did not follow, that all restrictions accompanying the sale of patented goods were deemed illegal.” (construing Bauer & Cie. v. O'Donnell, 229 U.S. 1 (1913); Straus v. Victor Talking Mach. Co., 243 U.S. 490 (1917); Boston Store v. American Gramophone Co., 246 U.S. 8 (1918); General Talking Pictures Corp. v. Western Elec. Co., 304 U.S. 175, aff'd on reh'g, 305 U.S. 124 (1938))).

n31 See, e.g., id. at 708, 24 U.S.P.Q.2d (BNA) at 1179 (“Adams v. Burke and its kindred cases do not stand for the proposition that no restriction or condition may be placed upon the sale of a patented article.”).

n32 See, e.g., id. at 706, 24 U.S.P.Q.2d (BNA) at 1178 (“The principle of exhaustion of the patent right did not turn a conditional sale into an unconditional one.”).

n33 The Federal Circuit said that the Supreme Court's decisions under the exhaustion doctrine must be understood to be limited to situations involving price-fixing and tie-ins (requiring the purchasing of non-patented products or related services with the purchase of patented devices). Id. at 704, 24 U.S.P.Q.2d (BNA) at 1176. However, since Congress amended the patent laws in 1988 to make patent-related tie-ins unlawful only when the patentee has market power, 35 U.S.C. § 271(d)(5) (1988), it would appear no longer to be the case that all such tie-ins are unlawful.

n34 See Mallinckrodt, 976 F.2d at 709, 24 U.S.P.Q.2d (BNA) at 1180.

n35 Id. at 708, 24 U.S.P.Q.2d (BNA) at 1179.

n36 See Richard H. Stern, The Unobserved Demise of the Exhaustion Doctrine in U.S. Patent Law, [1993] EUR. INTELL. PROP. REV. 460.

n37 Some intimation of this may be provided by Intel Corp. v. ULSI Sys. Tech., Inc., 995 F.2d 1566, 27 U.S.P.Q.2d (BNA) 1136 (Fed. Cir. 1993), cert. denied, 114 S. Ct. 923 (1994). Here, the Federal Circuit held that the patent license at issue did not limit the licensee's right to make private-label products for a third company (a competitor of the patentee). Id.. at 1569-71, 27 U.S.P.Q.2d (BNA) at 113840. As a result, the third company was not liable for patent infringement. Id.. at 1570, 27 U.S.P.Q.2d (BNA) at 1140. In so ruling, the Federal Circuit panel used the same sweeping language of pre-Mallinckrodt exhaustion doctrine decisions, perhaps in dictum given the facts of the ULSI case, about how a sale by a licensee placed the sold product beyond the reach of the patent. Id.. at 1568, 27 U.S.P.Q.2d (BNA) at 1138 (“The law is well settled that an authorized sale of a patented product places that product beyond the reach of the patent. The patent owner’s rights with respect to the product end with its sale, and a purchaser of such a product may use or resell the product free of the patent.” (citing Bloomer v. Millinger, 68 U.S. (1 Wall.) 340, 350-51 (1863); United States v. Univis Lens Co., 316 U.S. 241, 250, 252 (1942); Unidisco, Inc. v. Schattner, 824 F.2d 965, 968, 3 U.S.P.Q.2d (BNA) 1439, 1441 (Fed. Cir. 1987), cert. denied, 484 U.S. 1042 (1988))). The court then denied rehearing, en banc, over the dissent of several judges, including the author of the Mallinckrodt opinion. Id. at 1576, 27 U.S.P.Q.2d (BNA) at 1140.

n38 Supreme Court review of the Mallinckrodt decision itself was precluded by a settlement agreement under which the defendant agreed to go out of the challenged line of business and the patentee agreed to forego damages for past infringement. In the twelve years since the Federal Circuit was created as the exclusive federal court of appeals for patent cases, the Supreme Court has not once reviewed a Federal Circuit decision on a substantive issue of interpretation of patent law, although during that period the Court has repeatedly addressed substantive issues of interpretation of copyright law. Given its track record, the Supreme Court is not likely to review the Federal Circuit’s interpretation of when violation of a post-sale restriction is patent infringement.

n39 For a discussion of wealth transfers and their effect on consumers, see Robert H. Lande, Chicago’s False Foundation: Wealth Transfers (Not Just Efficiency) Should Guide Antitrust, 58 ANTITRUST L.J. 631 (1989) (arguing that preventing unfair wealth transfers from consumers to corporations who are exploiting their market power, and not economic efficiency, is the primary goal of antitrust laws). For an in-depth analysis of the economic effects of this kind of shift in power (the wealth transfer effect) see Alan A. Fisher & Robert H. Lande, Efficiency Consideration in Merger Enforcement, 71 CAL. L. REV. 1580 (1987).

n40 In addition, of course, the recycler, Medipart, will exit this business and thus stop making its $10 to $15 per unit on the transactions.

n41 See supra note 39.

n42 But see Cotton-Tie Co. v. Simmons, 106 U.S. 89 (1882) (holding that reuniting severed cotton-bale ties in violation of restrictive notice was not repair, but reconstruction, and therefore patent infringement).

n43 Some examples are discussed infra notes 84-87. See Ward S. Bowman, Jr., Tying Arrangements and the Leverage Problem, 67 YALE L.J. 19, 24 (1957) (explaining use of tying as a counting device).

n44 Toner cartridge structures are often patented. The refill procedure described in text, however, is not patent infringement. See Morgan Envelope Co. v. Albany Perforated Wrapping Paper Co., 152 U.S. 425, 433 (1894) (not patent infringement to supply paper rolls to purchasers of patented paper-dispensing device, where paper roll is an unpatented element of patented combination); accord Aro Mfg. Co. v. Convertible Top Replacement Co., 365 U.S. 336, 128 U.S.P.Q. (BNA) 354 (1961) (Aro I) (the right of repair includes the right to replace worn elements of patented combination).

n45 The effect of a tie-in restriction is that a seller is able to dictate to consumers where they will buy related products, thus forcing the consumers to do something they would not do in a competitive market. Eastman Kodak Co. v. Image Technical Serv., Inc., 112 S. Ct. 2072, 2075 (1992); Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 12 (1984).

n46 The form of the restriction is different because in a tie-in restriction the sale of two products (e.g., printer and toner) is linked, whereas a single-use restriction merely restricts a product (e.g., toner cartridge) to a single use.

n47 Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 708, 24 U.S.P.Q.2d (BNA) 1173, 1179 (Fed. Cir. 1992).

n48 See Automatic Radio Mfg. Co. v. Hazeltine Research, Inc., 339 U.S. 827, 836, 85 U.S.P.Q. (BNA) 378, 381 (1950) (holding no patent misuse because “this limited license for ‘home’ use production contains neither an express nor implied agreement to refrain from production for ‘commercial’ or any other [non‘home’ use as part consideration for the license grant”). Advanced Computer Servs. of Mich., Inc. v. MAI Sys. Corp., 845 F. Supp. 356, 30 U.S.P.Q.2d (BNA) 1443 (E.D. Va. 1994), is in the same vein. In Advanced Computer, the court rejected a tie-in claim between the sale of computers and maintenance services. Id. at 369, 30 U.S.P.Q.2d (BNA) at 1453. The computer manufacturer did not license competitive maintenance service suppliers under its own copyrighted operating system software that the manufacturer supplied to its computer customers. Id. at 361, 30 U.S.P.Q.2d (BNA) at 1446-47. This software included the internal programs necessary to start a computer up when its power was turned on. Id. at 360, 30 U.S.P.Q.2d (BNA) at 1446. There was no explicit agreement conditioning the sale of the computers on the purchase of maintenance services, the court said, and hence there was no tying agreement. Id. at 369, 30 U.S.P.Q.2d (BNA) at 1453. However, the decision had the effect of prohibiting any third-party maintenance service provider from repairing the computer, since if the provider turned the computer on, he would be committing copyright infringement. To the same effect but less explicitly, see MAI Sys. Corp. v. Peak Computer, Inc., 991 F.2d 511, 26 U.S.P.Q.2d (BNA) 1458 (9th Cir. 1993), cert. dismissed, 114 S. Ct. 671 (1994) (holding that computer maintenance service company engaged in unlawful reproduction of operating system software when it loaded the copyrighted start-up software into a computer's random access memory by turning on the computer).

On the other hand, courts have at times found patent tie-ins even in the absence of an explicit tying agreement. See, e.g., Leitch Mfg. Co. v. Barber Co., 302 U.S. 458, 462-63 (1938) (Brandeis, J.). Leitch dealt with the sale of unpatented tar emulsion to road builders by the owner of the patent on a process for curing pavement with emulsion. Id. at 460-61. There were no explicit licenses and no explicit tying agreements, but the patentee sued its competitors for causing a patent infringement when they supplied the emulsion to road builders. Id. at 459-60. The Court held that there was an implicit tie-in sale and therefore patent misuse. Id. at 462-63. Probably, the Leitch case would be decided differently today, under 35 U.S.C. § 271(d)(5) (1988), as amended in 1988, because of absence of market power. See also United States v. Masonite Corp., 316 U.S. 265, 278 (1942) (in applying the exhaustion doctrine, “this Court has quite consistently refused to allow the form into which the parties chose to cast the transaction to govern”). In the author’s opinion, it is very doubtful that the Federal Circuit would defer to the dictum from Masonite.

n49 Aro Mfg. Co. v. Convertible Top Replacement Co., 365 U.S. 336, 128 U.S.P.Q. (BNA) 354 (1961) (Aro I); see also Aro Mfg. Co. v. Convertible Top Replacement Co., 377 U.S. 476, 141 U.S.P.Q. (BNA) 681 (1964) (Aro II).

n50 See Surgical Laser Techs., Inc. v. Surgical Laser Prods., Inc., 25 U.S.P.Q.2d (BNA) 1806 (E.D. Pa. 1992). Here, the patentee designed only one component of the laser unit to be disposable. Id. at 1807. The court held that the patentee’s intent was immaterial. Id. at 1808.

n51 Aro I, 365 U.S. at 346, 128 U.S.P.Q. (BNA) at 685.

n52 In Mallinckrodt, the recycler sought to assert the repair doctrine as a defense, but the court said that it was unnecessary to reach the issue of permissible repair versus impermissible reconstruction. Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 702, 709, 24 U.S.P.Q.2d (BNA) 1173, 1175, 1180 (Fed. Cir. 1992). According to the court, since the hospitals’ licenses to use the devices terminated after one use, it made no difference whether what the recycler did was repair or reconstruction. Id. at 709, 24 U.S.P.Q.2d (BNA) at 1180. In either case, any further use was in excess of the license that Mallinckrodt granted to its customers, and therefore was patent infringement, whether it was the use of a repaired or of a reconstructed unit. Id. at 709, 24 U.S.P.Q.2d (BNA) at 1180. This is comparable to the case of repair of an unlicensed machine, which is considered a patent infringement. Aro II, 377 U.S. at 484, 141 U.S.P.Q. (BNA) at 683.

n53 See Wilbur-Ellis Co. v. Kuther, 377 U.S. 422, 425, 141 U.S.P.Q. (BNA) 703, 705 (1964).

n54 Id., 141 U.S.P.Q. (BNA) at 705.

n55 This might create a conflict with the policy of § 117 of the Copyright Act, 17 U.S.C. § 117 (1988 & Supp. V 1993), which gives owners of copies of computer programs a right (or privilege) to adapt and modify the programs to improve their usefulness to the owner. See Foresight Resources Corp. v. Pfortmiller, 719 F. Supp. 1006, 1009, 13 U.S.P.Q.2d (BNA) 1721, 1723 (D. Kan. 1989) (purchaser of a copy of computer software may modify the software to improve its usefulness to the purchaser). The Federal Circuit might well conclude, however, that no “right” exists, only a very limited immunity from suit under the Copyright Act, and no immunity at all from suit under the Patent Act. See Interpart Corp. v. Italia, 777 F.2d 678, 228 U.S.P.Q. (BNA) 124 (Fed. Cir. 1985), overruled by Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 9 U.S.P.Q.2d (BNA) 1847 (1989). In Interpart, the Federal Circuit held that the federal patent laws do not create a right to copy and use prior-art and public-domain products, although § 102 of the patent act, 35 U.S.C. § 102 (1988), excluded such products from patent protection: “The patent laws ‘say nothing about the right to copy or the right to use, they speak only in terms of the right to exclude.’ Interpart, 777 F.2d at 685, 228 U.S.P.Q. (BNA) at 129. But in Bonito Boats, the Supreme Court expressly rejected this position: “The federal patent laws do create a federal right to ‘copy and to use’ [those products in the public domain].” Bonito Boats, 489 U.S. at 165, 9 U.S.P.Q.2d (BNA) at 1857-58. Presumably, therefore, the Federal Circuit would deny that there is a federal adaptation ‘right&rsquot; for computer programs, and the Supreme Court would insist that there is such a right. If the latter view prevails, § 117 would be considered an affirmative grant of federal rights and would presumably supersede earlier patent legislation and case law interpreting it. See Vault Corp. v. Quaid Software, Ltd., 847 F.2d 255, 270, 7 U.S.P.Q.2d (BNA) 1281, 1295 (5th Cir. 1988) (holding state law preempted because it interfered with rights under § 117).

n56 See generally Winn L. Rosch, Math Coprocessors: Fact or Fantasy, PC MAG., Feb. 12. 1991 at 301. A clock doubler chip is used to increase the internal clock speed of a microprocessor rated at low clock speed. Similarly, the motherboard, itself, may be designed to run clock speeds higher than that which is printed on the microprocessor chip. In essence, such activity makes irrelevant the clock speed printed on the microprocessor chip.

n57 See generally id. A given microprocessor chip, such as the 68030 or 80386, is more valuable and commands a greater price if it can operate at a speed of 33 MHz than if it can operate only at a speed of 25 MHz. Id. (1991 price of 8-MHz 80287 math co-processor was $ 326, 10-MHz, $ 374). A personal computer with a 33 MHz microprocessor processes data much faster than one with a 25 MHz microprocessor, and commands a higher price in the market. (By the same token, a 50 MHz microprocessor commands an even higher price.) Two 80386 microprocessor chips may come off the same production line, one after the other, and may be identical in all respects, but one may be labelled and sold as a 25 MHz chip and the other as a 33 MHz chip. It is cheaper for a manufacturer to operate this way than to have two separate 80386 chip designs created to satisfy both the 25 and 33 MHz markets.

A similar business phenomenon has led to class-action litigation against the contact lens manufacturer, Bausch and Lomb. Roberts v. Bausch & Lomb, No. CV-94-C-1144-W (N.D. Ala. Nov. 1, 1994) (order granting class certification). Bausch & Lomb manufactures the Medalist, Optima, and SeeQuence2 lines of contact lenses. Kenneth N. Gilpin, Suit Against Bausch & Lomb is Judged a Class Action, N.Y. TIMES, Nov. 3, 1994, at D5. The Optima is sold for annual replacement at a price of $ 60 to $ 70 per pair. Class Action Complaint at 14, Roberts. The Medalist, priced at $ 7.50 per pair, are a premium grade marketed to users for discard and replacement every one to three months. Id. The lower priced SeeQuence2 lenses, designed to be discarded every two weeks, sell for $ 6.65 per pair. Id. Class action plaintiffs contend that Bausch & Lomb has defrauded consumers by failing to inform them that the lenses were all identical. Id. at 1, 14-15. A Bausch & Lomb spokesman stated that the lenses use the “same design and same materials,” but they are packaged differently, and doctors prescribe them differently. Gilpin, supra, at D5.

n58 See John McCormick, Fast 386 Computers, 8 GOV'T COMPUTER NEWS 59 (Aug. 21, 1989) (pushing 20-MHz Intel 386 microprocessor chips to 25-MHz became common practice during a shortage of 25-MHz chips because of the shortage and high price of the faster chips).

n59 See id. (quoting Intel’s 80386 product manager Bill Rash warning against the common practice of pushing 20-MHz Intel 386 chips to 25-MHz because the chip “cannot function correctly at speeds it is not tested for”). Here Intel knew of the practice but apparently felt that it could do nothing but issue dire warnings.

n60 See Barry Siegel, Faith Lost, a Doctor Turns Bitter, L.A. TIMES, Sept. 12, 1993, at A1 (the drug levamisole sold for six cents per pill when sold as an animal dewormer and six dollars per pill when marketed as a human drug). A patented drug typically sells for a much higher price in the “human” market than in the “animal” market. See id. A drug may have different market value when used for different indications, and therefore may be able to command different prices in the various indication niches. For example, a mediocre anti-flatulence agent facing many competitors might command a much higher price as a baldness cure if it had no rivals in the latter niche.

n61 Herbicides and insecticides may have different values when used against varying targets or when sold to different types of end users, for example, commercial growers versus home gardeners. See, e.g., Chemagro Corp. v. Universal Chem. Co., 244 F. Supp. 486, 489, 146 U.S.P.Q. (BNA) 466, 467-68 (E.D. Tex. 1965), discussed infra at note 98.

For an example of niche marketing in a geographic context, see LaPeyre v. F.T.C., 366 F.2d 117, 119-20 & n.3, 151 U.S.P.Q. (BNA) 79, 80-81 & n.3 (5th Cir. 1966) (patentee leased shrimp-peeling machines to Pacific Northwest shrimp processors at twice the price charged Gulf Coast processors to reflect price of the hand labor being replaced, because peeling smaller Northwest shrimp required twice the hand labor as peeling Gulf shrimp).

n62 A customer who buys the product at a lower price, because that product is in a market that could support only the lower price, engages in arbitrage by reselling the product at a higher price to another customer, whom the manufacturer would have charged a high price because the other customer is in a market that can support the higher price.. See Raymond P. Niro et al., Nonstatutory Defenses in Patent Infringement Suits: Where Did They Come From? Where Are They Going?, in PATENT LITIGATION 1991 at 711, 735 (PLI Patents, Copyrights, Trademarks, and Literary Property Course Handbook Series No. G4-3873, 1991) (discussing price discrimination in licensing or sale and the possibility of arbitrage). See also Chemagro, 244 F. Supp. at 489, 146 U.S.P.Q. (BNA) at 468, discussed infra at note 98.

n63 This restriction is similar to one found violative of the antitrust laws when used in the drug industry. See, e.g., United States v. Glaxo Group Ltd., 410 U.S. 52, 176 U.S.P.Q. (BNA) 289 (1973). In the Glaxo case the patentees used a “no resale as such” restriction to keep a drug chemical out of the hands of price cutting manufacturers (generic drug sellers). Id. at 55, 63, 176 U.S.P.Q. (BNA) at 290, 293. The perhaps important difference between the hypothetical case in the text and the Glaxo case is that the former involves several differentiated markets in which the patented technology or technology-embodying product has different values in the various niches, while the latter case is a single market in which prices are elevated by keeping the product out of the hands of price cutters. Whether the difference should make a legal distinction is difficult to resolve.

n64 See Unidisco, Inc. v. Schattner, 824 F.2d 965, 3 U.S.P.Q.2d (BNA) 1439 (Fed. Cir. 1987). The patentee, Schattner, granted a license to Girard to make, use, and sell without any right to assign or sublicense. Id. at 966, 3 U.S.P.Q.2d (BNA) at 1440. Girard then sold exclusively to Unidisco, who resold to the dental trade in a manner to which the patentee objected. Id., 3 U.S.P.Q.2d (BNA) at 1440. The trial court entered judgment for the patentee on a jury verdict of willful patent infringement. Id. at 967, 3 U.S.P.Q.2d (BNA) at 1441. The Federal Circuit reversed, holding that the exclusive distribution arrangement was not a sublicense, and that Unidisco had authority to resell the purchased products without liability for patent infringement, since it purchased them from a seller that the patentee had authorized. Id. at 968, 3 U.S.P.Q.2d (BNA) at 1441.

n65 Id. at 966, 3 U.S.P.Q.2d (BNA) 1440. The Mallinckrodt decision does not discuss Unidisco.

n66 See Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 708-10, 24 U.S.P.Q.2d (BNA) 1173, 1179-81 (Fed. Cir. 1992).

n67 Id. at 708, 24 U.S.P.Q.2d (BNA) at 1179-80. The Mallinckrodt court distinguished prior Supreme Court decisions under the exhaustion doctrine as being limited to situations in which patentees sought to impose price-fixing or tie-in restrictions on their customers. Id., 24 U.S.P.Q.2d (BNA) at 1179-80; see also supra note 33, and accompanying text.

n68 See United States v. General Elec. Co., 272 U.S. 476, 489-90 (1926); see also supra text accompanying notes 5-7.

n69 Windsurfing Int’l, Inc. v. AMF, Inc., 782 F.2d 995, 1001-02, 228 U.S.P.Q. (BNA) 562, 566-67 (Fed. Cir.), cert. denied, 477 U.S. 905 (1986).

n70 See id. at 1001 & n.9, 228 U.S.P.Q. (BNA) at 566-67 & n.9 (applying antitrust rule of reason to misuse).

n71 In the Mallinckrodt case, the hospitals sent the units offsite for recycling. Mallinckrodt, 976 F.2d at 702, 24 U.S.P.Q.2d (BNA) at 1174. It would appear that if it were impractical for the hospital, itself, to recycle the product, then manufacturing the product would have been equally impractical.

n72 See General Elec., 272 U.S. 476.

n73 See generally Jay P. Friedenson, Reasons to License and Not to License Technology and Patents, in TECHNOLOGY LICENSING AND LITIGATION 1991 at 101 (PLI Patents, Copyrights, Trademarks, and Literary Property Course Handbook Series No. G4-3860, 1991) (listing factors to be considered when deciding whether or not to license technology and patents).

n74 In United States v. Ciba-Geigy Corp., 508 F. Supp. 1118 (D.N.J. 1976), the patentee Ciba-Geigy licensed another drug company, Abbott, to manufacture a patented drug chemical, subject to various licensing limitations. Id. at 1129. But Ciba-Geigy sold the chemical to several other drug companies, subject to restrictions that were functionally similar to the limitations in Abbott's license. Id. at 1132-35 (both licensees and bulk buyers were prohibited from selling the product in bulk). Clearly, it was more economical for the customer companies to purchase the drug chemical from Ciba-Geigy than enter into its manufacture, at least in the first instance. See, e.g., id. at 1129 (Abbott wanted to get into the market as soon as possible without waiting to perfect manufacture of the product). Apparently, only Abbott found it economical to enter into the manufacture of the drug chemical for use in its products. See id. at 1126. Economies of scale would appear most plausibly to account for the different companies' respective approaches. See id. at 1155 (showing chart of market share). See also Robert P. Taylor, Licensing in Theory and Practice: Licensor-Licensee Relationships, 53 ANTITRUST L.J. 561, 601 & n.112 (1985) (showing hypothetical in which licensor's economies of scale would be lost by authorizing manufacture by licensees).

n75 The risk of loss of trade secrets may have been the explanation for the use of restrictive sales arrangements, rather than merely using limitations in manufacturing licenses, in United States v. Glaxo Group Ltd., 410 U.S. 52, 176 U.S.P.Q. (BNA) 289 (1973). Limited potential sales volume and consequent inability of the licensee/customer companies to reach sufficient volume for economies of scale may also have been important. The product (an antibiotic) was made by a fermentation process, which may have involved scale economies and trade secrets. Id. at 54, 176 U.S.P.Q. (BNA) at 289.

n76 Patentees could, for example, patent both the process and the apparatus of the invention. See generally In re Tarczy-Hornoch, 397 F.2d 856, 857, 158 U.S.P.Q. (BNA) 141, 142 (C.C.P.A. 1968) (holding that a process claim will not be rejected because the application discloses the apparatus that will inherently execute the process).

n77 See, e.g., Advanced Computer Servs. of Mich., Inc. v. MAI Sys. Corp., 845 F. Supp. 356, 30 U.S.P.Q.2d (BNA) 1443 (E.D. Va. 1994) (software developer licensed, but did not sell, its software, and tie-in was not explicit); see also text accompanying supra notes 3-12, 48.

n78 “An authorized sale of a patented product exhausts the patent monopoly as to that product. Thus, a purchaser of such product from a patent owner may use or resell the product free of control or conditions imposed by the patent owner.” Cyrix Corp. v. Intel Corp., 803 F. Supp. 1200, 1214 (E.D. Tex 1992), appeal dismissed, 9 F.3d 978 (Fed. Cir. 1993).

n79 Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 702 n.2, 24 U.S.P.Q.2d (BNA) 1173, 1174 n.2 (Fed. Cir. 1992).

A hypothetical such overall combination claim (a rewritten, simplified form of Mallinckrodt’s combination claim) could be in the following general form:

An apparatus for dispensing and collecting radioactive mist for diagnosis of pulmonary disease in a patient, said apparatus comprising:

    a container for holding a predetermined quantity of radioactive diagnostic fluid, said container having an output port;

    connected to said output port, a nebulizer for generating mist from said fluid and for dispensing said mist via a first tube, said first tube containing a one-way valve permitting airflow only away from said nebulizer;

    connected to said first tube, a breathing mask for covering a patient's nose and mouth, said mask having an exit tube containing:

    a one-way output valve permitting airflow only away from said mask; and a filter for trapping said mist and removing said mist from air flowing out of said exit tube; and

    a sealable, radiation shield for supporting and enclosing said container, nebulizer, and tubes, said shield having an “open” mode permitting access to said elements and a “closed” mode for reducing radiation transmitted to surroundings.

n80 See id. at 702, 24 U.S.P.Q.2d (BNA) at 1174.

n81 Id. at 703, 24 U.S.P.Q.2d (BNA) at 1175.

n82 See In re Tarczy-Hornoch, 397 F.2d 856, 158 U.S.P.Q. (BNA) 141 (C.C.P.A. 1968) (permitting method claims to be drawn to inherent functioning of apparatus).

n83 One such claim, to “a method for dispensing and collecting radioactive mist for diagnosis of pulmonary disease in a patient,” might be written in terms of “providing” each of the component elements of the apparatus (the container, the nebulizer, the mask, the filter, etc.) of the claim, supra note 79. Another form of method claim might recite the inherent functioning of the same apparatus (placing fluid in a container, generating a mist, trapping mist in a filter, etc.).

n84 See In re Yarn Processing Patent Validity Litig., 541 F.2d 1127, 192 U.S.P.Q. (BNA) 241 (5th Cir. 1976), cert. denied, 443 U.S. 910 (1977); Duplan Corp. v. Deering Milliken, Inc., 444 F. Supp. 648, 197 U.S.P.Q. (BNA) 342 (D.S.C. 1977), aff'd, 594 F.2d 979, 201 U.S.P.Q. (BNA) 641 (4th Cir. 1979), cert. denied, 444 U.S. 1015 (1980). The patentee licensed the machine patent to one set of licensees and the patents on the yarn or method of making it to a second set of licensees. Yarn Processing, 541 F.2d at 1131, 192 U.S.P.Q. (BNA) at 244-45. The first set of licensees made and sold the machines to the second set. Id., 192 U.S.P.Q. (BNA) at 245. In principle, it would appear to be of no significance whether the patentee itself made the machines instead of having the first set of licensees do so.

n85 In Yarn Processing and Duplan, the courts also upheld limitations or restrictions in the patentees’ apparatus-patent licenses to machine manufacturers, which required them to sell machines only to customers who had a method-patent license (or one to make the end product that the machines manufactured). Yarn Processing, 541 F.2d at 1135, 192 U.S.P.Q. (BNA) at 247; Duplan, 444 F. Supp. at 671-72, 197 U.S.P.Q. (BNA) at 364-65.

n86 The purpose of the upheld restrictions was to insure that yarn manufacturers paid a royalty based on amount of yarn production. There were other customer restrictions that the courts found illegal, because their purpose was not benign (e.g., price-fixing, tie-ins). Yarn Processing, 541 F.2d at 1135-37, 192 U.S.P.Q. (BNA) at 248-49.

n87 See also Cold Metal Process Co. v. McLouth Steel Corp., 41 F. Supp. 487 (E.D. Mich. 1941), aff'd on other grounds, 126 F.2d 185 (6th Cir. 1942). In that case the court upheld a license on a patented rolling mill, where the patentee sold the mill to the customer and also charged a running royalty based on output. Id. at 488. The court distinguished Motion Picture Patents Co. v. Universal Film Mfg. Co., 243 U.S. 502 (1917), on the ground that the exhaustion doctrine did not excuse the customer from paying a running royalty on use where that is merely a method of making the patentee’s compensation proportional to the value of the benefit conferred. Cold Metal, 41 F. Supp. at 490.

n88 See supra notes 84-87.

n89 See supra note 48.

n90 In the hypothetical case as stated, there is a sale of a product on which there is another patent, a product patent. Presumably, that patent is irrelevant. But see Ethyl Gasoline Corp. v. United States, 309 U.S. 436 (1940). In that case, Ethyl had a number of different method and product patents. Id. at 446. The Supreme Court did not accept Ethyl’s argument that the exhaustion doctrine did not apply because the restriction could be upheld as a license limitation when viewed as supported by certain of the patents whose subject matter the licensees manufactured. Id. at 457. However, the facts of Ethyl involved a blatant cartel for regulating prices. Id. at 458.

n91 See supra note 11 and accompanying text.

n92 See Windsurfing Int’l, Inc. v. AMF, Inc., 782 F.2d 995, 1001-02 & n.9, 228 U.S.P.Q. (BNA) 562, 566-67 & n.9 (Fed. Cir.) (applying antitrust rule of reason to misuse, citing Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36 (1977)), cert. denied, 477 U.S. 905 (1986).

n93 Id. at 1001-02, 228 U.S.P.Q. (BNA) 566-67.

n94 Id. at 1001, 228 U.S.P.Q. (BNA) 566-67.

n95 The defense of license or estoppel might be urged. Ordinarily, a patentee’s sale of a machine or other product having no substantial use but in the practice of an invention on which the patentee-seller owns a patent conveys an implied license to practice the invention or estops the patentee from suing the customer for infringement. Met-Coil Sys. Corp. v. Korners Unlimited, Inc., 803 F.2d 684, 687, 231 U.S.P.Q. (BNA) 474, 476-77 (Fed. Cir. 1986). Given an explicit notice from the patentee that a two-tier license system is in effect, however, a court would probably follow the decisions described supra in notes 84 & 87, and find no estoppel resulting from the sale.

n96 See supra note 83. There might, but need not, be additional claims dependent on the basic method claim, in the form: “The method of claim 1 wherein said signal is processed at a clocked rate of no less than 33 MHz.”

n97 This would be analogous to the licensing of big (commercial) and little (noncommercial) amplifiers separately in General Talking Pictures Corp. v. Western Elec. Co., 304 U.S. 175, aff'd on reh’g, 305 U.S. 124 (1938). The General Talking Pictures case involved licenses to make and sell a patented product for use in different fields of end use. Id. at 176. Here it is proposed to grant licenses to use a patented method in different fields. Courts would probably consider the distinction immaterial. The opinions in the General Talking Pictures and General Electric cases, assert that it is legitimate to license less than the entire scope of one's patent rights. General Talking Pictures, 304 U.S. at 181; United States v. General Elec. Co., 272 U.S. 476, 489-90 (1926). Use is all that one can license under a method patent. Id. at 489. Hence, there is no way to license less than all of one's rights under a method patent other than by in some way limiting the licensed use. Id. The same criteria for determining reasonableness of the limitations in the case of a product patent should apply to a method patent, for there appears to be no difference in the applicable principles.

If the method patent had separate claims on different speeds, different licensees would be licensed under different claims. See supra note 96.

n98 The same principle applies to herbicides and pharmaceuticals. See Chemagro Corp. v. Universal Chem. Co., 244 F. Supp. 486, 146 U.S.P.Q. (BNA) 466 (E.D. Tex. 1965). In addition to method patents, one can often get a product patent on a particular composition (formulation for using the active chemical). If different uses require different compositions, selling the chemical in two partially formulated forms may be a very effective way to prevent cross-selling between separately licensed niches. That reduces the “policing” problem inherent in such licensing schemes. For example, consider a pharmaceutical used in counteracting anaemia in humans and piglets, where the price in the first market can be a large multiple of the price in the second market. Suppose that the manufacturer adds phenol (carbolic acid) to the bulk chemical destined for the low-price “piglet” market; that will prevent purchasers from diverting the bulk product to the highprice “human” market. See Fisons Ltd. v. United States, 458 F.2d 1241, 172 U.S.P.Q. (BNA) 334 (7th Cir. 1972), cert. denied, 405 U.S. 1041, 173 U.S.P.Q. (BNA) 257 (1972) (decision on jurisdiction). (Fisons, however, not only added phenol to bulk products intended for piglet use, but it imposed customer restrictions as well. Id. at 1243, 172 U.S.P.Q. (BNA) at 335.)

The difficulties of keeping market niches separate when formulations for different niches are not incompatible are illustrated by the facts of Chemagro. Chemagro sold its patented insecticide in a 10% strength for use by commercial growers, and in a 2% , higher-priced formulation for home-garden use. Chemagro, 244 F. Supp. at 489, 146 U.S.P.Q. (BNA) at 468. The 10% product had a label notice prohibiting reformulation and stating that the product was sold for commercial use only. Id., 146 U.S.P.Q. (BNA) at 468. Universal purchased the lowpriced 10% formulation, diluted it fivefold, and sold the resulting product for home-garden use. Id., 146 U.S.P.Q. (BNA) at 468. Chemagro sued for patent infringement and prevailed. Id. at 490, 146 U.S.P.Q. (BNA) at 469. The Federal Government then sued Chemagro and its corporate parent for antitrust violation and required them to accept a decree prohibiting them from engaging in the restriction. United States v. Farbenfabriken Bayer AG, 1969 Trade Cas. (CCH) 1 72,918 (D.D.C.) (consent decree). Apparently, the restriction, supported by patent infringement sanctions, was the most effective means for Chemagro to prevent cross-selling between markets and thus maximize its revenue, but the antitrust decree prevented use of that expedient. Under Mallinckrodt, however, a patentee in Chemagro’s position would now be placed back in the position in which the court in the initial Chemagro case left Chemagro and from which the antitrust decree ejected it. The alternative transaction structure suggested in text would, if feasible, accomplish much the same result by substituting a use-license limitation for the overt customer restriction. So, too, would mixing the “downmarket” commercial grower product with an ingredient making it unsuitable for “upmarket” home-garden use, e.g., something with an unpleasant smell.

n99 The Telephone Cases, 126 U.S. 1 (1888), are illustrative. The patentee, Alexander Graham Bell, could not be certain of the ramifications of his inventions. “He astonished the scientific world.” id. at 539. Bell mistakenly thought that the magneto method of practicing his invention was superior to the variable resistance method, and directed most of his patent claims to it. Bell guessed wrongly, but luckily had included a method claim covering both the magneto and variable resistance methods. Id. at 536-37. The Court upheld this claim, ruling that it would not confine the patent of the magneto method at the expense of the variable resistance method. Id. at 537. Not every inventor is as lucky as Bell.

n100 Mallinckrodt merely placed a label on the package and sold it with the restrictive label “For Single Patient Use Only” and an insert with a similar statement. Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 702, 24 U.S.P.Q.2d (BNA) 1173, 1174 (Fed. Cir. 1992). Could this lead to a situation analogous to the celebrated “battle of the forms?” See U.C.C. 2-207 cmt. 1 (1966).

n101 U.S. CONST. art. 1, 8, cl. 8; United States v. Masonite Corp., 316 U.S. 265, 278 (1942); Pennock v. Dialogue, 27 U.S. (2 Pet.) 1, 19 (1829); see also Feist Publications, Inc. v. Rural Tel. Serv. Co., 499 U.S. 340, 349, 18 U.S.P.Q.2d (BNA) 1275, 1279 (1991) (primary objective of copyright law is to advance progress of science and useful arts rather than to reward labor of authors).

n102 Sony Corp. of Am. v. Universal City Studios, Inc., 464 U.S. 417, 429, 220 U.S.P.Q. (BNA) 665, 672-73 (1984) (fundamental purpose of copyright and patent statutes is to serve public good, rather than to further private gain of individuals).

n103 See generally Herbert Hovenkamp, Positivism in Law and Economics, 78 CAL. L. REV. 815 (1990) (supporting Chicago School view of economics, and asserting that an increase in equity will yield an increase in welfare). Hovenkamp terms this “efficiency” and asserts that “a move is efficient if those who gain from the move are able to compensate the losers fully for their losses but still remain better off.” id. at 833. Query: does the gainers’ ability to compensate the losers imply that the gainers in fact do so?

n104 Eleanor M. Fox & Lawrence A. Sullivan, Antitrust Retrospective and Prospective: Where Are We Coming From? Where Are We Going?, 62 N.Y.U. L. REV. 936, 959 (1987).

n105 See id. at 961-63 (exploring the Critical Legal Studies (CLS) movement’s criticism of the Chicago School and applying the CLS view of law formation as a function of class interests — i.e., who wins and who loses — to antitrust). Fox and Sullivan conclude that “disregard of fairness . . . is the stuff of which disaffections are made.” id. at 964. The implication is that such disaffection invites social and political upheaval. Id.

n106 See supra text accompanying notes 76-100.

n107 Economies of scale have led scholars to declare that “the single largest factor tending to undermine perfect competition is the presence of significant economies of scale.” HERBERT HOVENKAMP, ECONOMICS AND FEDERAL ANTITRUST LAW 1.4, at 25 (1985). “Economies of scale” exist whenever the costs per unit of some input decrease as volume increases. Id. at 25.

n108 This is, of course, a simplified reduction of the economic theory and there are many more factors that will go into determining and analyzing “economics of scale.” See id. at 26-27. It is often very difficult to determine the most efficient minimum size of a single plant, for example, and conclude that a firm that operates such a plant has achieved all available economics of scale. Id.

n109 Can Semiconductors Survive Big Business, BUS. WK., Dec. 3, 1979, at 66 (unit cost reduction of 25%-30% with each doubling of production).

n110 See supra text accompanying note 57.

n111 The Federal Circuit made a similar point in the Mallinckrodt opinion as an argument for not following prior decisions of the Supreme Court. Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 705, 24 U.S.P.Q.2d (BNA) 1173, 1177 (Fed. Cir. 1992). In that context, the argument is not persuasive. Ordinarily, the wisdom of Supreme Court decisions does not enter into whether or not an inferior federal court should follow them as precedents or, in effect, overrule them.

n112 See supra text accompanying notes 76-100.

n113 See Straus v. Victor Talking Mach. Co., 243 U.S. 490, 501 (1917) (“restraints . . . such as have been hateful to the law from Lord Coke’s day to ours, because [they are] obnoxious to the public interest”); 2 COKE ON LITTLETON 360 (post-sale restrictions on goods are “against trade and traffique, and bargaining and contracting betweene man and man”); see also Continental T.V., Inc. v. GTE Sylvania, Inc., 433 U.S. 36, 66-69 (1977) (White, J., concurring).

n114 This is the same paradigm as that of the automobile manufacturer trampling on the rights of the individual car owner in British Leyland Motor Corp. v. Armstrong Patents Co., [1986] 1 All E.R. 850 (H.L.) (holding that copyright law must not be utilized in derogation of car owner”s property right in car purchased from copyright owner; consequently, copyright owning car manufacturer cannot require that customers buy copyrighted replacement parts only from manufacturer's licensees).

n115 Mallinckrodt, 976 F.2d at 708, 24 U.S.P.Q.2d (BNA) at 1179.

n116 Another formulation that the Federal Circuit used was that, if the restriction “was within the scope of the patent grant or otherwise justified, then violation of the restriction may be remedied by action for patent infringement.” id. at 709, 24 U.S.P.Q.2d (BNA) at 1180 (emphasis added). The use of the word “or” here indicates that a restriction will be enforced if either criterion is satisfied.

n117 Id. at 704, 24 U.S.P.Q.2d (BNA) at 1176.

n118 The 1926 General Electric case, United States v. General Elec. Co., 272 U.S. 476 (1926), has never been overruled. See United States v. Huck Mfg. Co., 382 U.S. 197, 147 U.S.P.Q. (BNA) 404 (1965) (4-4 decision); United States v. Line Material Co., 333 U.S. 287, 305, 76 U.S.P.Q. (BNA) 399 (1948) (4-4 decision). General Electric allowed price-fixing in a patentee’s manufacturing license and prohibited it when a post-sale restriction. General Elec., 272 U.S. at 488, 494. It is unclear what the Federal Circuit would make of that distinction. Presumably, the Federal Circuit would not hold price-fixing illegal in circumstances in which General Electric held it legal, so that some price-fixing is not illegal per se.

Since a 1988 amendment to the patent statute, tie-ins are unlawful only if the patentee has “market power” in the relevant market for the tying item. 35 U.S.C. § 271(d)(5) (1988). The Federal Circuit would hardly place patentees in a worse position than the patent statute places them.

Few post-sale restrictions, therefore, would be unlawful under the Federal Circuit’s test.

n119 The latter would be lawful under federal patent law. 35 U.S.C. § 271(d)(4) (1988).

n120 In addition, a sensible rule would not accord disparate legal treatment to functionally equivalent, commercially similar practices. This would mean giving comparable legal treatment to equivalent restrictions and limitations. Adopting that approach would, of course, require courts not to let the expedients discussed in section III.B of this Article mesmerize them. Indeed, the system should have built-in safeguards against that.

n121 See National Soc’y of Professional Eng'rs v. United States, 435 U.S. 679, 696 n.22 (1978).

n122 See supra text accompanying notes 5-7.

n123 See supra text accompanying notes 116-119; see also supra note 11 and accompanying text.

n124 See supra text accompanying notes 5-7.

n125 See Robert W. Kastenmeier & Michael J. Remington, The Semiconductor Chip Protection Act of 1984: A Swamp or Firm Ground?, 70 MINN. L. REV. 417, 422-23, 440-42 (1985) (describing appropriate test to determine whether to recognize additional intellectual property rights).

n126 Mitchel v. Reynolds, 1 P. Wms. 181, 24 Eng. Rep. 347 (1711).

n127 1 P. Wms. 181, 191-92, 24 Eng. Rep. 347, 351 (1711) (“All contracts where there is a bare restraint of trade and no more, must be void; but there can be no reason why, in cases where the special matter appears so as to make it a reasonable and useful contract, it should not be good . . . .“); accord United States v. Addyston Pipe & Steel Co., 85 F. 271, 281 (6th Cir. 1898), aff’d, 175 U.S. 211 (1899).

n128 Such restraints are those restrictions that are ancillary to a contract for a lawful undertaking and are necessary to bring about the lawful main purposes of the contract. See Addyston Pipe, 85 F. at 282 (A restriction must be “merely ancillary to the main purpose of a lawful contract, and necessary to protect the covenantee in the enjoyment of the legitimate fruits of the contract, or to protect him from the dangers of an unjust use of those fruits by the other party. . . . If the restraint exceeds the necessity presented by the main purpose of the contract, it is unlawful . . . .”); see also Interface Group, Inc. v. Massachusetts Port Auth., 816 F.2d 9, 10 (1st Cir. 1987) (Breyer, J.) (“ 'Unreasonableness' in antitrust law has a rather special meaning . . . that the anticompetitive consequences of a particular action or arrangement outweigh its legitimate business purposes.”).

n129 See Aro Mfg. Co. v. Convertible Top Replacement Co., 365 U.S. 336, 339-40, 128 U.S.P.Q. (BNA) 354, 356-57 (1961) (Aro I) (claims are measure of grant).

n130 In Copperweld Corp. v. Independence Tube Corp., 467 U.S. 752, 777 (1984), the Supreme Court held that a corporation could not “conspire” with itself (in that case, a subsidiary) in violation of the antitrust laws, because two separate entities were not involved. Presumably, the same notion would apply to a principal and agent. See United States v. General Elec. Co., 272 U.S. 476, 484-85 (1926), in which the Supreme Court held that the defendant did not violate the antitrust laws by fixing the prices at which its agents would resell patented light bulbs, because the principal and agent were one, so that the transfer of the light bulbs from the defendant to its agents was not a sale subject to a restriction. Id. at 484-85. The Supreme Court has at other times refused to treat similar transactions between a seller and its so-called agent as beyond the reach of the antitrust laws. See, e.g., Simpson v. Union Oil Co., 377 U.S. 13, 24 (1964); United States v. Masonite Corp., 316 U.S. 265, 280 (1942).

The present proposal does not attempt to sort out this issue, although the author considers the opinions in the later cases to state the better view. Perhaps, courts must decide on the basis of each case's individual facts whether a sham or genuine agency relationship is involved, but those terms are not self-defining.

n131 Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 703-09, 24 U.S.P.Q.2d (BNA) 1173, 1175-80 (Fed. Cir. 1992).

n132 The principal defenses to a claim of patent infringement are listed in 35 U.S.C. § 282 (1988 & Supp. IV 1992). See generally 5 DONALD S. CHISUM, PATENTS § 19.01, at 19-4 to 19-5 & n.1 (1994) (discussing four major defenses — invalidity, fraudulent procurement, misuse or violation of anti-trust laws, and laches or estoppel — as well as such defenses as implied license or experimental purpose).

n133 See Constant v. Advanced Micro-Devices, Inc., 848 F.2d 1560, 1563-64, 7 U.S.P.Q.2d (BNA) 1057, 1058-59 (Fed. Cir.), cert. denied, 488 U.S. 892 (1988).

n134 See Spindelfabrik Suessen-Schurr v. Schubert & Salzer Maschinenfabrik AG, 829 F.2d 1075, 1080-82, 4 U.S.P.Q.2d (BNA) 1044, 1048-50 (Fed. Cir. 1987) (discussing legal estoppel as a defense to patent infringement), cert. denied, 484 U.S. 1063 (1988).

n135 See Met-Coil Sys. Corp. v. Korners Unlimited, Inc., 803 F.2d 684, 687, 231 U.S.P.Q. (BNA) 474, 476 (Fed. Cir. 1986).

n136 See, e.g., In re Yarn Processing Patent Validity Litig., 541 F.2d 1127, 192 U.S.P.Q. (BNA) 241 (original opinion), 194 U.S.P.Q. (BNA) 121 (corrections to opinion) (5th Cir. 1976), cert. denied, 443 U.S. 910 (1977); see also Duplan Corp. v. Deering Milliken, Inc., 444 F. Supp. 648, 197 U.S.P.Q. (BNA) 342 (D.S.C. 1977), aff’d, 594 F.2d 979, 201 U.S.P.Q. (BNA) 641 (4th Cir. 1979), cert. denied, 444 U.S. 1015 (1980). See supra notes 84-86 and accompanying text for a discussion of these cases.

n137 Bonito Boats, Inc. v. Thunder Craft Boats, Inc., 489 U.S. 141, 156-57, 9 U.S.P.Q.2d (BNA) 1847, 1854 (1989).

n138 Such a sanction would be an indirect enforcement of a limitation, condition, or restriction that Federal law has said should not be enforced.

n139 Immunization of licensees and customers, as well as patentees, may be desirable to effectuate the proposed compromise. There is authority that acquiescence in unlawful agreements may suffice to establish a combination and conspiracy in violation of the antitrust laws. See, e.g., United States v. Bausch & Lomb Optical Co., 321 U.S. 707 (1944). In Bausch & Lomb, the Supreme Court rejected the defendant’s contention that it acted unilaterally in imposing price and customer license restrictions, so that no conspiracy or combination in restraint of trade existed. Id. at 719-23. The Court said, “Whether this conspiracy and combination was achieved by agreement or by acquiescence of the wholesalers coupled with assistance in effectuating its purpose is immaterial.” id. at 723; see also United States v. Paramount Pictures, Inc., 334 U.S. 131, 161, 77 U.S.P.Q. (BNA) 243, 245 (1948) (holding that it is not a defense that large customers “fathered the illegal practices and forced them onto the defendants” because “acquiescence in an illegal scheme is as much a violation of the Sherman Act as the creation and promotion of one”).

It would be inconsistent with the logic of the proposed compromise to immunize patentees and yet hold their licensees and customers liable for acquiescing in the patentee’s immunized conduct.

n140 433 U.S. 36 (1977).

n141 The decision in United States v. Glaxo Group Ltd., 410 U.S. 52, 176 U.S.P.Q. (BNA) 289 (1973), is not to the contrary. In that case, the antitrust violation was a restriction on the sale of the patented drug product that was designed to keep the product out of the hands of price-cutting, generic-drug manufacturers, and thus was a device to fix or tamper with price levels for the drug. Id. at 62-63, 176 U.S.P.Q. (BNA) at 62. The facts are thus comparable, albeit on a lesser scale, with industry-regimenting licensing schemes such as that involved in Ethyl Gasoline Corp. v. United States, 309 U.S. 436, 449-50 (1940).

n142 See supra text following note 125.

n143 The term “on sale” is a term of art in patent law, meaning roughly that a mature, tangible version of the product has been offered for sale. 35 U.S.C. § 102(b) (1988); see, e.g., UMC Elec. Co. v. United States, 816 F.2d 647, 652, 2 U.S.P.Q.2d (BNA) 1465, 1468 (Fed. Cir. 1987), cert. denied, 484 U.S. 1025 (1988).

n144 85 F. 271 (6th Cir. 1898), aff'd, 175 U.S. 211 (1899); see supra note 128 (quoting the passage in the decision corresponding to the language in text).

n145 U.S. Dep't of Justice, Antitrust Guidelines for the Licensing and Acquisition of Intell. Prop. § 4.3.2, example 7 (draft Aug. 8, 1994), reprinted in Trade Reg. Rep. (CCH) ¶ 50,141 at 49,071-73.

n146 See supra notes 84-86 and accompanying text.

n147 See text accompanying supra notes 122-24.

n148 United States v. Addyston Pipe Co., 85 F. 271 (6th Cir. 1898), aff'd, 175 U.S. 211 (1899); see also supra note 127.

n149 See, e.g., Summit Health, Ltd. v. Pinhas, 500 U.S. 322, 330 (1991) (“proper analysis focuses, not upon actual consequences, but rather upon the potential harm that would ensue”); McLain v. Real Estate Bd. of New Orleans, Inc., 444 U.S. 232, 243 (1980) (collecting cases); United States v. Columbia Steel Co., 334 U.S. 495, 522 (1948).

n150 See, e.g., Jefferson Parish Hosp. Dist. No. 2 v. Hyde, 466 U.S. 2, 26-29 (1984) (tie-in where market power exists); Hartford-Empire Co. v. United States, 323 U.S. 386, 406-07 (1945) (product market allocation scheme); Ethyl Gasoline Corp. v. United States, 309 U.S. 436, 457-59 (1940) (cartel or market regimentation scheme); United States v. Krasnov, 143 F. Supp. 184, 202-03, 110 U.S.P.Q. (BNA) 411, 423-24 (E.D. Pa. 1956), aff'd per curiam, 355 U.S. 5 (1957) (groupboycott, veto-power license).

n151 Mallinckrodt, Inc. v. Medipart, Inc., 976 F.2d 700, 708, 24 U.S.P.Q.2d (BNA) 1173, 1179 (Fed. Cir. 1992).

n152 Id. at 709, 24 U.S.P.Q.2d (BNA) at 1180.

n153 Id. at 704-05, 24 U.S.P.Q.2d (BNA) at 1176.

n154 Id. at 708, 24 U.S.P.Q.2d (BNA) at 1179; see supra note 151.

n155 Mallinckrodt, 976 F.2d at 709, 24 U.S.P.Q.2d (BNA) at 1180; see supra note 152.

n156 See supra text following note 119.

n157 It should be accepted that, considered over a wide range of instances, making it easier to market or commercialize a product is equivalent to making it possible to market some products (those near the margin) that otherwise would not have been marketed at all.

n158 See supra text accompanying notes 74-75.