En Banc Activity / Featured / Opinions / Panel Activity / Petitions

Jeffrey A. Lefstin serves as a Professor of Law at the University of California, Hastings College of the Law. Prior to serving as a professor, he clerked for Federal Circuit Judge Raymond C. Clevenger III. Prof. Lefstin holds a Ph.D. in Biochemistry from the University of California San Francisco and a J.D. from Stanford Law School. He has written extensively and testified before Congress concerning the doctrine of patent eligibility.

Though described by the majority as “narrow,” the American Axle v. Neapco panel opinion sets forth two far-reaching expansions in the law of patent eligibility.

First, the panel opinion holds that a patentee’s alleged failure to describe how to implement an invention defined by a claim renders the claim ineligible under § 101—without needing to resort to the factual inquiries associated with the written description or enablement requirements of § 112. Second, the panel opinion holds that a claim may be “directed to” an ineligible law of nature in step one of the Mayo/Alice inquiry, even though neither the claim nor the specification recites it directly.

Authority for the first proposition is said to be the “O’Reilly test” for patent eligibility articulated by the Supreme Court in O’Reilly v. Morse in 1854.[1] That case, however, represented a conventional application of the statutory disclosure requirements now lodged in § 112, as I explain in detail below. As authority for the second proposition, the revised panel opinion invokes Neilson v. Harford, the famous case on James Neilson’s hot-blast patent, decided by the Court of Exchequer in 1841. But, again, in seeking to force the American Axle claims into the mold of the Mayo/Alice test, the Federal Circuit has imagined in old precedent a resemblance that does not exist. 

The Federal Circuit’s Misplaced Reliance on Neilson v. Harford

The revised American Axle panel opinion justifies finding a natural law, in a claim that does not itself recite one, by asserting that Neilson v. Harford involved the very same situation.[2] Quoting from the Neilson case, Judge Dyk’s opinion tells us that Neilson “was not aware of the nature and principle of his discovery,” that hot air would promote ignition better than cold. This “principle” was supposedly not in the patent itself, but was nonetheless embodied in the patented machine. Following Mayo, Judge Dyk then explains that Neilson’s patent was held eligible only because it implemented the principle in an inventive way. Not so.

In invoking the authority of Neilson v. Harford, American Axle follows a long tradition in American patent law. Neilson and other cases on Neilson’s hot-blast patent were the foundation for many of the Supreme Court’s nineteenth-century cases that laid down the basic principles of American patent law, such as Le Roy v. Tatham, O’Reilly v. Morse, Winans v. Denmead, Tilghman v. Proctor, Risdon Iron & Locomotive Works v. Medart, and Boyden Power-Brake Co. v. Westinghouse.[3] More recently, Neilson was the ostensible authority for the ‘inventive application’ requirement for patent eligibility announced in Parker v. Flook and revived in Mayo v. Prometheus.[4] It seems as true today as when Henry Merwin wrote in his 1883 treatise, that for American courts and commentators Neilson’s “correctness has never been doubted or denied.”[5]

But American Axle’s account is a peculiar representation of the case. The inference in American Axle seems to be that the Exchequer questioned the eligibility of Neilson’s patent because the patent was directed to his discovery of a law of nature, even though—since Neilson was ignorant of his discovery—his patent was not directed to a law of nature on its face. But the “principle” of Neilson’s invention was his discovery that hot air was superior to cold air in the blast furnace. While it is true that the underlying physical and chemical mechanisms of the hot blast were not elucidated until years later, it was clear to everyone involved in the case—the patentee, the defendants, and the judges of the Exchequer—that Neilson had explicitly disclosed and claimed the advantage of hot air over cold. Indeed, the problem with Neilson’s patent was that Neilson had disclosed almost nothing else except the “principle” that hot air was superior to cold.

The quote that Neilson was “not aware of the nature and principle of his discovery” comes not from the Exchequer’s opinion, but from  the defendants’ principal argument:[6] that Neilson’s patent was invalid because he had failed to describe any practical means for heating air in the blast furnace. Neilson’s specification provided approximate dimensions for heating vessels suitable for small-scale forges and furnaces; for a blast furnace, he merely directed that “air vessels of proportionably increased dimensions and numbers are to be employed.” However, practical use of the hot blast for iron smelting would require a much more complex arrangement of tubes to maximize the surface area of, and limit thermal expansion in, the heating vessel. The defendants’ argument at trial was that these refinements in the heating apparatus were the real essence of the hot blast as applied to iron smelting. Neilson had not hinted at any such requirement and, according to the defendants, “had not the least idea of the important principles that are involved in the extension of this discovery to larger and other matters”[7] when he enrolled his specification.

So Neilson was certainly not a case that questioned the eligibility of a patent for alleged monopolization of an unrecited natural law. But the misuse of Neilson in the law of patent eligibility runs much deeper than American Axle’s new error. Judge Dyk repeats the profound misreading on which Mayo based its requirement for inventive application: that Neilson’s patent was held eligible because it “explained how the principle could be implemented in an inventive way.”[8]

As I have explained elsewhere,[9] nothing could be further from the truth. Neilson’s patent was sustained not because he implemented his discovery in an inventive way, but because the application of his discovery to the blast furnace required nothing beyond routine, well-known, and conventional activity. The defendants’ primary attack on the Neilson patent was enablement: they alleged that Neilson’s failure to describe the heating apparatus, as well as other refinements necessary to implement the hot blast in a blast furnace, rendered the specification defective.[10] Both the patentee and the court deflected that attack by pointing out that all the instrumentalities necessary to implement Neilson’s discovery were “perfectly well known” to those of skill in the art.[11] Indeed, in English law, Neilson became the primary exemplar for the proposition that applying a new discovery by entirely old and conventional means sufficed for patentability.[12]

The question of whether Neilson’s failure to specify the shape and form of the heating vessel invalidated his patent was ultimately resolved by the special findings of the jury, which determined that “a man of common understanding, of ordinary skill and knowledge of the subject . . . would be enabled to construct, according to the specification alone,” an apparatus with beneficial effect—regardless of the shape and form of the heating vessel.[13] So in the end Neilson was a case where the patent was challenged because it did not, in American Axle’s words, identify how a result could be achieved “by limiting the claim scope to structures specified at some level of concreteness.”[14] That challenge, however was resolved in the old-fashioned way: by asking a jury whether one of ordinary skill in the art could practice the invention based on the directions in the specification.

The Federal Circuit’s Misplaced Reliance on O’Reilly v. Morse

But the Federal Circuit’s failure to recognize Neilson as an enablement case points out American Axle’s more profound mistake: its assumption that O’Reilly v. Morse was similarly decided on eligibility grounds rather than enablement.  That assumption is American Axle’s keystone, and its justification for treating the patentee’s alleged failure to describe a means of implementation as a matter of § 101 rather than § 112. The premise that American Axle merely implements “the O’Reilly test” is the centerpiece of Judge Dyk’s and Judge Chen’s defenses of the panel opinion in the denial of en banc rehearing, and even those Federal Circuit judges who argued for en banc rehearing focus more on the facts of O’Reilly rather than on the basic question of whether the discussion in O’Reilly concerned eligibility at all. But the O’Reilly Court’s extensive discussion of Neilson makes clear that O’Reilly was an enablement case. The Court premised the invalidity of Morse’s eighth claim not on the ground that natural laws in the abstract were unpatentable—that was an old and trite conclusion well before O’Reilly was decided—but on the grounds that Morse had claimed more broadly than the statutory disclosure requirement would permit.

Neilson was Morse’s principal support for the validity of his eighth claim. For if Neilson’s patent could encompass the use of the hot blast without being bound to any particular apparatus he had disclosed, then why could Morse’s patent not encompass the use of electromagnetism for writing at a distance, without being bound to the telegraphic apparatus he had disclosed? In answer, the Court noted that although Neilson had not claimed any particular mode or form of heating apparatus, the Exchequer had construed his patent as claiming a machine, rather than an abstract principle.[15] More significantly, the Court pointed out the jury’s finding that one of ordinary skill in the art could, based on Neilson’s specification alone, readily adapt the preheating principle to “all cases of forges, cupolas, and furnaces, where the blast is used.”[16] Thus Neilson’s disclosure enabled the universal application of his discovery, no matter what form of heating or blowing apparatus was employed. Morse’s disclosure, by contrast, did not enable the practice of his invention independent of form, as the Court explained:

For Neilson discovered, that by interposing a heated receptacle between the blower and the furnace, and conducting the current of air through it, the heat in the furnace was increased. And this effect was always produced, whatever might be the form of the receptacle, or the mechanical contrivances for heating it, or for passing the current of air through it, and into the furnace.

But Professor Morse has not discovered, that the electric or galvanic current will always print at a distance, no matter what may be the form of the machinery or mechanical contrivances through which it passes.  . . .  To produce that effect, it must be combined with, and passed through, and operate upon, certain complicated and delicate machinery, adjusted and arranged upon philosophical principles, and prepared by the highest mechanical skill. [17]

The difference between Neilson’s patent and Morse’s was simply this: Neilson’s disclosure enabled the universal application of his discovery, beyond the form of heating or blowing apparatus he employed. Morse’s disclosure did not.

But one need not even appreciate Neilson as an enablement case to recognize that O’Reilly was one as well. For the remainder of the Court’s opinion—strangely ignored by all of the opinions in American Axleis devoted to showing that Morse’s eighth claim was invalid because it did not comply with the disclosure provisions of the Act of 1836.[18] Those provisions, in language that hardly differs from § 112, required the inventor to provide a written description of his invention and an enabling disclosure.[19] As the Court summed up when addressing whether a disclaimer could save the remaining claims:

The act of Congress above recited, requires that the invention shall be so described, that a person skilled in the science to which it appertains, or with which it is most nearly connected, shall be able to construct the improvement from the description given by the inventor.

Now, in this case, there is no description but one, of a process by which signs or letters may be printed at a distance. And yet he claims the exclusive right to any other mode and any other process, although not described by him, by which the end can be accomplished, if electro-magnetism is used as the motive power. That is to say—he claims a patent, for an effect produced by the use of electro-magnetism distinct from the process or machinery necessary to produce it. The words of the acts of Congress above quoted show that no patent can lawfully issue upon such a claim. For he claims what he has not described in the manner required by law.[20]

If the Court had wanted to express more clearly that the “O’Reilly test” was simply one of compliance with the statutory requirements of disclosure, I can scarcely imagine how it might have done so.


Once upon a time—by which I mean roughly from 1952 to 2012—patent law was relatively simple. If a claim seemed to do nothing much that was not done before, we asked whether the claim was non-obvious over the prior art. We called that a § 103 analysis. If a claim seemed vague or overbroad, we asked whether the claim specifically delineated the invention, whether the disclosure taught one of skill in the art how to make and use what the claim had delineated, and whether what was delineated by the claim was described by the disclosure. We called that a § 112 analysis. True, it was not always easy to assess whether the claim met those statutory requirements set forth in Title 35, but we had a reasonably clear idea of what we were supposed to be doing.

Yet now, some new doctrine has emerged from beneath the waves. Some claim that it evolved from § 101 of the Patent Act, though if so the drafters of that provision would have great difficulty recognizing its mutant offspring. Others maintain that it is nothing but the descendant of the ancient doctrines that came across the Atlantic when the patent law was young. Whatever its origin, it has a ravenous appetite not only for patents (some of which we admittedly might not mourn), but for the statutory requirements of patentability as well. First it ate § 103. Not sated with that provision of the statute, in American Axle it seems to have completed its digestion of § 112 as well.

[1] O’Reilly v. Morse, 56 U.S. (15 How.) 62 (1854).

[2] American Axle & Mfg., Inc. v. Neapco Holdings LLC, (Fed. Cir. Jul. 31, 2010), slip. op. at 27.

[3] For discussion of Neilson’s influence on American patent law, see Jeffrey A. Lefstin, Neilson v. Harford: Shape and Form in Patent Law, in Forgotten Intellectual Property Lore: Creativity, Entrepreneurship and Intellectual Property 77, 104-113 (Shubha Ghosh ed., 2020).

[4] Both Flook and Mayo relied on a severe misunderstanding of the hot-blast cases. See Jeffrey A. Lefstin, Inventive Application: A History, 67 Fla. L. Rev. 565, 581–93 (2015).

[5] Henry Childs Merwin, The Patentability of Inventions 539 (1883).

[6] The quote is actually the plaintiff’s paraphrase of the defendant’s contention. Neilson v. Harford, Webster’s Patent Cases 295, 344 (1841).

[7] Id. at 339-40. That is, they argued that Neilson had no conception of the improvements that would be necessary to implement the hot blast in a working blast furnace.

[8] American Axle, slip. op. at 27 (quoting Mayo, 566 U.S. at 83).

[9] See Lefstin, Inventive Application, at 581-593.

[10] In particular, Neilson had stated that the shape and size of the heating vessel was immaterial to the effect.

[11] Neilson, Webster’s Patent Cases at 337 (Alderson, B); id. at 344 (patentee’s argument).

[12] See Otto v. Linford [1882] 46 LT (N.S.) 35, 39-40.

[13] Neilson, Webster’s Patent Cases at 327 (jury findings); id. at 372 (Parke, B., for the Exchequer, explaining that court was bound by jury findings).

[14] American Axle, slip. op.  at 27-28.

[15] O’Reilly v. Morse, 56 U.S. at 116. The argument that Neilson claimed an abstract principle was not an argument pressed by the defendants in the case, who focused on the inadequacy of the specification; the judges of the Exchequer raised it of their own accord.

[16] Id. at 115.

[17] Id. at 116-17.

[18] Id.at 118-120.

[19] See id.

[20] Id. at 121.