Revised Paper on Overbreadth In Canadian Patent Law

I have just posted a revised version of my draft paper on Overbreadth In Canadian Patent Law. In the earlier version I had somehow overlooked Apotex v AstraZeneca 2017 FCA 9, a key FCA decision on the “roads to Brighton” problem. That section of the paper is re-written, but my overall conclusions are not changed; on the contrary, Apotex v AstraZeneca strengthens my view that overbreadth is not required as an independent ground of invalidity

Claims May Be Construed with Reference to the Entire Specification

Tetra Tech EBA Inc v Georgetown Rail Equipment Co 2019 FCA 203 Dawson JA: Pelletier, Webb JJA rev’g Georgetown Rail Equipment Co v Rail Radar Inc 2018 FC 70 Fothergill J
             2,572,082 / 2,766,249

In Georgetown Rail v Rail Radar (blogged here), Fothergill J held that Georgetown’s 082 and 249 patents were valid and infringed by Tetra Tech EBA Inc, dismissing the main validity attack, that was based on obviousness. The FCA has now reversed.

The appeal relating to the 249 patent turned on claim construction. The FCA reversed Forthergill J on claim construction, presumably applying a correctness standard,given that claim construction is a matter of law [89]. In NOVA v Dow 2016 FCA 216 [15] the FCA suggested that in some circumstances, particularly where claim constructions turns on the trial judge’s assessment of expert evidence concerning the meaning of terms used in the claims, trial judges might be given “some leeway” in claim construction. Dawson JA made no reference to this quasi-deferential standard, but her decision is not inconsistent with it. In this case, the disputed construction did not turn on an appreciation of expert evidence, but rather on the interplay between the disclosure and claims; that is, a quintessentially judicial function of interpreting a written document. The obviousness of the claims alleged to have been infringed followed straightforwardly from the new claim construction [125]-[130].

A controversial point related to claim construction is whether having recourse to the disclosure as an aid to construing the claims is always permissible, or is permissible only if the claims are ambiguous when read on their own: see here and here. Dawson JA did not directly address the controversy, but she began her analysis by remarking (my emphasis):

[104] I begin the construction exercise with the Patent’s disclosure, mindful that while claims may be construed with reference to the entire specification, the disclosure may not be used to enlarge or contract the scope of a claim as written.

This passing remark strikes me as a succinct statement of the correct approach. It is always permissible to construe claims with reference to the disclosure, but it is never permissible to use the disclosure to change the scope of the claims, once the claims have been construed in light of the disclosure. (That is, absent any invocation of essential elements or Actavis-style expanded protection.) The line between construing the claim in light of the disclosure and importing limitations from the disclosure into the claims may sometimes be a bit blurry, but the law is full of lines that are sometimes blurry. This particular distinction strikes me as conceptually sound and easy enough to apply in most cases.

The appeal relating to the 082 patent turned on obviousness, with the key point being that the common general knowledge must be taken into account in assessing obviousness [13]. This, of course, does not represent new law.

The patents related to an automated system and method for inspecting railroad track using a laser and camera to collect information about the railroad track, plus a processor to analyze the information according to a specified algorithm: FC [16], [31]. The individual components were known and there was no suggestion of inventive ingenuity in adapting those components to implement the algorithm: FC [146]. One of the problems addressed by the 082 patent was identifying “plate cut” which occurs when a tie plate, which fastens the rail to the wooden crosstie, cuts into the crosstie as a result of the vibration of the passing trains. Fothergill J noted that all of the components and technology used by the claimed invention were known in the prior art, but “none of the prior art identifies plate cut or rail seat abrasion as problems to be solved, or suggests solutions similar to those disclosed by the 082 and 249 Patents”: FC [146]. However, as Dawson JA emphasized, this statement was based on the six items of prior art that were submitted in Tetra’s expert report, and did not consider whether those elements were found in the common general knowledge as accepted by Fothergill J: FCA [60], [62]. So, plate cut itself was well known in the prior art, and indeed was required to understand the undefined term “sunken plate” (a plate that has sunk into the tie through plate cut) used in the patent itself [18], so it must have been obvious to identify plate cut as a problem [66]. There was also a problem in how to identify plate cut, because plate cut is hidden when viewed from above since it occurs under the tie plate. The 082 patent solved this problem by comparing the height of the crosstie with the height of the tie plate [67]. But, as Dawson JA noted, “This was the very measurement human inspectors took when trying to identify plate cut” [68]; Georgetown’s own expert testified to that effect [70]. Thus the method of identifying plate cut was known in the prior art, and Forthergill J had found that there was nothing inventive in the way that that known solution was implemented by the automated system.

The strength of these points is illustrated by Georgetown’s strained response, that “although defects such as plate cut and rail seat abrasion were defects known to ‘railway people’, these defects were not held by the Court to be within the common general knowledge of the person skilled in the art” [22]. That is, the problem and solution adopted by the patent would not have been obvious to a person such as myself, wholly ignorant of railways. But Dawson JA had no difficulty in dismissing this argument, both in light of Fothergill J’s express findings regarding the common general knowledge, and in light of the general principle that “[t]he identification of the ordinary person skilled in the art should be consistent with the specification of the patent” [26].

Forthergill J’s failure to address obviousness in light of the common general knowledge was an error of law [63]. The principle applied by the FCA is well established and the decision does not reflect any change in the law.

Patent Remedies and Complex Products: Toward a Global Consensus

I’m very pleased to announce the release by Cambridge Univ. Press of Patent Remedies and Complex Products: Toward a Global Consensus (Brad Biddle, Jorge L. Contreras, Brian J. Love & Norman V. Siebrasse eds., Cambridge Univ. Press 2019). Here is the description:

Through a collaboration among twenty legal scholars from eleven countries in North America, Europe and Asia, Patent Remedies and Complex Products presents an international consensus on the use of patent remedies for complex products such as smartphones, computer networks and the Internet of Things. It covers the application of both monetary remedies like reasonable royalties, lost profits, and enhanced damages, as well as injunctive relief. Readers will also learn about the effect of competition laws and agreements to license standards-essential patents on terms that are 'fair, reasonable and non-discriminatory' (FRAND) on patent remedies. Where national values and policy make consensus difficult, contributors discuss the nature and direction of further research required to resolve disagreements.

The book is the work of the project on International Patent Remedies for Complex Products (INPRECOMP), which brought together the aforementioned twenty scholars for two workshops, plus innumerable rounds of collaborative work in drafting the individual chapters. At the second workshop, a panel of prominent judges and practitioners with experience in complex products litigation provided feedback on a draft version of the work, and the book is aimed not only at academics but also lawyers, judges, and other professionals involved in the global patent system.

Patent Remedies and Complex Products is available in print from Amazon or direct from Cambridge U Pr and also in an open-access version which provides the full text in pdf and html for free download and viewing.

The project was managed through the Center for Law, Science and Innovation (CLSI) at Arizona State University's Sandra Day O'Connor College of Law and primarily funded by a gift to CLSI from Intel.

While I’m on the subject of patent remedies for complex products, I’ll mention my book chapter with Professor Tom Cotter, "Judicially Determined FRAND Royalties," in Jorge L Contreas (ed), The Cambridge Handbook of Technical Standardization Law: Competition, Antitrust, and Patents (Cambridge University Press, 2017), Chapter 23. It is a comprehensive review of the basis for calculation of FRAND royalties in all the cases that were decided by the time it was written (which still includes most of the leading cases), and I suggest it would be of interest to any lawyer involved in FRAND royalty litigation

Recovery of Loss from Displaced Sales

Arysta Lifescience North America, LLC v AgraCity Crop & Nutrition Ltd 2019 FC 530 Pentney J
            2,346,021 / flucarbazone sodium herbicide / EVEREST

In Arysta v AgraCity the patentee, Arysta, sought, and was granted, an interim injunction, on the basis that AgraCity’s sale of a generic herbicide would infringe Arysta’s 021 patent. Yesterday’s post discussed the ability-to-pay aspect of the irreparable harm issue. On a separate point, AgraCity argued Arysta would not suffer substantial damages because the 021 patent only covers the granular or powder form of the herbicide.

AgraCity’s product is a generic of the EVEREST 1.0 granular product formerly sold by Arysta [9], but Arysta no now only sells a new and improved liquid product, EVEREST 3.0 [54]. AgraCity argued that because Arysta no longer sold a product embodying the patented invention, “any losses that may flow to Arysta from the sale of its generic product must be limited to a reasonable royalty” [54]. Pentney J rejected this argument, essentially on the basis that the patentee is entitled to any losses caused by the infringement; if the sale of the generic granular product displaced sales of Arysta’s liquid product, those losses would be recoverable [68]. I won’t go through Pentney J’s analysis in detail, except to say that it strikes me as entirely correct.

Standard for Proving Irreparable Harm Due to Inability to Pay

Arysta Lifescience North America, LLC v AgraCity Crop & Nutrition Ltd 2019 FC 530 Pentney J
            2,346,021 / flucarbazone sodium herbicide / EVEREST

It is notoriously difficult for a patentee to obtain an interlocutory injunction in Canada, primarily because of the stringent test applied in the Federal Courts in respect of irreparable harm, the second branch of the tripartite test for an interlocutory injunction set out in RJR-MacDonald [1994] 1 SCR 311. In this decision, the patentee, Arysta, was able to overcome that hurdle. Pentney J granted Arysta’s motion for an interim injunction, after finding that Arysta had established irreparable harm [72], the other elements of the RJR-MacDonald test were satisfied [40], [80] and also that the matter was urgent, as is required for an interim injunction [3], [81].

Despite the result, this decision does not, on its face, mark any relaxation of the stringent test for irreparable harm. Pentney J noted and accepted the FCA jurisprudence that establishes the threshold for irreparable harm as “very high” [21], and he emphasized the “particular circumstances of this case” [3], [81]. The decision nonetheless raises a question as to whether any different approach will be applied when the alleged irreparable harm stems from the defendant’s inability to pay a damages award, as it was here.

The case can be distinguished straightforwardly on the facts from the more typical Federal Court patent cases in which an interlocutory injunction is denied. In such cases the key question on irreparable harm is whether damages will be fully compensatory: see here for a review of some of the recent decision. In principle, of course, damages are intended to provide full compensation, so the question is whether, on the facts of the case, damages are so difficult to assess as to make it likely that the patentee will be undercompensated, should it win at trial. In this case, the focus was not the court’s ability to assess damages accurately, but rather on the defendant’s ability to pay any damages which might be awarded.

At first blush, such a distinction makes perfect sense: even if the court assesses damages perfectly, the plaintiff will clearly suffer irreparable harm if it is unable to actually recover the amount awarded. However, it is not clear that the distinction supports a different approach to irreparable harm in a case such as Arysta. The high threshold for irreparable harm in the Federal Courts stems, in principle, from an evidentiary principle. The FCA has consistently held that “the evidence as to irreparable harm must be clear and not speculative”: ICI v Apotex (1989) 27 CPR (3d) 345 (FCA) 351. That is, the jurisprudence does not say that difficulty in assessing damages can never in principle amount to irreparable harm, but rather that the difficulty in assessing damages must be established on a firm evidentiary basis. The rule that “the evidence as to irreparable harm must be clear and not speculative,” on its face applies to any allegation of irreparable harm, whether the substantive issue is difficulty of quantification or ability to pay.

So, when the patentee alleges irreparable harm based on the defendant’s inability to pay a damages award, it would seem that this inability to pay must be established on the basis of evidence which is clear and not speculative. However, Pentney J’s analysis did not turn on this general principle. Instead, he reviewed the jurisprudence on the evidentiary standard relevant to inability to pay in particular:

[46] The concern about an inability of the defendant to pay a damages award will be difficult to establish, since the information relevant to the question will be largely in the control of the defendant. The onus on the plaintiff for establishing irreparable harm on this basis has been variously expressed as: “a very serious doubt” (Bulman Group Ltd v Alpha One-Write Systems British Columbia Ltd (1981), 54 CPR (2d) 171, at p 172); or “strong indications that a substantial judgment would be uncollectable” (Dyckerhoff & Widmann AG v Advanced Construction Enterprises Inc, [1986] 1 FC 526, at p 534 (TD); or “reasonable grounds for concern that it is unlikely that any substantial monetary judgment… could be collected or enforced” (Titan Linkabit Corp v SEE See Electronic Engineering Inc (1993), 48 CPR (3d) 62, at p 78 (FCTD)).

It’s reasonable enough to cite authorities on the issue of ability to pay specifically, as that is a narrower subset of the more general question of irreparable harm, and so cases on that point should be more directly applicable. But cases on a narrower point should nonetheless reflect the application of the broader principle on the more specific facts, and it is not clear to me that the standard articulated in these cases is consistent with the more general requirement that the evidence of irreparable harm must be established on the basis of evidence that is “clear and not speculative.” The standard of “reasonable grounds for concern” (Titan Linkabit) seems to me to be clearly lower. “Strong indications” (Dyckerhoff) also strikes me as a lower standard, though perhaps “a very serious doubt” (Bulman) is closer to “clear and not speculative.” None of these cases explains how these standards fit with the modern requirement that the evidence of irreparable harm must be “clear and not speculative,” and indeed none of them mentions that standard, nor any of the current leading cases. No doubt this is simply because the “clear and not speculative” standard was not fully established when they were decided.

Nor does Pentney J clearly explain the relationship between these standards. On the facts, he held that irreparable harm had been established because Arysta had demonstrated “serious reason to doubt” whether AgraCity would be able to pay a damages award [50], and also that “Arysta has demonstrated, through clear and non-speculative evidence, that there is a substantial risk” that AgraCity would not be able to pay a damages award [53].

The phrase “serious reason to doubt” used at [50] seems to me to be a clearly lower standard than “clear and not speculative.” The evidence reviewed by Pentney J related to AgraCity’s ability to pay at [47], strikes me as somewhat speculative, at least by the standards normally used in establishing irreparable harm, though certainly adequate to raise a serious doubt. This suggests that Pentney J was applying a “serious reason to doubt” standard rather than a “clear and not speculative” standard, though of course he may simply have viewed the evidence differently, particularly given that I don’t have the benefit of the full record.

Pentney J’s statement at [53] seems to imply a distinction between the degree of risk (“substantial”) and the evidence (“clear and not speculative”). Perhaps that is a sound basis for reconciling the various standards, but I’m not sure such a distinction is consistent with the jurisprudence dealing with cases where the concern is quantification, rather than ability to pay.

If there is some kind of different standard, perhaps it is justifiable on the basis that “the information relevant to the question will be largely in the control of the defendant” [46]. A similar point was made in Bulman, when the court remarked that the serious doubt arose because “[t]here was in effect a refusal [by the defendant] to divulge information as to financial responsibility” (172). This does strike me as a principled evidentiary reason for distinguishing cases involving ability to pay from those involving difficulty of quantification. It is also an evidentiary point, which might therefore be taken as a refinement of the general evidentiary rule. Perhaps the principle is that the plaintiff must establish irreparable harm on the basis of evidence that is clear and not speculative, unless the relevant evidence is largely in the control of the defendant, in which case it is enough to raise a serious doubt. With that said, even if this point is consistent with Bulman, I don’t see it as driving the result in Bulman, which took a more holistic approach — indeed, Bulman considered ability to pay in the context of the balance of convenience as much as in the context of irreparable harm. And I don’t see this evidentiary point as explaining Titan Linkabit or Dyckerhoff, neither of which mentioned such any evidentiary difficulty.

In summary, Pentney J’s analysis of the significance of ability to pay is consistent with the cases he cited, which were on point, but those cases are relatively old, and it is not clear to me they are consistent with the modern requirement that the evidence of irreparable harm must be clear and not speculative. I do not mean this as any criticism of Pentney J. On the contrary, apart from this narrow issue, his application of the relevant principles to the facts strikes me as entirely sound. And to the extent that the decision illustrates a tension in the standard for irreparable harm, I am not suggesting that the older cases should be rejected, as I have long been critical of the current approach to irreparable harm: see Norman Siebrasse, ‘Interlocutory Injunctions and Irreparable Harm in the Federal Courts’ (2010) 88 Canadian Bar Review 517.

Obviousness Analysis Need not Focus on the Inventive Concept

Tearlab Corporation v I-MED Pharma Inc 2019 FCA 179 de Montigny JA: Nadon, Gleason JJA aff’g 2018 FC 164 Manson J
            2,494,540 / TearLab System / i-Pen System

In the decision under appeal, Manson J held certain claims of TearLab’s ‘540 patent to be infringed, but invalid due to anticipation and obviousness: see here. On appeal, TearLab essentially just argued that Manson J had erred in his claim construction and in his conclusions, with the usual attempt to elevate these criticisms into an error of law, eg by arguing that Manson J had erred in how he had applied purposive construction on the facts [2], [36]. The FCA has dismissed the appeal, affirming Manson J’s holdings, either on the basis of the applicable deferential standard of review [27], or on the basis that Manson J’s conclusions were correct [80].

There are a few legal points of passing interest. In light of the uncertainty as to the nature of the “inventive concept” and its role in the obviousness inquiry, the recent FCA decisions have “downplayed the importance of the ‘inventive concept’ as an analytical tool in the context of an obviousness analysis,” and consequently it is not an error for the trial judge to focus on the claims rather than the inventive concept in the obviousness analysis [76]-[78].

Second, as discussed here, in Aux Sable 2019 FC 581 Southcott J held that the state of the art for the purposes of an obviousness attack includes all prior art, and not just that discoverable in a reasonably diligent search. In the TearLab appeal, TearLab argued that “the judge erred in not determining whether the prior art at issue was citable, ie whether it would have been located using a reasonably diligent search” [54]. The FCA dismissed this point on the basis that the relevant evidence was indeed discoverable [55]–[58], which implies that the reasonably diligent search test is indeed applicable. However, nothing much can be made of this, as the legal point does not appear to have been argued, and was not addressed by the FCA.

Third, on another point related to obviousness, TearLab argued that “the judge failed to take into account relevant secondary indicia of obviousness, such as the ‘long-felt want or need’” [67]. It is sometimes suggested that secondary evidence is necessarily substantively secondary to the “primary” evidence of expert witnesses who testify as to whether the invention would have been obvious to a POSITA: see eg Mölnlycke AB v Procter & Gamble Ltd [1994] RPC 49 (CA), 113. I have argued that this position is unsound; secondary evidence has both advantages and disadvantages as compared to the evidence of experts, and one or the other might be more persuasive on the particular facts of a case: see ‘“Secondary” Evidence of Obviousness is Not Secondary’ (2012) 28 CIPR 279 (draft version here); and various blog posts. The FCA dismissed TearLab’s argument, but not on the basis that secondary evidence is necessarily of lesser importance, rather because “the commercial success of a patented product is never conclusive, in and of itself, and is clearly not sufficient to save an obvious claim,” [68] and more specifically because

[69] to the extent the evidence of commercial success and industry praises put forward by the appellant related not to the invention disclosed in the patent itself, but rather to the TearLab System, and that no nexus was clearly shown between the two, it was irrelevant to the obviousness inquiry.

This illustrates one of the substantive limitations of commercial success, namely, for commercial success to be relevant, it must be shown that the success was due to the patented invention, and not to other aspects of the product embodying that invention.

Professional Judgment and Subject Matter

Aux Sable Liquid Products LP v JL Energy Transportation Inc 2019 FC 581 Southcott J
            2,205,670

A final point raised by Aux Sable concerns patentable subject matter. Southcott J held Claims 9-10 to be invalid as being addressed to non-patentable subject matter, on the basis that “practice of claims 9-10 to a successful outcome within the ranges of compositions prescribed does depend on the exercise of professional judgment” [238]. In so holding, Southcott J relied on the dosage range cases such as Axcan 2006 FC 527 [232], [238]. I have criticized this line of cases at some length in this post, and I note that in Cobalt v Bayer [YAZ] 2015 FCA 116 [101] the FCA remarked that my posts had “forcefully advanced arguments of policy and logic against the current position.” Southcott J’s discussion was very brief — a single paragraph — and he remarked that “little turns on the outcome of this analysis,” as he had already held these claims invalid for other reasons [238]. Consequently, Aux Sable does not add any weight to the existing line of cases, particularly in light of the concerns expressed by the FCA.

A Person of Ordinary Skill Is Not a Person of Ordinary Skill

Aux Sable Liquid Products LP v JL Energy Transportation Inc 2019 FC 581 Southcott J
            2,205,670

I’ve already written posts discussing issues related to obviousness, utility, overbreadth and anticipation that were raised in Aux Sable, but there are still a couple of outstanding issues of minor interest.

One is that when it comes to obviousness, the legal person of ordinary skill in the art is not a person of ordinary skill in the art. Several of the expert witnesses in this case were confused on this point (emphasis added):

[27] However, I do find merit to the concern that JL Energy raises about [Aux Sable’s expert] Dr. Sharma’s understanding of the Skilled Person. Dr. Sharma was clear in his testimony that his conception of the Skilled Person was an engineer with an average level of inventiveness. This conflicts with the description of the Skilled Person by Mr. Justice Rothstein, in Sanofi-Synthelabo Canada Inc. v. Apotex Inc., 2008 SCC 61 [Sanofi] at para 52, as a “… technician skilled in the art but having no scintilla of inventiveness or imagination …”.

A similar error was made by JL Energy’s experts, Dr. Monnery [34] and Dr Ryan [39]–[40]. All three made the error of thinking the POSITA would have some degree of inventiveness; this error is quite understandable, given that the words “person of ordinary skill in the art” do suggest, well, a person of ordinary skill in the art, and all three experts believed that an ordinary engineer of average skill would have some degree of inventiveness — a belief no doubt shared by anyone who has ever interacted with actual engineers or scientists.

This view, while entirely understandable, is an error of law, as pointed out by Southcott J [34]. The POSITA does not reflect a person of ordinary skill, but rather a “mythical” being, described in Beloit (1986), 8 CPR (3d) 289, 294 (FCA) as having “no scintilla of inventiveness or imagination; a paragon of deduction and dexterity, wholly devoid of intuition; a triumph of the left hemisphere over the right” – that is, a being nothing much like an actual ordinary scientist or engineer. Southcott J capitalized Skilled Person in his reasons, which helps remind us that we are not talking about a skilled person.

The legal point made by Southcott J is not new, but the experts’ confusion should give us pause. The original statement of the test for obviousness, which is the ultimate source of the test set out in the Act, was the famous Cripps question, formulated by Stafford Cripps, KC, in argument in Sharp & Dohme v Boots Pure Drug Co (1928) 45 RPC 153 (CA), 162–63 and adopted by Lord Hanworth MR at 173:

‘The real question is: was it for all practical purposes obvious to any skilled chemist in the state of chemical knowledge existing at the date of the patent which consists of the chemical literature available (a selection of which appears in the Particulars of Objections) and his general chemical knowledge, that he could manufacture valuable therapeutic agents by [the process specified in the patent]…’

This, on its face, appears to be a factual inquiry about a skilled chemist, who is hypothetical in representing an average person, rather than any particular individual, but otherwise actually ordinary in the degree of skill. How then was the “skilled person” of patent law transformed from a person of average ingenuity, to one with no scintilla of inventiveness? To answer this question is beyond the scope of a blog post, but the confusion of the experts in Aux Sables is a reminder of this puzzle.

Amfac Approach to Overbreadth Applied

Aux Sable Liquid Products LP v JL Energy Transportation Inc 2019 FC 581 Southcott J
            2,205,670

The overbreadth doctrine holds that a claim that exceeds the scope of the invention disclosed in the specification is invalid. A key question in assessing overbreadth is therefore how to determine what constitutes “the invention.” In my draft paper on Overbreadth in Canadian Law, I argue that an approach to overbreadth which focuses solely on the disclosure in assessing the nature of the invention is likely to result in arbitrary and erroneous results. I particularly criticized Amfac Foods (1986), 12 CPR (3d) 193 (FCA), as being wrongly decided, both in the result and, more importantly, in its approach. In Aux Sable, Southcott J held two claims to be invalid, primarily on the basis of overbreadth. In so holding, Southcott J relied on Amfac [57], [73], and in my view, his analysis reflects the weaknesses identified in my article. While Amfac is regularly cited for the general proposition that a claim will be invalid if overbroad, I believe that Aux Sable is the first time since Amfac that the method used in Amfac has been applied.

The invention at issue in Aux Sable relates to the transportation of natural gas by pipeline. Finding ways to pump gas more efficiently, thereby using less energy, is evidently desirable. Natural gas is primarily methane, CH4, with one carbon atom. Heavier weight hydrocarbons, such as ethane and propane are referred to as C2 and C3 hydrocarbons because they have two and three carbon atoms, respectively. Ethane and propane require more energy to transport than methane because they are heavier, but they are more compressible, which makes them easier to pump. The inventors of the 670 patent discovered that because of this tradeoff, it turns out that the right amount of heavier weight hydrocarbons can provide a net benefit, so the gas requires less energy to pump [5]-[6]. Recall that Aux Sable, the accused infringer, is the plaintiff in this impeachment action, while JL Energy, the patentee, is the defendant [10].

The key parameter for determining the energy reduction is the “zMw product” (the product of the molecular weight (Mw) and the compressibility (z) [7]). Claim 1 of the 670 patent claims a method of transporting natural gas by adding sufficient C2 and C3 hydrocarbons so as to ensure the zMw product is lowered [51]. Claims 2-8 are dependent claims particularizing various parameters. All of these claims were held to be valid [2]. Claim 9 is to the gas mixture itself, specified in terms of temperature, pressure and hydrocarbon composition. Claim 10 is dependent with further particularization. Crucially, Claims 9 and 10 do not include the two elements of adding C2 and/or C3 and evaluating zMw [52]-[54].

Both Claims 9 and 10 were held to be invalid, primarily for overbreadth [2], in the sense that it was the first ground considered by Southcott J, and he remarked that “it is not strictly necessary for me to consider other grounds of invalidity” [74], though he did so nonetheless for completeness.

Most of Southcott J’s analysis of overbreadth was spent in dismissing various objections by JL Energy, such as the argument that the promise doctrine was being improperly imported [62]-[71]. The substance of the overbreadth analysis was quite brief. The key passage from Amfac was paragraph 32, quoted by Southcott J at [57] (my emphasis):

32 … The weakness in the claim in issue here is that the claim failed to mention essential elements disclosed as part of the invention. . . . While there can be no question that a patent must be fairly construed, if such fair construction reveals that an essential element (in this case a limitation) has not been claimed, the omission is fatal to the claim’s validity.

This is fine so far as it goes, but the crucial question is how to determine the essential elements of the claim. This is equivalent to defining “the invention”.

[58] Aux Sable’s argument, in relation to claims 9 and 10 of the 670 Patent, is that the claims are broader than the invention disclosed in the patent. As claims 9 and 10 claim natural gas mixtures regardless of whether ethane or propane was added to the mixture and regardless of whether the zMw product is reduced after such addition. Aux Sable argues that the elements necessary to limit claims 9 and 10 to the invention of the 670 Patent are missing.

Southcott J then observed that the evidence was to the effect that the addition of C2 and/or C3 and the zMw limitation were essential elements: for example, “Dr. Ramsay’s report expressly refers to each of the addition of C2 and/or C3 and the zMw limitation as an ‘essential limitation’ that is omitted from claims 9-10” [73]. Southcott J’s conclusion that the claims were invalid for overbreadth followed directly [74].

So, Claims 9-10 were held to be invalid simply because they did not contain any reference to the addition of C2/C3 and monitoring zMw. Now, in one sense it is true that adding C2/C3 and monitoring zMw were essential to the invention, for all the reasons given by Southcott J and the experts on which he relied. But they were essential in the sense that they reflect the inventive concept. Thus Southcott J’s holding amounts to saying that the inventive concept must be explicitly present in the claims. There is no such requirement.

For example, a compound with no known use is unpatentable, even if it is novel. If the inventors discover it is useful in treating cancer, the compound may be patented as such; the claim itself need not mention the use. That is true even if the compound is easy to synthesize, so that the inventive concept lies solely in the discovery of the new use; if the compound is new, it may be claimed as such, without reference to the use. As another example, if the invention is an enantiomer of a known racemic compound which was known to be useful, and the inventive step lay in discovering a way to separate the enantiomers, the enantiomer may be claimed as such, without reference to the method of separation: Sanofi 2008 SCC 61; Lundbeck [2009] UKHL 12. In Aux Sable, the discovery of the importance of the zMw product is the inventive concept, and Claims 9-10 to the gas mixture itself, as the product of that inventive method, are analogous to claims to the compound per se, which are produced by an inventive process of adjusting the zMw product. (Claims 2-8 are analogous to claims to the method of separating enantiomers.) (I should add that there is a discussion at [66]-[68], that I find difficult to understand, but which might be taken to suggest that the claimed gas mixture was not in fact the product of the inventive method of adjusting zMw. If so, then the claims should normally be invalid for lack of utility in any event, as Southcott J suggests. This was not a key point in Southcott J’s analysis.)

In the great majority of cases in which overbreadth is invoked, as discussed in my Overbreadth paper, the “essential elements” means the elements of the claim that ensure it is new, useful and non-obvious. So, in BVD Co v Canadian Celanese Ltd [1937] SCR 221, 237 the Court remarked that “the inventor did not state in his claims the essential characteristic of his actual invention,” which was the use of a thermoplastic woven into yarn, but the result was that the claims also encompassed thermoplastic in the form of sheets, and the Court went on to emphasize that the specification made it clear that the use of thermoplastic woven into yarn was an essential aspect of the invention, and that the claims themselves omitted any reference to this essential feature (228-30). But the reason all of this was relevant is that “[u]nless the claims…can properly be narrowed by the introduction of a limitation to the use of the cellulose derivative in the form of yarns, filaments or fibres, they are, we think, clearly anticipated [by two prior art patents]” (230). The Court then spent more than several pages detailing how the prior art patents anticipated the patent at issue, if the claims were not limited to the use of woven thermoplastic yarns (230-33). So, exactly the same facts which meant the claim was anticipated also meant it was overbroad. The essential element was the element that made the disclosed invention new; omitting this element was fatal to the claim because it meant the invention claimed was not new, even though the invention disclosed in the specification was new. In short, the claims were overbroad because they were anticipated, and they were anticipated because they were overbroad. A claim may similarly be overbroad because it claims material which is obvious, or useless, or not sufficiently disclosed. That is true in the great majority of cases invoking overbreadth; in effect, overbreadth is used as a compendious way of saying that the claim claims subject matter which is not an invention.

Used this way, the “essential elements” of the claim are those which would be identified as essential in a claim construction exercise under Free World. Southcott J considered this point and rejected it:

[72] I appreciate that, as explained in Free World Trust v Électro Santé Inc., 2000 SCC 66 [Free World Trust] at paras 20-23, it may be necessary, for consideration of both validity and infringement allegations, to identify essential and non-essential elements of the claims of a patent in the course of the antecedent claim construction. However, this is clearly not the same analysis that is contemplated by paragraph 32 of Amfac Foods. That paragraph refers to essential elements disclosed as part of the invention and the need to consider whether such elements are mentioned in the patent’s claims. Such analysis does not involve whether a particular element is an essential or non-essential component of the claims (the process contemplated by Free World Trust), but rather considers whether the element is found in the claims at all.

Southcott J is correct in his observations regarding Amfac. This is not inconsistent with my remarks above. The problem is how to identify the “essential elements” for the purpose of overbreadth. I said that in “the great majority of cases” the term is used to mean the elements which are essential to patentability. There is also a handful of overbreadth cases, raising the “roads to Brighton” problem, in which a patentee who has discovered one method of achieving an end result which is known to be desirable, claims the end-result itself: see eg Biogen [1996] UKHL 18, Lundbeck v Apotex 2009 FC 146, aff’d 2010 FCA 320. And then there is Amfac, which is an outlier. As I read the cases, Amfac is the only case to hold claims invalid for overbreadth based solely on the identification of the essential elements from the specification, without regard to whether the putative overbreadth also resulted in anticipation, obviousness or some other statutory ground of invalidity.

Now, in some ways, Southcott J’s approach is more principled than that in Amfac. Southcott J at least identified the essential elements with the inventive concept as identified by the expert evidence, whereas in Amfac, the essential elements were identified as those which appeared to be emphasized by the specification, without regard to whether it actually reflected the inventive concept. Nonetheless, both approaches raise the same problem, in that they potentially invalidate claims to subject-matter which is new, useful and non-obvious. If the specification sufficiently discloses how to make and use that subject matter, the patentee has fulfilled its end of the patent bargain. What justification is there for striking down the claims? The question is particularly pressing given that overbreadth has no clear basis in the Act. In cases such as BVD, the answer is clear, as the justification for striking down the claim for overbreadth is exactly the same as that for striking it down for lack of novelty. But that straighforward justification was not available in Amfac, as there was no suggestion that the claims which were held to be overbroad would otherwise have been invalid on any statutory basis.

Now, Southcott J also held Claims 9-10 to be invalid for anticipation, lack of utility, and unpatentable subject matter, but these were independent considerations, which might well have gone the other way on the facts. This is different from cases such as BVD, along with most other overbreadth cases, in which the basis for holding the claims overbroad was exactly the same as the basis for holding it to be anticipated.

In summary, the overbreadth analysis set out in Amfac can result in invalidation of claims to subject-matter which is new, useful, non-obvious and sufficiently disclosed. No justification for this result is evident. The key to the problem is the identification of “essential elements” in the context of overbreadth. Aux Sable is the first case to actually follow the Amfac approach, as opposed to simply citing Amfac for the proposition that a claim is overbroad if essential elements are omitted, and the results are similarly problematic.

Rule Against Joinder in NOC Actions Is Strict

Teva Canada Innovation v Pharmascience Inc 2019 FC 595 Tabib J
            2,702,437 / glatiramer acetate

In this motion to strike, some unusual facts have given Tabib J the opportunity to clarify aspects of the relationship between an action under the new PM(NOC) Regulations and an action under the Patent Act. In particular, Tabib J held that it is clear that the prohibition on subsequent actions for infringement under s 6.01 applies only to a drug made in accordance with the specific SNDS which was the subject of the NOA; it is also clear that the rule against joinder of non-NOC actions is strict, with no room for discretion; but it is not clear – at least not clear enough to survive a motion to strike – whether a s6(1) action continues to be predicated on the allegations in the NOA, or whether it now looks to the content of the SNDS and whether a drug made in accordance with that submission would infringe a listed patent. (Note that I have a couple more posts on Aux Sable that aren't ready yet.)

Teva is the licensee* of the 437 patent, which claims glatiramer acetate for use in treating multiple sclerosis. Teva holds NOCs in relations to 20mg/mL and 40mg/mL dosages, but it only listed the 40mg dosage on the Register [3]. Pharmascience obtained an NOC for the 20mg dosage without having to address the 437 patent, and has been marketing it under the name Glatect for a couple of years [4], [6]. Pharmascience has now sought an NOC for the 40mg dosage, and has accordingly served Teva with an NOA addressing the 437 patent by alleging non-infringement and invalidity [6]. In its statement of claim responding to the NOA, Teva sought a declaration that both Glatect 20 mg and 40 mg would infringe, and seeking damages — not merely an order of prohibition – for past and current infringement in respect of Glatect 20mg [7]. Pharmascience moved to strike all aspects relating to Glatect 20mg [8].

The principal issue, therefore, was “whether it is reasonably arguable that Teva has a cause of action pursuant to section 6(1) of the Regulations in respect of Glatect 20 mg, and if so, whether that cause of action includes a cause of action for past and current infringement” [14].

The first issue is whether Teva has a reasonably arguable cause of action in respect of Glatect 20 mg pursuant to section 6(1), even though the 437 patent was not listed on the Register against the 20mg dosage, and Pharmascience had already obtained an NOC for it. Teva relied on the plain wording of s 6(1), to argue that “The cause of action contemplated in section 6(1) is. . . no longer predicated on the content of the NOA and on whether the allegations are justified, as it was under the old regime, but on the content of the submission and on whether a drug made in accordance with that submission would infringe a listed patent” [21]. Pharmascience used a purposive interpretation to argue that the scheme of the Regulations require “a clear relationship between the rights of action conferred by section 6(1) and the obligation of second person to serve a NOA” [22]. Tabib J refused to strike, not on the basis that Teva’s argument was correct, but that it was a complex issue of statutory interpretation which should not be decided on a motion to strike while the law respecting the new Regulations is in an “embryonic state” [23].

The next question was whether Teva has a reasonably arguable cause of action under section 6(1) in respect of past or current infringement by Glatect 20 mg. The answer is pretty clearly “no” on the face of s 6(1), but Teva made a purposive argument that the answer must be “yes” because otherwise it would be barred by s 6.01 from subsequently asserting those claims in an ordinary infringement action [26]. Section 6.01 is part of the scheme for addressing the “two kicks at the can” problem which arose under the old scheme, where the same issues could be litigated in relation to the same drug and same patent, once under the Regulations and again under the Act. In broad terms, s 6.01 provides that if the patentee does not respond to an NOA served under the Regulations in respect of a particular SNDS, it cannot subsequently bring an infringement action. Teva argued that in this case, if it could not bring its claim for damages under the Regulations, it would be barred from doing so in an action under the Act. Tabib J rejected this argument, holding that it was clear on the text of the provision that the prohibition on subsequent actions for infringement is only in respect of a drug made in accordance with the specific SNDS which was the subject of the NOA [28]-[30], in this case the 40 mg dosage.

A third question was whether Teva’s allegations related to Glatect 20 mg should be struck as improper joinder, contrary to s 6.02 of the Regulations, which, broadly, prohibits joinder of an ordinary infringement action with an NOC proceeding. Teva argued that on the facts of this case, joinder should be permitted, given that “the same patent, the same invalidity allegations, the same medication and substantially the same infringement allegations are at issue” [45]. Tabib j rejected this argument, noting while the Federal Court Rules give the court discretion in joinder decisions:

Section 6.02 is clear and mandatory. Unlike many other provisions of the Regulations, it does not provide for the exercise of the Court’s discretion to modify or depart from its provisions. Accordingly, the Court is bound to apply it, irrespective of any efficiency that might otherwise be gained [47].

However, Tabib J suggested that “scheduling the trial of common issues to proceed simultaneously” would not be contrary to s 6.02 [47].

*The patentee is Yeda, and strictly, the decision says only that Teva has obtained NOCs and listed the 437 patent on the Register with the consent of the Yeda, without referring to a formal licence [3]. However, Teva also asserts that it would be able to bring an infringement action under the Act and this was not contested [26]. This implies Teva is a person “claiming under” the patentee, presumably a licensee.

Blogging Break

I'm taking a blogging break this week, as I'm on vacation (in beautiful Squamish, BC, home to some of the best mountain biking in the world). I'll be back to blogging in the middle of next week with a post on the overbreadth issue in Aux Sable.

Is Overbreadth the new Promise Doctrine?

I have a new draft paper available on SSRN, “Overbreadth in Canadian Patent Law.” Here is the abstract

Under the overbreadth doctrine, a claim that exceeds the scope of the invention disclosed in the specification is invalid. While the doctrine is well established, it is redundant in the great majority of cases in which it is invoked, as an overbroad claim typically encompasses subject-matter which is not new, lacks utility, or is obvious. When overbreadth is not redundant, a puzzle arises: what is the principled justification for striking down a claim to an invention which is in fact new, useful, non-obvious and sufficiently disclosed? In such a case, how can it be said that the claim is broader than the invention? This article argues that overbreadth properly arises as an independent ground of invalidity in the context of the “roads to Brighton” problem, in which the question is whether the first inventor to achieve a result known to be desirable may claim the result itself or may claim only their particular method of achieving it, but the answer to this question is controversial. Overbreadth was also applied as a truly independent ground of invalidity by the Federal Court of Appeal in Amfac Foods Inc v Irving Pulp & Paper, Ltd. This article argues that Amfac was wrongly decided, both on its facts, and in its approach to overbreadth. The article warns that the Amfac approach, if widely adopted, risks invalidating patents for inventions which are new, useful and non-obvious, on the basis of an arbitrary parsing of the disclosure, in a manner reminiscent of the promise doctrine.

Teaching Away Irrelevant to Anticipation

Aux Sable Liquid Products LP v JL Energy Transportation Inc 2019 FC 581 Southcott J
            2,205,670

Recall that the invention at issue relates to a method of adjusting the composition of natural gas for more efficient transportation by pipeline. Claims 1-8 related to the method, while Claims 9-10 claim “a gas mixture, for use in a pipeline,” with specified ranges of the proportion of various gases, as well as specified temperature and pressure ranges. Prior posts have dealt with the state of the art and utility. Aux Sable also raises a couple of points of minor interest regarding novelty in respect of Claims 9 and 10.

In attacking the patent for lack of novelty, Aux Sable relied on prior art disclosing a gas mixture falling within the claimed range. Southcott J noted that the prior disclosure of a point within the prescribed range is anticipatory [90]. Moreover, he held that so long as the prior art provided an enabling disclosure of a point falling within the prescribed range, it was irrelevant that the prior art taught away from the prescribed range [98], [108]. This follows from the basic rule that enabling disclosure establishes anticipation [108], and once that test is satisfied, additional information that might be found in the prior art is irrelevant. As support, Southcott J at [98], [108] cited Snider J’s decision in Schering-Plough 2009 FC 1128 [87] which held that it was irrelevant that the prior art might teach non-infringing formulations.

I’d add only a couple of points. First, I happened to come across this pithy statement to the same effect, by Stafford Cripps KC in argument at trial in Sharp & Dohme Inc v Boots Pure Drug Co Ltd (1927) 44 RPC 367 (Ch) at 378:

“When one is dealing with anticipation, a, single document that tells one what is in the patent specification is sufficient to defeat the patent, even if the other documents say that it is nonsense.

The point was accepted by Astbury J at 395, though not in quite those terms.

The second point is that Southcott J also relied on Merck & Co v Pharmascience Inc, 2010 FC 510, [166]-[168] in which Hughes J held that a claim to the use of finasteride for the treatment of baldness in a dosage of about 1.0 mg, was anticipated by a prior patent disclosing the use of finasteride in a range of dosages, with doses from 5 mg/day to 2000 mg/day disclosed as examples. I must say that in my view Hughes J was very clearly wrong in so holding, in light of the well-established rule of the highest authority that “A signpost, however clear, upon the road to the patentee’s invention will not suffice. The prior inventor must be clearly shown to have planted his flag at the precise destination before the patentee”: General Tire [1972] RPC 457 (CA) at 486; Free World 2000 SCC 66, [26]; Sanofi 2008 SCC 61, [21]. While I agree that the law as stated by Southcott J was correct, his point was not aided by citing this particular decision.

Turning to a different legal issue, another piece of prior art was a Handbook entitled “Properties of Natural Gases and Volatile Hydrocarbon Liquids.” It was undisputed that this was widely used by those working in the field of transporting natural gas by pipeline [105]. The Handbook disclosed a gas falling within the specified range, but did not expressly state that the gas was for use in a pipeline, which was an explicit element of the claim. JL Energy submitted “that one cannot employ an inference in performing an anticipation analysis” [114]. This submission was no doubt based on the rule that trial and error is permitted at the enablement stage but not at the disclosure stage. Southcott J rejected this submission, noting “the inference represents simply the process by which the Skilled Person’s [sic] employs the CGK to interpret the prior art. Anticipation is assessed based on the prior art as the Skilled Person would understand it, and the Skilled Person can use his or her CGK as part of the analysis” [114]. That must be right – the Handbook, after all, was all about gases for use in a pipeline, and it is a good illustration of the point made by Lord Hoffmann in Synthon [2005] UKHL 59 [32] that at the disclosure stage “the skilled person is taken to be trying to understand what the author of the description meant. His common general knowledge forms the background to [that] exercise.”

When Is Utility “Related to the Nature Of” the Invention?

Aux Sable Liquid Products LP v JL Energy Transportation Inc 2019 FC 581 Southcott J
            2,205,670

Monday’s post dealt with the issue of what constitutes the state of the art for an obviousness attack, which was raised as a pure question of law. A number of other interesting legal issues were raised on the facts. This post provides an overview of the facts and discusses one such issue, namely what it means for utility to be “related to the nature of the subject-matter” of the invention.

The invention relates to the transportation of natural gas by pipeline. Finding ways to pump gas more efficiently, thereby using less energy, is evidently desirable. Natural gas is primarily methane, CH4, with one carbon atom. Heavier weight hydrocarbons, such as ethane and propane, are heavier, and so require more energy to transport, but they are also more compressible, which makes them easier to pump. The inventors of the 670 patent discovered that because of this tradeoff, it turns out that the right amount of heavier weight hydrocarbons can provide a net benefit, so the gas requires less energy to pump [5]-[6].

Ethane and propane are referred to as C2 and C3 hydrocarbons because they have two and three carbon atoms, respectively. The key parameter for determining the energy reduction is the “zMw product” (the product of the molecular weight (Mw) and the compressibility (z) [7]). Claim 1 of the 670 patent claims a method of transporting natural gas by adding sufficient C2 and C3 hydrocarbon so as to ensure the zMw product is lowered [51]. Claims 2-8 are dependent claims particularizing various parameters. All of these claims were held to be valid [2]. Claim 9 is to the gas mixture itself, specified in terms of temperature, pressure and hydrocarbon composition. Claim 10 is dependent with further particularization. Claims 9 and 10 do not include the two elements of adding C2 and/or C3 and evaluating zMw [52]-[54]. Both Claims 9 and 10 were held to be invalid for overbreadth, inutility, anticipation and non-patentable subject matter [2].

While the obviousness issue discussed in the last post related to Claims 1-8, most of the other interesting legal issues related to Claims 9-10. One of these concerns utility. On all these points, nothing turns on the difference between Claims 9 and 10, so I will refer to Claim 9. Recall that Aux Sable, the accused infringer, is the plaintiff in this impeachment action, while JL Energy, the patentee, is the defendant [10].

Claim 9 claims “a gas mixture, for use in a pipeline,” with specified ranges of various gases, in particular “from 68 to 92% by volume of methane, from 6 to 35% by volume of ethane, from 0 to 9% by volume of propane” [50]. Ranges for other gases, as well as temperature and pressure ranges were also specified. First, “JL Energy acknowledges that if there is any combination of composition, pressure and temperature, within the ranges prescribed by either of these claims, that lacks that scintilla of utility, then the claim is invalid” [81]. I’m not sure such a concession was warranted, particularly if the inoperable embodiments could have been identified without undue effort: see here, here and here. But since it was conceded, I will accept the point for the remainder of the discussion, as did Southcott J. Moreover, because of this concession, the facts were not developed in detail, and it may be that there was a substantial range which did not deliver benefits, or that identifying the beneficial combinations would have been unduly difficult.

Recall that the aim of the 670 patent was improved efficiency in transportation of gas in pipelines. The unchallenged evidence of Aux Sables established that the mixture specified in Claim 9 did not have improved efficiency across the full range [82]. JL Energy argued that this does not matter, and that utility is established if the claimed gas mixture is capable of being transported by pipeline [83], which it clearly was. The question then, is whether, to establish utility, it is enough that the specified mixture was capable of being transported by pipeline, or was it required that the mixture had improved efficiency, at least to some modest degree.

As Southcott J noted [76], in AstraZeneca 2017 SCC 36 [53] the SCC held that only a scintilla of utility is required, but it must be a scintilla “related to the nature of the subject-matter” of the invention: “a proposed invention cannot be saved by an entirely unrelated use. It is not sufficient for an inventor seeking a patent for a machine to assert it is useful as a paperweight.”

The question then is whether usefulness in being capable of being pumped through a pipeline is related to the nature of the invention. Southcott J held not, on the view that “the subject matter of the invention . . . relates to the efficient transport of natural gas” [85]. JL Energy’s argument that the mere fact that the gas mixtures claimed are capable of being transported in a pipeline is sufficient “is not quite as extreme as submitting that an otherwise useless machine has utility as a paperweight (per paragraph 53 of AstraZeneca). However, in my view, it reaches in that direction” [87].

My intuition is Southcott J’s analysis is right, but the problem is a tricky one that raises some basic conceptual questions regarding “the nature of the invention” for the purposes of utility. It is now clear law that to be useful, the invention must have a scintilla of utility that is related to the nature of the invention. This raises two question: (1) how much is a “scintilla”? and (2) what is the nature of the invention? We have many cases exploring the first issue, but few exploring the second. In most cases in which the invention lacks utility, the only utility asserted clearly is related to the nature of the invention, and utility is lacking because it has not even a scintilla of such utility. So, with pharmaceuticals, the question is whether the drug is useful in treating a disease. In Wandscheer [1948] SCR 1, the question was whether the snow-blower could actually throw snow. These cases raise the question of how much is a “scintilla”.

The problem here concerns branch (2), the nature of the invention. The gas mixture at issue, even if it did not deliver efficiency advantages, could certainly be pumped through a pipeline and burned at the other end. This is much more than a scintilla of utility – if that type of utility is relevant. The question of the nature of the utility is normally raised using hypothetical examples, such as using the invention as a paper weight, or a throw-away utility, such as using a complex invention as landfill. These examples make the basic point, but are so extreme as to be unhelpful when it comes to nuances. As Southcott J noted, this case is more difficult. Being pumped through a pipeline is much closer to the “nature of the invention” in this case than being used as a paperweight.

The “nature of the invention” should be given a purposive interpretation, consistent with the purpose of the utility requirement, which is to prevent speculative claiming [AstraZeneca 56]. This might suggest that the mixture should be held to lack utility if it does not deliver even a scintilla of efficiency improvement. Otherwise, a patentee might patent all sorts of novel gas mixtures, so long as they could actually be pumped through a pipeline, in the hopes that some of those mixtures would later prove to have some advantageous properties. On the other hand, perhaps it is enough that the new mixture gives “a useful choice”: Consolboard [1981] 1 SCR 504, 525. I must admit that I find the approach adopted in the seminal EPO decision in T 0939/92 AgrEvo, to be very intuitively appealing, at least on these facts. AgrEvo concerned claims to a genus of compounds, some of which, at least, were said to have herbicidal activity. However, it was not plausible that all the compounds in the genus would have the alleged herbicidal activity [2.2.2], and the Board held the claims invalid, not for lack of utility (industrial applicability), but for lack of an inventive step, on the view that the compounds which were not herbicidal were not an inventive solution to the problem of providing herbicidal compounds [2.7], but rather “a mere arbitrary choice” [2.5.3]. On the AgrEvo approach, the gas mixtures at issue here would be considered to have utility, because they could be pumped through the pipeline, but would nonetheless be invalid, as being an arbitrary selection. But I won’t push this too far; maybe AgrEvo itself should have been decided on utility grounds.

I won’t pursue this further. For now, suffice it to say the facts raise a very nice example for considering what constitutes the nature of the invention for the purpose of utility.