join, dissenting from the court’s denial of rehearing en banc.
Almost exactly ten years ago, the Federal Circuit was formed in order to eliminate inconsistent law among circuits in patent cases. It was expected that the Federal Circuit would produce an internally consistent body of law of national effect, so that technological industry could rely on its decisions, in turn benefitting technological innovation. It was surely not intended that different panels of the Federal Circuit would deliberately issue conflicting law.
Since differences of panel opinion are not impossible, the Federal Circuit adopted the usual en banc procedures for resolving internal conflicts. However, the panel in Atlantic Thermoplastics Co. v. Faytex Cor*1282poration, 970 F.2d 834, 23 USPQ2d 1481 (Fed.Cir.1992), having disagreed with a point of law in Scripps Clinic & Research Foundation v. Genentech, Inc., 927 F.2d 1565, 18 USPQ2d 1001 (Fed.Cir.1991), has refused to follow this procedure for conflict resolution. Thus these overtly conflicting decisions will repose in the official reporters, I suppose some day to be resolved, but meanwhile to place this law in disarray. This does not serve the public, or litigants, or trial judges, who are entitled to know how the Federal Circuit will interpret a certain class of product-by-process claims, without depending on the luck of the draw of the appellate panel.
Thus I respectfully dissent from the court’s decision not to consider this issue en banc. I must dissent also from the Atlantic panel’s reading of precedent, and from its policy-driven conclusion.
I
The Claimed Inventions
In Scripps, supra, this court held that the “product-by-process” claims there at issue were properly interpreted as product claims, independent of how the product was made. The Scripps court followed the precedent of the Court of Customs and Patent Appeals and the Court of Claims,1 and is consistent with Supreme Court precedent. I shall discuss this precedent in some detail, for it illustrates the varieties of product/proeess claims, the different ways the law interprets them, and the inappropriateness of applying the interpretation of one class of claim to the facts of another class of claim.
The Scripps claims are of the class sometimes called “true” product-by-process claims, in that their patentability and validity depends on the novelty and unobviousness of the product, and it is immaterial whether the process is also patentable. Such inventions have had formal recognition at least since 1891, and continue to be made by inventors, and patented in accordance with the procedures set forth in the Manual of Patent Examining Procedure.
This class of claim usually appears when the invention is a chemical or biological product of such structural complexity that the product can not be defined in independent structural terms. The premise of such claims has been called the Rule of Necessity, for it provides a way of patenting inventions or discoveries whose structure is not sufficiently known or knowable to be described objectively. Such a discovery is illustrated by the complex blood clotting protein that is the subject of the product-by-process claims in Scripps.
The Scripps class of product-by-process claim is quite different from the claims now placed at issue by the Atlantic panel. The Atlantic claims here at issue are “product of the process” claims, such as may be allowed when the process is found patentable. Faytex Corporation pointed out that Atlantic was estopped by the prosecution history from taking the position that these claims are free of process limitations, for Atlantic had argued to the examiner that these claims were patentable solely because the process had been examined and found patentable. Allowance was obtained on that basis.
The applicant in Atlantic had separated his product claims into a divisional patent application, as a result of the examiner’s requirement for restriction between product and process claims. The Atlantic process claims and “product of the process” claims were grouped together and were issued in one patent, based entirely on examination of the process.2 These process claims were the only claims at issue at the Atlantic trial. Atlantic’s prod*1283uct claims were not issued until two years and nine months later. The Atlantic panel was told at oral argument that these claims have been rejected on reexamination.
Applying the normal rules of claim interpretation, it is apparent that there is no need whatsoever for the conflict that the Atlantic panel created. The claims in Atlantic and those of Scripps are different classes of claims, with a long history recognizing the difference. This history is exemplified not only in the decisions of the Court of Customs and Patent Appeals and the Court of Claims, but also in the decisions of the Supreme Court and the regional circuit courts. The stretch of the Atlantic panel to reach out for the Scripps decision is unnecessary in law, as well as incorrect in practice. It is unnecessary to reject (or accept) the Scripps decision in order to decide Atlantic. I shall discuss this further, but first I touch on the policy aspects, for the Atlantic panel tells us that it is public policy that has driven its action.
Policy Aspects
The Atlantic panel tells us that it is contrary to the public interest to allow the inventor of a novel and unobvious biological product, for which the structure is not objectively definable because it is not completely known, to claim the product without limitation to the process by which it was made. The panel believes that the inventor of such a complex chemical or biological composition must be limited to the same process by which the product was obtained and which is described by the inventor, lest the inventor and his/her patent get in the way of commercializers who might later use a different process to make the same product.
This policy action cuts to the heart of the patent system, raising questions of innovation incentive and fairness, as well as law, that were not mentioned in the briefs dealing with the foamed plastic innersole that is the subject of the Atlantic litigation. The panel’s view that it is contrary to the public interest to permit an inventor to have useful patent protection for a new product when the product can not be distinguished in the claim other than by reference to how it was made, is surely not an implementation of 35 U.S.C. § 101:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
The Court of Customs and Patent Appeals recognized that a new composition of matter may be of such complexity that in order to “particularly point out and distinctly claim” it, as required by 35 U.S.C. § 112, reference may be made to how it was obtained. In this way the statutory purpose of enabling the inventor of a new and unob-vious product to “obtain a patent therefor”, 35 U.S.C. § 101, can be implemented.
I need not belabor that the purpose of the United States patent law, to promote the progress of science and useful arts, is not served by eliminating useful patent protection for new technological products. When one reads that the cost of research and development of a new drug is well over two hundred million dollars, one must wonder at the public interest served now that the Federal Circuit has announced that a new product of biotechnology, if incapable of independent structural definition, may or may not have useful product patent protection, depending on which Federal Circuit panel is drawn.3
It has long been recognized that there are products for which novelty and unobvi-ousness can be established, but for which the structure of the product is not fully known. This mode of patenting such prod*1284ucts has existed for over a century. Such a product can be patented, like other new products, only when it meets the requirements of product patentability. Applying the same claim interpretation to patentability, validity, and infringement — an important contribution of the Federal Circuit to stability in the law — the claim is enforceable like other product claims. This follows from CCPA and Court of Claims precedent, and contravenes no law or precedent.
A product patent provides the standard opportunity to the inventor to develop the invention, to commercialize the product or license it to others, to publish the scientific results, and in general to go about the business of innovation. The patent law is intended to foster these activities, with accompanying benefits to the public. A policy change aimed at complex chemical and biological inventions, a change that may deprive such inventions of useful product patent protection, if enacted by judges instead of legislators, as a minimum should be done en banc.
II
PRECEDENT
The cases that the Atlantic panel states support its conclusion that Scripps was wrongly decided do not relate to the class of “true” product-by-process claim that was at issue in Scripps. Most of the cases that the Atlantic panel relies on do not even concern new products. Indeed, most courts have avoided the pitfall that has befallen our colleagues in Atlantic; for most courts, interpreting claims in accordance with the classical criteria of the specification, the prosecution history, and the prior art, have simply looked to see what the invention was, and interpreted the claims accordingly.
As the cases illustrate, claims that contain both product and process terms appear in an assortment of factual situations, of which the most common are:
(1) when the product is new and unobvious, but is not capable of independent definition;
(2) when the product is old or obvious, but the process is new;
(3) when the product is new and unobvious, but has a process-based limitation (e.g. a “molded" product).4
Type (2) includes the Atlantic class of claim; such claims are examined as process claims, their validity depends on the novelty and unobviousness of the process, and they are infringed only when the process is used. Type (1) is the Scripps class of claim; such claims are examined as product claims, their validity depends on the novelty and unobviousness of the product, and they are infringed by the product however made. Indeed, claims of types (2) and (3) are not properly called “product-by-process” claims, if that term is used with precision.
All of these classes, and variations thereof, have occurred in litigated cases. The form of claim that is used depends, of course, on the nature of the invention. In each litigated case the courts have applied the law that is appropriate to the facts of the particular invention. The CCPA, the Court of Claims, and the regional circuit courts have interpreted these various types of claims, in accordance with their substance. So has the Supreme Court. The courts have approached claim interpretation in the classical way: they have considered the specification, the prosecution history, and the prior art. The Atlantic panel has simply lumped all of these classes and claims and inventions into a one-rule-fits-all law, in a distressingly superficial treatment.
*1285I shall illustrate with the cases relied on by the Atlantic panel.
The CCPA and Court of Claims
The Court of Customs and Patent Appeals and the Court of Claims have both dealt with what the patent office calls, in the Manual of Patent Examining Procedure, a “product-by-process” claim. The need for such a claim was recognized at least as early as In re Painter, 1891 C.D. 200, 57 O.G. 999 (Comm’r of Pats. 1891):
It requires no argument to establish the proposition that as a rule a claim for an article of manufacture should not be defined by the process of producing that article. On the other hand, when a man has made an invention his right to a patent for it, or his right to a claim properly defining it, is not to be determined by the limitations of the English language. When the case arises that an article of manufacture is a new thing, a useful thing, and embodies invention, and that article cannot be properly defined and discriminated from prior art otherwise than by reference to the process of producing it, a case is presented which constitutes an exception to the rule.
1891 C.D. at 200-01 (emphasis original and added).
This holding recognized that a new and inventive article can be patented as an article, although it is defined “by reference to the process of producing it”. This is as important today as it was a century ago. The obligation that the patent statute imposes on the patent office and the courts is nicely met in this way, for it permits the “new and useful thing” that was invented to be claimed as such, despite the inadequacy of language or science to “properly define and discriminate” it. Painter formalized what is still known by scientists, despite the great advances that have been made: that there is more in heaven and earth than we know how to define in objective structural terms.
The CCPA respected the Painter principle. In In re Hughes, 496 F.2d 1216, 1218, 182 USPQ 106, 107-08 (CCPA 1974) the court remarked that “at least since 1891” it has been permissible to claim a new product by describing it with process terms, when the inventor
finds that his invention is incapable of description solely by structure or physical characteristics.
Id. at 1219, 182 USPQ at 108.
In In re Bridgeford, 357 F.2d 679, 149 USPQ 55 (CCPA 1966) the issue before the court was whether there was same-invention double patenting between product claims and product-by-process claims. The CCPA held that there was. The court explained that product-by-process claims are claims to the product unlimited by the process steps in the claim:
the invention so defined [by the product-by-process claim] is a product and not a process.
Id. at 682, 149 USPQ at 58 (emphases in original).
The Bridgeford court discussed the various ways in which a chemical product can be claimed:
Whether the invention be defined in terms of the structure of the compound, or its novel physical characteristics, or by defining it in terms of the process by which it is produced or in a proper case, by employing more than one of these methods of defining the invention, the right to a patent on the invention is the ultimate consideration, subject to conditions set forth in 35 U.S.C. 112. “In a proper case” does not mean that an applicant must prove the impossibility of defining the invention other than by using more than one of the above methods.
Id. at 682-83, 149 USPQ at 58. The court explicitly rejected the position that a product-by-process claim is simply a broader version of a process claim,5 clarifying an earlier decision by stating:
*1286While there is some language in [In re] Freeman [166 F.2d 178, 76 USPQ 585 (CCPA 1948) ] to support the contention that a product-by-process type claim differs only “in scope” from a process type claim and they therefore “are directed to a single invention,” (76 USPQ at 587) so far as this is inconsistent with our holding here it must be overruled.
Id. 357 F.2d at 683 n. 6, 149 USPQ at 58 n. 6.
In In re Brown, 459 F.2d 531, 173 USPQ 685 (CCPA 1972) the claims were for an oxygen-activated catalyst, and the claims included a description of the steps by which the catalyst was made. The court held that Brown had not shown that the product was patentable. The court explained that the product-by-process method of claiming a product is a
perfectly acceptable one so long as the claims particularly point out and distinctly claim the product,,
and that
it is the patentability of the product claimed and not of the recited process steps which must be established.
Id. at 535, 173 USPQ at 688 (emphases in original).
The court in Brown recognized the difficulties faced by the patent office in determining the patentability of a product that can not be described other than by the way it is made, and explained that there is a lowered burden on the examiner in showing prima facie obviousness for this class of claim:
It must be admitted, however, that the lack of physical description in a product-by-process claim makes determination of the patentability of the claim more difficult. ... We are therefore of the opinion that when the prior art discloses a product which reasonably appears to be identical with or only slightly different than a product claimed in a product-by-process claim, a rejection based alternatively on either section 102 or 103 of the statute is eminently fair and acceptable.
Id. at 535, 173 USPQ at 688. The court thus viewed it as “eminently fair”, under these circumstances, to shift to the applicant the burden of showing that the claimed and prior art products are patent-ably distinct.
In In re Pilkington, 411 F.2d 1345, 162 USPQ 145 (CCPA 1969) the court had explained that the question of patentability is whether the product itself is new and unob-vious, stating that a recitation of process steps, even of a new process, can not impart patentability to an otherwise old product:
“The addition of a method step in a product claim, which product is not patent-ably distinguishable from the prior art, cannot impart patentability to the old product.”
* * * * * *
Thus our inquiry here must be directed to a comparison of appellant’s sheet glass with the sheet glass of the references to determine whether any differences between them are such that the subject matter as a whole would be unob-vious.
Id. at 1348, 162 USPQ at 147 (citations omitted). The court quoted In re Dilnot, 300 F.2d 945, 950, 133 USPQ 289, 292 (CCPA 1962), a case in which the product was known, and thus the product claim, although it included a new method step, was held unpatentable.
The court in Pilkington again observed that when “conventional” product claiming is not available, the statute allows claiming the product in the best available way:
“By statute, 35 U.S.C. 112, Congress has placed no limitations on how an applicant claims his invention, so long as the specification concludes with claims which particularly point out and distinctly claim that invention.” With that in mind, it appears that applicant has distinctly pointed out and claimed what he regards as his invention in accordance with 35 U.S.C. 112.
Id. 411 F.2d at 1349-50, 162 USPQ at 148 (quoting In re Steppan, 394 F.2d 1013, 1019, 156 USPQ 143, 148 (CCPA 1968)) (emphasis in Steppan). The court concluded:
*1287[T]he differences between [applicant’s] glass and the glass of the prior art do not appear to us to be particularly susceptible to definition by the conventional recitation of properties or structure. Under the circumstances, it seems to us that the present product-by-process claim satisfies the requirements of 35 U.S.C. 112 and is appropriate here.
Id. 411 F.2d at 1349-50, 162 USPQ at 148 (footnote omitted). The court pointed out that the product must be independently patentable, and that use of the product-by-process form to describe the product satisfied the requirements of section 112.
In In re Fessmann, 489 F.2d 742, 180 USPQ 324 (CCPA 1974) the court again explained the use of product-by-process claims when the invented product is not susceptible of independent definition. The court referred to the extra burden that is placed on the applicant to establish patenta-bility of such a product, and stated:
In Brown, the court was in effect saying that the Patent Office bears a lesser burden of proof in making out a case of prima facie obviousness for product-by-process claims because of their peculiar nature than would be the case when a product is claimed in the more conventional fashion.
Id. at 744, 180 USPQ at 326. The court held that the appellant had not overcome the examiner’s prima facie case of obviousness based on similar prior art compounds, and therefore that the product-by-process claims were unpatentable, whether or not the process was new.
In re Hirao, 535 F.2d 67, 190 USPQ 15 (CCPA 1976) is relied on by the Atlantic panel. The claim in Hirao was:
1. A process for preparing foods and drinks sweetened mildly, and protected against discoloration, Strecker’s reaction, and moisture absorption, which comprises:
[1] adding a 1,6-glueosidase and b-amylase, under such conditions and in a quantity sufficient to produce straight chain amylose ...
[2] subjecting the resulting amylose to the action of b-amylase and purifying and drying to obtain high purity maltose in crystalline powder form ...
[3]adding said high purity crystalline maltose powder to foods and drinks as the essential added sweetener.
Id. at 67-68, 190 USPQ at 16. The Solicitor is stated to have argued, creatively, that the third step was analogous to a “product-by-process” claim, that the product used in the third step was obvious in light of the first two steps, and that “some courts in infringement suits” limit “product-by-process claims to the process used”. The CCPA remarked that:
This court ... does not construe product-by-process claims in an ex parte case as limited to the product formed by the specific process recited.
Id. at 69 n. 3, 190 USPQ at 17 n. 3.
The claim in Hirao is a plain process claim. Although this case is far removed from the Scripps class of claim and the premises on which it rests, in response to the Solicitor’s argument the court said:
A product-by-process claim, although reciting the subject matter of the claim in terms of how it is made, is still a product claim.
Id. at 69, 190 USPQ at 17.
In re Luck, 476 F.2d 650, 177 USPQ 523 (CCPA 1973) is also relied on by the Atlantic panel. In Luck the applicant claimed a coated lamp bulb wherein the claim included a description of how the coating was applied:
1. A hollow light-transmitting lamp-bulb-shaped glass member adapted to surround a source of radiations, a coating carried on the external surface of said glass member, said coating comprising a mixture of:
(a) a polymer consisting essentially of polymethyl-methacrylate having a tack point ...
(b) ... an organofunctional silane ...
(c) ... an additive organic substance
(d) said coating having been affixed to said glass member by applying thereon a liquid organic solvent having dissolved therein said polymer, said organofunc-*1288tional silane and said additive organic substance, and said coated glass member thereafter being baked.
Id. at 651, 177 USPQ at 524. The CCPA explained that process steps may be included in a claim in order to define the product, and to aid in distinguishing the product from the prior art:
As for the method of application, it is well established that product claims may include process steps to wholly or partially define the claimed product. See In re Brown, and the cases cited therein. To the extent these process limitations distinguish the product over the prior art, they must be given the same consideration as traditional product characteristics.
Id. at 653, 177 USPQ at 525 (emphasis in original). The court held that the process terms did not distinguish the product from the prior art. It is not easy to see how Luck purportedly supports the Atlantic thesis. Indeed the CCPA, while holding that Luck’s invention was unpatentable based on prior art, reaffirmed that
it is well established that product claims may include process steps to wholly or partially define the claimed product.
Id.
The Atlantic panel also cited In re Avery, 518 F.2d 1228, 186 USPQ 161 (CCPA 1975), wherein a claim at issue was:
23. The magnesium product of the me-tallothermic process of claim 1, characterized in composition by a content of metallic impurities of less than about 1500 ppm, and of silicon of between about 50 and 300 ppm.
Id. at 1230, 186 USPQ at 162. The product was found to be old, although Avery argued that his process was different from the prior art process, and therefore that the claimed product was new and unobvious. The CCPA held that “[t]he claimed product is completely disclosed in the prior art,” stating that an unpatentable product is not saved by a process limitation in the claim. Id. at 1234, 186 USPQ at 165-66.
It is indeed hard to see how this relates to the Scripps type of product-by-process claim, where the product was found to be new and unobvious, and was patentable independent of the process.
The Manual of Patent Examining Procedure states the special burden of proof for examining product-by-process claims:
M.P.E.P. § 706.03(e) When the prior art discloses a product which reasonably appears to be either identical with or only slightly different than a product claimed in a product-by-process claim, a rejection based alternatively on either section 102 or 103 of the statute is appropriate. As a practical matter, the Patent and Trademark Office is not equipped to manufacture products by the myriad of processes put before it and then obtain prior art products and make physical comparisons therewith. A lesser burden of proof is required to make out a case of prima facie obviousness for product-by-process claims because of their peculiar nature than when a product is claimed in the conventional fashion.
M.P.E.P.Rev. 6, Oct. 1987 (citations omitted). Although commentator Donald Chisum states that a 1974 revision of the M.P.E.P. “squarely rejects the Painter necessity rule”, Donald S. Chisum, Patents § 8.05[3], at 8-91 (1991), the logic of this conclusion is obscure, particularly because Chisum also observes that two months later the CCPA in Hughes reaffirmed the Painter rule. Id. at 8-92.
The Painter rule is a practical solution to an important problem, whereby law and practice are adapted to complex products in order to implement the purposes of the patent law. This rule continues to serve the purposes of the law.6
The Court of Claims routinely recognized that product-by-process claims were product claims, unlimited by how the product was made. In Tri-Wall Containers, Inc. *1289v. United States, 408 F.2d 748, 750-51, 187 Ct.Cl. 326, 161 USPQ 116, 118-19, cert. denied, 396 U.S. 828, 90 S.Ct. 78, 24 L.Ed.2d 79 (1969), the court applied the on-sale bar to the product, although the product that had been on sale was made by a different process from that stated in the claim. The court stated that “As the findings of fact show, the prior art product and the claimed product are structurally identical”, and explained:
It is well established that a product claimed as made by a new process is not patentable unless the product itself is new. The Wood-Paper Patent, 90 U.S. (23 Wall.) 566, 596 [23 L.Ed. 31] (1874), Cochrane v. Badische Anilin & Soda Fabrik, 111 U.S. 293, 311 [4 S.Ct. 455, 464, 28 L.Ed. 433] (1984).... the addition of a method step in a product claim, which product is not patentably distinguishable from the prior art, cannot impart patentability to the old product.
Id. (some citations omitted). The Court of Claims cited, supra, two Supreme Court cases which held that the product itself must be new in order to patent it as a product per se. These cases appear to be misconstrued by the Atlantic panel, as I shall discuss, as holding that a new product can never be patented as a product per se if it is defined in process terms.
The Federal Circuit has similarly treated this class of claim. In In re Marosi, 710 F.2d 799, 801, 218 USPQ 289, 291 (Fed.Cir.1983), some claims were directed to “a zeol-ite manufactured by the claimed process.” The court, although reversing the Board’s rejection of the process claims over the prior art, affirmed the rejection of the product claims because no unobvious difference was shown between the claimed product and the prior art product:
Where a product-by-process claim is rejected over a prior art product that appears to be identical, although produced by a different process, the burden is upon the applicants to come forward with evidence establishing an unobvious difference between the claimed product and the prior art product....
Id. at 803, 218 USPQ at 292-93 (citations omitted).
In Scripps the Federal Circuit continued to apply the established law of the CCPA and the Court of Claims, in accordance with the mandate of South Corp. As observed supra at n. 3, the issue of infringement of these Scripps claims was remanded for trial. This court expressed no opinion on the merits of that issue, on which the factual record was “undeveloped”.7
The Supreme Court Decisions
Although the Supreme Court has not dealt with any case having the Painter/Bridgeford/Brown/Fessmann/Scripps class of claim, the Court’s decisions show its understanding that claims are interpreted in accordance with the nature of the claimed invention, as elucidated by the specification, the prior art, and the prosecution history. The Court’s group of 1870’s and 1880’s cases, on which the Atlantic panel places so much reliance, are in full accord with this mode of claim interpretation, despite the old-fashioned style of claiming.
Consider, for example, Merrill v. Yeomans, 94 U.S. (4 Otto) 569, 24 L.Ed. 235 (1877), relied on by the Atlantic panel. The issue before the Supreme Court was the “correct construction of plaintiff’s patent.” Id. 94 U.S. (4 Otto) at 569. The claim was for
the above-described new manufacture of hydrocarbon oils, ... by treating them substantially as hereinbefore described.
Id. at 571. The Court, examining the specification, found that the product was an old oil made by a new process. The Court concluded that the invention was only in the process, and that the most that the patentee could claim was this oil made by this particular process. The Court stated *1290that its obligation is to uphold “that which was really invented”:
[WJhere it appears that a valuable invention has really been made, this court, giving full effect to all that is found in the application on which the Patent Office acted, will uphold that which was really invented, and which comes within any fair interpretation of the patentee’s assertion of claim.
Id. at 573. It is hard to see how this decision, where the Court found the product old and interpreted the claim accordingly, affects the Scripps decision or negates the policy implemented in Painter.
Let us next look at the claim in Goodyear Dental Vulcanite Co. v. Davis, 102 U.S. (12 Otto) 222, 26 L.Ed. 149 (1880). That claim was written in the same then-standard style, a style long since discontinued:
The plate of hard rubber or vulcanite, or its equivalent, for holding artificial teeth, or teeth and gums, substantially as described.
Id. 102 U.S. (12 Otto) at 223. The Atlantic panel describes this claim as “not drafted in contemporary product-by-process terms, but instead incorporated the specification.” The Atlantic panel states that this claim form added the process limitations of the specification, turning this claim into a “product-by-process” claim of the Scripps variety. The Atlantic panel bases its strong criticism of Scripps on this reading.
In fact, the Court in Goodyear simply considered the specification and the prior art and decided what it was that Goodyear had invented. In earlier litigation of the same patent, Smith v. Goodyear Dental Vulcanite Co., 93 U.S. (3 Otto) 486, 23 L.Ed. 952 (1876), the Court had held that it was the process described in the specification, of mounting artificial teeth in vulcanized rubber, which distinguished Goodyear’s invention over the prior art, since a plate for holding artificial teeth was previously known. Id. 93 U.S. (3 Otto) at 493. Thus the Court held that since the accused infringer did not use Goodyear’s process to make a celluloid dental plate, no infringement could lie.
The “substantially as described” claim in the Goodyear cases is not comparable to the product-by-process claim of Scripps, by any stretch of its reading. The words “substantially as described” did not mean that a claimed product must be limited to the process in the specification. The Atlantic panel states:
Under the old practice for claims with the phrase “substantially as described”, the Supreme Court treated the specification as part of the claim. Thus, the specification added process limitations to the claimed product.
That is incorrect. See, e.g., II W.C. Robinson, Robinson on Patents (1890), at § 750:
The phrase “substantially as described” in the claim, signifies much or little, as the patentee evidently designed. This phrase is always supposed to be incorporated in the claim, whether or not it is actually inserted.
Robinson, at § 517, states:
In stating Claims certain phrases are frequently employed to which a special importance seems to be attached by applicants. Among these are the phrase “substantially as described” and others of the same meaning. These phrases import the same thing when used in a Claim as when elsewhere employed. They are neither necessary nor technical. The reference they make to the Description is always implied, and relates only to the essential features of the invention as therein delineated. They add nothing, therefore, to the certainty of the Claim, nor do they detract from it unless the claimant carelessly inserts them as a substitute for a more clear and definite statement of his invention.
(Footnotes omitted).
Goodyear illustrates the classical claim interpretation wherein the Court ascertained what the invention was, and determined infringement accordingly. The Goodyear cases do not require that a novel and unobvious product such as in Scripps, when claimed in the product-by-process form under the Rule of Necessity, must ipso facto be limited to the process terms used to identify the product. The broad *1291language of the Court in Goodyear does not reach forward to prejudge this different case.
The Atlantic panel relies similarly on Cochrane v. Badische Anilin & Soda Fabrik, 111 U.S. 293, 4 S.Ct. 455, 28 L.Ed. 433 (1884). The claim was:
Artificial alizarine, produced by either of the methods herein described, or by any other method which will produce a like result.
The Court stated:
It was an old article. While a new process for producing it was patentable, the product itself could not be patented.
Id. at 311, 4 S.Ct. at 464. The Court also found that the accused alizarine was made by a process described in the prior art, not by the process described in the patent, and that the products were different. The Court simply treated the process of making the alizarine as an evidentiary matter, stating:
[Ujnless it is shown that the process of [the specification] was followed to produce the defendants’ article, or unless it is shown that the article could not be produced by any other process, the defendants’ article cannot be identified as the product of the process of [the specification].
Id. at 310, 4 S.Ct. at 464.
Further, the claim used the phrase “herein described” instead of “substantially as described”, a difference to which Robinson appears to attach significance:
When used, the word “described” refers to the Claim as well as the Description, and either may thus aid in the interpretation of the other. Whether expressed or implied, they limit the general terms of the Claim to the specific features set out in the Description, though neither these phrases, nor the one “as specified”, confines the Claim to that mode of using the invention which has been previously described. The phrase “substantially as and for the purposes set forth” is of a different character. These are words of limitation, and should not be used unless they are intended to have their special signification. They refer to the Description for a qualification of the general statements of the Claim, and confine the invention within the purposes and operations there specifically named.
Robinson on Patents at § 517 (footnotes omitted).
The Atlantic panel, having described the claim in Cochrane v. BASF as a product-by-process claim, presumably of the Scripps mode, states that Cochrane v. BASF established the legal rule that a product-by-process claim is always limited to the process described in the specification. The panel applies the broad statement made by the Court to whatever the facts of the future case. The Court in Cochrane v. BASF decided the case on its facts, holding that since the claimed product was old, and the accused product was made by a process that was in the prior art, it could not be an infringement. This holding did not make meaningless the ensuing century of patentability decisions that adapted the law to the progress of technological invention. Cochrane v. BASF is not fairly read to prejudge the question confronted in Painter, and to restrict to the capabilities of 1884 the march of science and the evolution of the common law of patents.
It is perhaps helpful to remember that the claims of that period did not set the periphery of the patent, as today, but simply stated the core of the invention. See A.W. Deller, 1 Patent Claims, § 25 at 44-45 (2d ed. 1971):
In the old days of central definition, the use of the above [“substantially as described”] or similar words was customary. Under the present peripheral system, they are practically never employed.
The courts decided whether there was infringement, in light of what had really been invented.
The holding of Plummer v. Sargent, 120 U.S. 442, 7 S.Ct. 640, 30 L.Ed. 737 (1887), also relied on by the Atlantic panel, again illustrates this judicial approach. The Plummer product was claimed as a new manufacture, a bronzed iron, “substantially as described”:
*1292What I claim ... is the new manufacture hereinabove described, consisting of iron ornamented in imitation of bronze by the application of oil and heat, substantially as described.
Id. at 445, 7 S.Ct. at 642. The prior art described iron that was bronzed by the “application of oil and heat”, the words in the claim. The circuit court had found non-infringement because the defendant had simply followed the process of the prior art. Id. at 448, 7 S.Ct. at 643. The Supreme Court agreed, in a straightforward claim interpretation in view of the prior art:
It seems necessarily to follow from this view either that the Tucker patents are void by reason of anticipation practiced by [the prior art], or that the patented process and product must be restricted to exactly what is described....
Id. at 449, 7 S.Ct. at 644. The Court thus limited infringement to practice of the same process described by the patentee, not because of any per se rule, but to sustain the validity of the patents in view of the prior art. The accused product, in turn, could not be independently distinguished without showing how it was made. Although this case was also selected by the Atlantic panel as directing the result in Scripps, it does not take to so expansive a reading.
General Electric Co. v. Wabash Co., 304 U.S. 364, 58 S.Ct. 899, 82 L.Ed. 1402, 37 USPQ 467 (1938) is the last of the Supreme Court cases relied on by the Atlantic panel. This case, although more recent than the others, did not involve product-by-process claims. In General Electric a typical claim was:
25. A filament for electric incandescent lamps or other devices, composed substantially of tungsten and made up mainly of a number of comparatively large grains of such size and contour as to prevent substantial sagging and offsetting during a normal or commercially useful life for such a lamp or other device.
Id. at 368, 58 S.Ct. at 901, 37 USPQ at 468. The Court held the claim “invalid on its face” under 35 U.S.C. § 33 (the predecessor of 35 U.S.C. § 112) for failing to provide a “distinct and definite statement of what he claims to be new, and to be his invention.” Id. at 369, 58 S.Ct. at 901, 37 USPQ at 468. Specifically, the Court found the description of the grains as “of such size and contour as to prevent substantial sagging and offsetting” to be “inadequate as a description of the structural characteristics of the grains.” Id. at 370, 58 S.Ct. at 902, 37 USPQ at 469. The Court stated that the claims could not be saved by reference to the specification because the specification too was inadequate:
Even assuming that definiteness may be imparted to the product claim by that part of the specification which purportedly details only a method of making the product, the description of the [patented] process is likewise silent as to the nature of the filament product.
Id. at 373, 58 S.Ct. at 903, 37 USPQ at 470 (footnote omitted). The Court held that the product as claimed was not distinguished from the prior art, pointing out that tungsten filaments of large grains were known in the prior art, and that neither the product nor the process was sufficiently described in the specification. The Court also criticized the use of functional terms, stating that:
It may be doubted whether one who discovers or invents a product he knows to be new will ever find it impossible to describe some aspect of its novelty.
Id. at 373, 58 S.Ct. at 903, 37 USPQ at 470. Perhaps. I remark, however, that whether the description of “some aspect” will satisfy section 112 can not have been prejudged for all technologies in General Electric. It is in this context that the Court made the statement quoted by the Atlantic panel:
[A] patentee who does not distinguish his product from what is old except by reference, express or constructive, to the process by which he produced it, cannot secure a monopoly on the product by whatever means produced.
Id. This statement is not inapt on the facts of General Electric. But it can not be deemed to have decided all claims, such as the product-by-process claims authorized in *1293Painter, and the situation illustrated in Scripps where the product was, without dispute, “distinguished from what was old”. Indeed, it would eliminate all functional claiming, and a good deal more not contemplated by the General Electric court.
None of the Supreme Court cases dealt with claims of the type at issue in Painter or Scripps or Bridgeford or Brown, or considered the problems that confronted these tribunals, or applied the patent law to those problems. Modern technology may be capable of providing proofs not available a century ago, or half a century ago. In Scripps, for example, the new product was described as having a “fingerprint” obtained by gel electrophoresis, 927 F.2d at 1569 n. 3, 18 USPQ2d at 1004 n. 3, although the structure was not fully known. It is curious indeed to see the Atlantic panel reading Cochrane v. BASF and General Electric as eliminating for all time the possibility of proving infringement of complex molecules. The decision in Painter, only seven years after the decision in Cochrane v. BASF, could not have interpreted BASF as negating the very type of claim that was recognized in Painter. Neither the Painter rule nor its century of practical life could have ensued on the reading given to these cases by the Atlantic panel. These Supreme Court cases, of another era of technology, did not decide the question before us, or freeze the tracks of the common law into the mold of 1884.
The Circuit Courts
Although the regional circuit court decisions are not precedent in the Federal Circuit, I point out that they do not support the one-rule-fits-all approach of the Atlantic panel. The results reached in the regional circuits, before formation of the Federal Circuit, vary as broadly as do the facts on which they were based. Again, I will limit my review to cases cited by the Atlantic panel, for my purpose in this dissent is not to decide what is the correct law for Painter/Scripps situations, but to explain why the question should have been taken en banc.
The regional circuits have considered many forms of process-limited claims, for they come in many forms. The circuits did not hold that the inventor of a new and unobvious complex product, who identified the product with reference to the process by which it was made, could never protect the product per se. The regional courts simply interpreted the claims on the facts of each case, as had the Supreme Court, the CCPA, and the Court of Claims.
I start with Paeco, Inc. v. Applied Moldings, Inc., 562 F.2d 870, 194 USPQ 353 (3d Cir.1977), for this case is cited in Atlantic as holding that a claim containing process limitations can not be infringed unless “the claimed process or its equivalent” is used. (The Atlantic panel continues to call all claims with process terms “product-by-process” claims, thus begging the question.) It is instructive to look at the claim in Paeco. The issue was the meaning of the emphasized word “volume ” in the following claim:
2. As an article of manufacture, replica wooden beams having the feel, appearance and stain characteristics of wood fabricated of rigid urethane foam molded in a mold having a volume less than the free rise volume of said urethane foam resin, the outer skin surface thereof being non-porous and dense, and a portion of said skin being removed to define an exposed porous surface whereby the easy mounting of said beam is facilitated, said beam stained to provide colora-tions corresponding to a wooden beam.
Id. at 872, 194 USPQ at 355. The question before the court was invalidity based on anticipation, which in turn depended on whether the “volume” referred to the foam or to the mold. Stating that “because the claim is capable of two interpretations, we are justified in looking behind the words of the claim to the specification”, id. at 876, 194 USPQ at 358, the court concluded that the volume referred to the mold, not the foam.
The court, having resolved its “suspicion that the words may describe the process, not the product”, held that the claim could not be infringed unless the foam were *1294made in a mold having the volume stated in the claim. The Paeco court, yielding to the temptation of appellate tribunals, generalized this straightforward claim interpretation into the dictum that “A patent granted on a product claim describing one process grants no monopoly as to identical products manufactured by a different process.” Id. Thus the holding of Paeco, based on its particular facts, was converted into the generalized pronouncement that appears in 3 E.B. Lipscomb, Lipscomb’s Walker on Patents, § 11.19 (3d ed. 1985), at p. 438, cited by the Atlantic panel. Lipscomb relies on Paeco for his broad statement that the same process must always be used, whatever the circumstances, neglecting to distinguish among new and known products and other significant claim and invention differences. See n. 3, supra, and accompanying text.
The other cases relied on by the Atlantic panel are of much older vintage than Pae-co. In National Carbon Co. v. Western Shade Cloth Co., 93 F.2d 94, 97, 35 USPQ 404, 407 (7th Cir.1937), cert. denied, 304 U.S. 570, 58 S.Ct. 1039, 82 L.Ed. 1535 (1938) the claims were to a carbon product (used in gas masks). The product was claimed solely in terms of functional limitations, viz., that the carbon had a retentive value of not less than 20 per cent and a life service of 225 minutes. The court invalidated the claims because they were phrased solely in functional terms:
[T]he attempt to broaden product claims by describing the product exclusively in terms of its use or function is subject to the same vice as is the attempt to describe a patentable device or machine in terms of its function.
Id. 93 F.2d at 98, 35 USPQ at 407. The court observed that it was only when the stated process was used that a carbon product having the claimed function was obtained, and held that a product can never be described solely in terms of its function.
In Buono v. Yankee Maid Dress Corp., 77 F.2d 274, 26 USPQ 57 (2d Cir.1935) the claimed product was a type of stitch produced by the disclosed sewing machine, but “the claims cover[ed] the product produced by other machines and processes”. Id. at 279, 26 USPQ at 61. The court concluded that the claimed stitch was not new, and held the patent invalid. Judge Learned Hand, writing for the court, stated:
Conceivably it might be possible to patent a product merely as the product of a machine or process, even though it were anticipated if made in other ways.... it would in that case not be infringed by anything but the product of the machine or of the process, ...
Id. The court thus speculated on the pat-entability of claims to old products (those “anticipated if made in other ways”) if limited to the new machine or process by which it was made. Buono does not deal with new products, even in dictum. Although the Atlantic panel makes much of this case, one must wonder at its relevance. I would indeed be interested in knowing Judge Hand’s view of the patentability of new products of chemical or biological structure, when the structure is not fully describable but the product has been proved to be new and unobvious. This case, however, does not offer us that view.
The Buono reasoning, as that case was decided, is consistent with that of Downes v. Teter-Heany Development Co., 150 F. 122, 123 (3d Cir.), cert. denied, 205 U.S. 543, 27 S.Ct. 790, 51 L.Ed. 922 (1907), also relied on by the Atlantic panel. Typical product claims were:
8. As an article of manufacture, an electrical conductor having a compact, uniform covering of asbestos fiber, wound spirally thereon, matted, compressed and waterproofed, substantially as described.
9. As an article of manufacture, an electrical conductor having a thin, compact, uniform covering of asbestos fiber wound thereon and attached by adhesive substance, the fibers being matted and compressed into a coherent fabric, substantially as described.
Id. 150 F. at 123. The Downes court was of the view that “the process and the product are substantially one invention or discovery,” having examined the prior art, which it found
*1295included the following steps: (1) coating the wire with a suitable adhesive gum; (2) winding the wire with an asbestos yarn or thread; (3) compressing or compacting the covering about the wire; and (4) applying a suitable water-proof coating or shield.
Id. at 124. The court found that what distinguished the invention over the prior art was an additional (fifth) step of brushing the asbestos yarn so as to smooth out “lumps and inequalities”. The Court reasoned that this fifth step was the only new feature, and since the defendant did not employ this step, the defendant was simply practicing the prior art.
The Downes court did not hold that no product claim can ever be infringed when the identical product is made by a different process from that described in the specification. The court did hold that the product was different because the process was different, and that the process used by the defendant was in the prior art. This case has nothing to do with the Painter or Bridgeford or Brown or Scripps type of product claim, or any of the issues raised in those cases.
A brushing step was again involved in United States Gypsum Co. v. Consolidated Expanded Metal Cos., 130 F.2d 888, 55 USPQ 247 (6th Cir.1942), cert. denied, 317 U.S. 698, 63 S.Ct. 441, 87 L.Ed. 558 (1943), where the court simply held the claim invalid for want of “novelty”. The product claim was:
16. Metallic fabric of the slit and expanded type having strands and strand interconnecting bonds forming mesh work openings having relatively sharp corners at points where strands and bonds are interconnected, said fabric having the edges and corners throughout the entire expanded metal body brushed smooth.
Id. 130 F.2d at 897 n. 3, 55 USPQ at 251 n. 3. The court found no novelty in a brushed smooth product where a smooth product was known in the art, and held the claim invalid for lack of invention. The court referred to
the rule that a patentee who does not distinguish his product from what is old except by reference, express or constructive, to the process by which he produced it, cannot secure a monopoly on the product by whatever means produced.
Id. at 893, 55 USPQ at 252. Although expansive when read out of context, the court simply stated the rule that an old product can not be patented. This is not another all-purpose ruling that “the process limits a product-by-process claim”, as the Atlantic panel states. Indeed, the claim in Gypsum was not a product-by-process claim.
Parke, Davis & Co. v. American Cyanamid Co., 207 F.2d 571, 99 USPQ 237 (6th Cir.1953) also does not provide the asserted support for the Atlantic panel’s rule. The claim was to a novel chemical, one constituent of which was obtained from mammalian liver tissue:
1. A compound of the class consisting of an organic acid, its salts and its esters, said acid being the acid derived by auto-lysis of mammalian liver tissue and being free from pantothenic acid....
Id. at 571 n. 1, 99 USPQ at 238 n. 1. The examiner had required that the claim be so limited during prosecution, and the applicant acquiesced. The court stated the issue as one of claim construction:
The controlling question is whether the claims are so limited by the specification, the file wrapper proceedings and their own language that they do not cover appellee’s product.
Id. at 573, 99 USPQ at 239. The court held that actions during prosecution estopped the patentee from interpreting the claim to include non-mammalian liver tissue. Id. at 573-74, 99 USPQ at 240. Although the issue is treated by the Atlantic panel as interpretation of a product-by-process claim, it bears a closer resemblance to a question of the doctrine of equivalents. Surely this ruling created no all-purpose rule, as the Atlantic panel states, that would negate treating the Painter/Scripps form of claim as a claim to the product per se. Indeed, the Painter-type issue does not appear in this case.
*1296In Hide-Ite Leather Co. v. Fiber Products Co., 226 F. 34 (1st Cir.1915), the claim was:
A waterproof leatherboard made from pulp containing disintegrated fibers of tanned leather, and having insoluble waterproofing compounds deposited by precipitation upon and thereby intimately mixed with the fibers of which it is composed, substantially as described.
Id. at 35. The court said, in the portion quoted by the Atlantic panel:
It is also a well-recognized rule that, although a product has definite characteristics by which it may be identified apart from the process, still, if in a claim for the product it is not so described, but is set forth in the terms of the process, nothing can be held to infringe the claim which is not made by the process.
Id. at 36. The Hide-Ite Leather court thus held that when a product may be identified apart from the process, but was deliberately claimed only in terms of the process, the patentee voluntarily limited the claim to the process. Invoking estoppel to assert otherwise, based on the patent prosecution history, the court stated that such a claim would not be infringed if the claimed process were not used because that “was the understanding of the Patent Office in allowance, [as] is fully shown by the file wrapper”. Id. at 37. This case of estoppel is entirely unrelated to the class of claim for which patentability is as a new product, independent of the process, and with no estoppel in the prosecution history.
Although not cited by the Atlantic panel, I mention Pfizer, Inc. v. International Rectifier Corp., 538 F.2d 180, 190 USPQ 273 (8th Cir.1976), wherein the court noted that:
A “product-by-process claim” is a form of product patent for a compound that cannot be described other than by reference to the process of making it.
Id. at 188 n. 15, 190 USPQ at 280 n. 15 (emphasis in original). Also Ralston Purina Co. v. Far-Mar-Co., Inc., 586 F.Supp. 1176, 222 USPQ 863 (D.Kan.1984), aff'd in part & rev’d in part on other grds., 772 F.2d 1570, 227 USPQ 177 (Fed.Cir.1985):
A claim for a product defined by the process of making the product, or product-by-process claim, is proper and not indefinite when the product is not fairly susceptible to description by its properties or structure, [citing In re Pilkington]. Beyond this, product-by-process claims are proper even if the product is capable of description in an allowable product claim, if the product is incapable of description by product claims which are of a different scope, [citing In re Hughes ].
586 F.Supp. at 1219, 222 USPQ at 896. Also Fromson v. Advance Offset Plate, Inc., 720 F.2d 1565, 1570, 219 USPQ 1137, 1141 (Fed.Cir.1983) (“That a process limitation appears in a claim does not convert it to a product by process claim.”)
The cases decided by the regional circuits illustrate the great variety of forms of patent claims. Viewed together these decisions negate, rather than support, the Atlantic panel’s thesis that all claims that contain process terms, whatever the novelty of the product or the process and whatever the facts of the case, must be read in a single way. Indeed, the cases cited by the Atlantic panel reinforce the rule that the interpretation of claims depends on the particular invention, in light of the specification and prosecution history and prior art for the specific case.
Ill
The Federal Circuit
One of the fruits of formation of the Federal Circuit is that the same law is now applied to issues of patentability, validity, and infringement. This new consistency has contributed usefully to the stability of the law.
For example, the Federal Circuit has held that claims, construed for the purpose of determining validity, must be construed the same way for the purpose of determining infringement. See, e.g., W.L. Gore & Associates, Inc. v. Garlock, Inc., 842 F.2d 1275, 1279, 6 USPQ2d 1277, 1280 (Fed.Cir.1988). If the identical product in the prior art anticipates the product-by-process claim, *1297then the identical product will infringe the claim. The Atlantic panel, holding that the identical product may not infringe, departs from this law. The Atlantic panel’s relegation of CCPA precedent to patent office examination, holding that this precedent is to be ignored in validity and infringement determinations, is a dramatic retrogression, reinstating the “nose of wax” of patent claims, criticized in White v. Dunbar, 119 U.S. 47, 51, 7 S.Ct. 72, 74, 30 L.Ed. 803 (1886).
The Painter class of product-by-process claim arose not to enlarge the scope of an applicant’s invention, but to allow its claiming commensurate with the applicant’s invention. When the invention is a new product, this form allows its patenting as a new product. The law of the CCPA and the Court of Claims, the precedent that binds us, is that the class of product-by-process claim that was recognized in Painter, Bridgeford, Pilkington, Brown, Fess-mann, and Tri-Wall Containers, is a product claim. These cases established that for this class of invention the process terms serve as a mechanism for identifying the product in order to comply with 35 U.S.C. § 112. These courts held, as a matter of the law of claim interpretation, that such products must meet the requirements of product patentability, independent of the way the product was made.8 Thus the Federal Circuit, applying the rule that pat-entability, validity, and infringement are governed by the same law of claim interpretation, held in Scripps that when the requirements of patentability as a product are met, such claims are enforceable independent of the process by which the product is made.
Most if not all of the cases cited by the commentators who are quoted by the Atlantic panel do not involve the class of claim of Painter and Bridgeford and Brown and Tri-Wall Containers and today typified by Scripps, for most of these cases relate to known or unpatentable products wherein, at most, the process was novel. See, e.g., Cochrane v. BASF, the Goodyear cases, etc.9 The facts of relatively few cases relate to the class of product-by-process claim at issue in Scripps. Indeed, in the Atlantic opinion, as well as the view of some commentators, one observes a collection of dicta lifted out of context, until a new structure has been built on the most tenuous of supports.
Not all commentators are so flawed. In Product-by-Process Claims: A Practical Perspective, 68 J.Pat. & Trademark Off. Soc’y at 41, Mirabel states, after lengthy analysis, that:
the conclusion of the commentators that “[t]he courts have uniformly held that only a product produced by the claim-designated process may be held to infringe” is not supported by the decisions cited.
In 2 Patent Law Perspectives, § 2.6[10] (1992), at 2-796.38—2-796.39, M.J. Adelman states:
Theoretically, the patentability of product patents resides in the discovery of attributes of the product rather than in its method of manufacture, although to the extent that the patentability of a product can be predicated on a nonob-vious method of manufacture, this need not be the case. Of course, if the claim is limited to products made by the defined process, then product-by-process claims serve principally to solve problems such as use of the process abroad to make a product that is then imported into the United States. However, if the inventor makes a new product and is unable to define a unique footprint for that *1298product, but does discover a utility for the product that would support the pat-entability of the product whatever its method of manufacture, then to further the policy of granting product patents to new and nonobvious products such claims should cover the product however made. Thus, limiting product-by-process claims to products made by the identified process seems strict if it can be shown that no other way was available at the time of grant to define the product.
(Emphasis added).10
All of the forms of claim that are illustrated in the cases serve a purpose. It is incorrect to apply the law of one form to the factual situation of the other. Just as the factual situation in Scripps placed it in the context of the Painter, Bridgeford, Brown, Tri-Wall Container, and other precedent governing this specific type of claim, so the factual situation in Atlantic places it in the class where the patentee is estopped to state that the product is claimed independent of the process.
The experience of technological innovation, and today’s dramatic discoveries involving biological products, show the continuing role of the Painter class of claim. It is inappropriate for a panel of this court, on the basis of a shoe innersole, to decide that the discoverer of a novel and unobvious biological product whose structure is not objectively known can not claim that product per se. Should this court now wish to restrict patent protection for inventions of complex biological products that are necessarily identified by how they were made, it should be done en banc so that all who are concerned with such inventions will know the law that will be applied. Our court was formed to stabilize, not destabilize, the law.11
“En banc courts are the exception, not the rule.” United States v. American-Foreign Steamship Corp., 363 U.S. 685, 689, 80 S.Ct. 1336, 1339, 4 L.Ed.2d 1491 (1960). “They are convened only when extraordinary circumstances exist that call for authoritative consideration and decision by those charged with the administration and development of the law of the circuit.” Id.
If not now, when?
. In South Corporation v. United States, 690 F.2d 1368, 215 USPQ 657 (Fed.Cir.1982) (en banc), the Federal Circuit adopted as precedent the law of the Court of Customs and Patent Appeals and the Court of Claims.
. Counsel for Atlantic argued to the PTO:
Such product-by-process claims are process claims with patentable process limitations and would not conflict with any divisional claims containing patentable, structural and compositional limitations. The patentable process limitations would not be acceptable in the divisional application claims.
Joint Appendix at 3066 (emphases in original).
. In matters of biotechnology the so-called "public” interest is heavily intertwined with commercial interests. This is not an abstract point. Commercial interests are illustrated in Scripps, where the inventor had obtained the patented product by complex techniques that consumed several years, while the accused infringer was alleged to have "cloned” the Scripps product for commercial production. That dispute was remanded to the district court, and is sub judice. The Atlantic panel's collateral attack on the law of a pending case, while declining to allow en banc review, does not enhance confidence in the judicial process.
. In E.P. Mirabel, Product-by-Process Claims: A Practical Perspective, 68 J.Pat. & Trademark Off. Soc'y 3 (1986), the cases are divided into three categories which he defines as follows:
(1)where the product is "defined by setting forth the process by which it is made. It is important to bear in mind that 'the invention so defined is a product and not a process’ ”, quoting In re Bridgeford, 357 F.2d 679, 682, 149 USPQ 55, 58 (CCPA 1966);
(2) "product claims with a process limitation”; and
(3) cases where “a process limitation has been construed as structural”, and therefore part of the product.
Id. at 3-4.
. The CCPA recognized that "some courts have construed such claims as covering only a product made by the particular process set forth in the claim and not to the product per se”, 357 F.2d at 682 n. 5, 149 USPQ at 58 n. 5 (citations omitted), and indeed there are some claims for which this is the correct construction — such as the Atlantic claims.
. Accommodation of the patent law to evolving technologies is not unusual. Another example is seen in the deposit law, whereby certain biological products may be described, in compli-anee with section 112, simply by making a sample publicly available. See In re Argoudelis, 434 F.2d 1390, 168 USPQ 99 (CCPA 1970).
. The Atlantic panel misdescribes Scripps in its usage of “undeveloped”. That usage did not refer to the law governing interpretation of these product-by-process claims, which was well briefed and developed. What was undeveloped, as the Scripps opinion states, were the facts relating to the question of infringement of these claims. Scripps, 927 F.2d at 1583, 18 USPQ2d at 1015 ("on an undeveloped record, we consider only the questions of law”).
. While the CCPA and this court have endorsed the procedure whereby patent examiners view claims broadly during the examination process, this is an administrative mechanism for improving examination. The law is established by the statute and the courts.
. This dissent is not the place in which to argue whether these cases controlled the result in Scripps. The correct place was, in the first instance, when the Scripps opinion was before the court, on hearing and rehearing and suggested rehearing en banc; and secondarily, when the Atlantic opinion was before the court, on hearing and the requested rehearings en banc. This public debate may add sunshine to the court’s "percolations”, but failure to resolve the debate does not add stability to the law.
. This text appears between the two quotations from Adelman that the Atlantic panel presents to contrary purpose.
. In South Corp. the Federal Circuit stated:
For every panel of judges of this court to examine anew every issue presented would be a practice devoid of counterbalancing advantage. ... We choose therefore to begin on a readily available and clearly identifiable base, maintaining at the same time a controlled capacity for change when change is compelled.
As a court of nationwide geographic jurisdiction, created and chartered with the hope and intent that stability and uniformity would be achieved in all fields of law within its substantive jurisdiction, we begin by adopting as a basic foundation the jurisprudence of the two national courts which served not only as our predecessors, but as outstanding contributors to the administration of justice....
690 F.2d at 1370-71, 215 USPQ at 658.