3 - Informed Subject Matter

Summary

This chapter looks at the way that patent law resolved some of the other problems that chemical subject matter posed, particularly in working out how to define the boundaries of what was being examined or protected. While patent law employed a number of different strategies to delimit a fickle, empirically based, and changing chemical subject matter, this chapter highlights the central role that the materiality of chemical compounds played in helping the law to accommodate chemical inventions.

Introduction

One of the challenges that a fickle, empirically based, and rapidly changing subject matter posed for nineteenth-century patent law was working out how to define the boundaries of what was being examined or protected. While patent law employed a number of different strategies to delimit chemical subject matter, one stood out. This was the decision to treat intangible chemical property as if it was coextensive with the material chemical compound described in the patent, that is, with the compound itself. Unlike with mechanical inventions, where protection extended beyond the machine as described in the patent to include inventions that were considered to be legally equivalent to the patented invention, the intangible property in a patented chemical compound was not only treated as if it was coextensive with the material chemical compound, it was also treated as if it was a ‘chemical individual’¹ or a ‘singular point in the general field of matter’.²

The decision to deal with some of the idiosyncrasies of organic chemistry by treating the intangible chemical property as if it was coextensive with the individual chemical compound was evident in the way that chemical subject matter was interpreted for the purpose of determining whether it met the doctrinal requirements of novelty and non-obviousness. While in considering the novelty of other types of subject matter, the law was willing to analogise and extrapolate away from the prior art; this was not the case where the prior art consisted of chemical compounds, where the information was restricted to the material chemical compound. The reason for this was that while it was often possible with mechanical inventions to predict whether the operation of a novel material substituted in an old combination could achieve similar results, this was not the case with chemical compounds where prediction could not occur without experiment.³

Similarly, as it was not possible to predict what chemical subject matter would be like without creating it first, the courts were cautious about assuming the obviousness of a step taken by a chemist, even when the prior art was very close.⁴ This can be seen, for example, in the 1928 Supreme Court decision of Corona Cord Tire Company v. Donovan Chemical Corporation, which concerned a patent for a chemical compound (an ‘accelerator’) that improved the elasticity, tensile strength, and other desirable commercial qualities of finished rubber products. The patent in question was for the discovery of a new type of accelerator, called diphenylguanidine. As the patent noted, diphenylguanidine was closely related chemically to another type of guanidine – triphenylguanidine – which had already proven itself useful as an accelerant. Drawing on the fact that the chemical compositions resembled each other, the petitioner argued that the patent was invalid because the utility of diphenylguanidine as an accelerator was ‘plainly indicated by general chemical knowledge’.⁵ That is, a person skilled in the art could have discovered diphenylguanidine merely using knowledge which ‘was fully and readably available to everybody in the art and without exercising any inventive faculty whatsoever’. Chief Justice Taft disagreed saying that ‘the catalytic action of an accelerator cannot be forecast by its chemical composition, for such action is not understood and is not known except by actual tests’.⁶

The decision to treat the intangible chemical property as if it was coextensive with the material chemical compound was also evident in the way that chemical subject matter was construed when questions of infringement arose. In many areas of patent law, a desire to protect the equity of an invention – that is a desire to ensure that third parties were unable to avoid an accusation of infringement by making minor (non-inventive) changes to the invention – led to patents being read in a way that extended protection beyond a strict literal reading of the claims to include similar or equivalent inventions. Under the doctrine of equivalents, for example, protection extended to include situations where a would-be infringer replaced or substituted part of an invention with something that performed the same or equivalent function. As Robinson said, an equivalent meant the ‘interchangeability of agencies which are known in the arts to be capable of serving the same purpose as integral parts of the invention. By identity is meant, not the identity of tangible embodiments, but identity of effect, function or means’.⁷ What occurs in these situations is that the scope of the subject matter is shaped by legal considerations such as fairness to the patentee or a desire to protect the equity in the patent. Whatever term is used, the result is the same: the scope of the subject matter extends beyond a literal scientific or technical understanding to include things which are deemed to be legally equivalent.

While the doctrine of equivalents was widely applied in patent law, the idiosyncratic nature of organic chemistry meant that it was not extended to chemical compounds.⁸ Specifically, the lack of prevision that was a characteristic of organic chemistry meant that it was not possible to abstract away from the invention as specified in the claim to include equivalent inventions. This was because while it was possible to ‘predict with confidence in mechanics, in chemistry you almost entirely fail … as you can not anticipate the result’.⁹ As one commentator noted, with mechanical inventions it was generally possible to state with certainty whether a mechanical element was equivalent to another element. It was possible to ‘substitute a gear by a pulley, or a cam by a crank and obtain exactly the same result’. However, ‘in chemistry no one element or reagent is universally equivalent to another, and in most cases it is not possible to predict that they are absolutely equivalent except by actual experiment’.¹⁰ This was because ‘[y]ou cannot, because sulphuric acid will succeed, tell at all that nitric acid will succeed, or that any other acid will succeed, until you have tried. You cannot anticipate the result: it is a mere question of result upon experiment’.¹¹ Or, as Ruby put it, the fact that a chemical compound was ‘an embodied utterly unique indissoluble ensemble of properties’ meant that there were ‘no equivalents of a chemical compound’: the only equivalent of a chemical compound was the compound itself.¹² This meant that unlike the case with mechanical inventions, where protection extended beyond the machine as described in the patent to include inventions that were considered to be legally equivalent to the patented invention, the intangible property in a patented chemical compound was treated as if it was coextensive with the material chemical compound.

The decision to treat chemical subject matter as if it was a closed, singular, and bounded object that was coextensive with the material chemical compound reached its highpoint in the procedural requirement that as part of the application process patentees were required to deposit specimens of their chemical inventions with their written descriptions at the Patent Office (see Figures 3.1–3.3). This longstanding requirement, which exists in a modified form today, was first introduced into US patent law by the 1793 Act to Promote the Progress of Useful Arts.¹³ The 1793 law provided that where the invention was of a composition of matter, every inventor was required to submit ‘specimens of the ingredients, and of the composition of matter; sufficient in quantity for the purpose of experimentation’.¹⁴ Patent Office practice of exhibiting models and specimens was formalised in the 1836 Patent Act, which provided that ‘models and specimens of compositions … patented and unpatented deposited in the Patent Office, should be arranged in suitable galleries and kept open for the inspection of the public’.¹⁵ The rules in relation to specimens remained unchanged until the 1870s when because of a growing concern about the cost of storing and exhibiting models and specimens, the law was changed so that applicants were only required to ‘furnish specimens of the composition, and of its ingredients, sufficient in quantity for the purpose of experiment, if required by the Commissioner’.¹⁶ Despite this change, the Patent Office rules retained the proviso that where an article was not perishable, ‘a specimen in the composition claimed, put up in proper form to be preserved by the office must be furnished’.¹⁷

Figure 3.1 Ferrous carbonate patent specimen

Josiah Lilly, ‘Composition for the Production of Ferrous Carbonate’ US Patent No. 876,366 (14 Jan 1908).

Courtesy of the Division of Medicine and Science, National Museum of American History, Smithsonian Institution.

Figure 3.2 Explosive compound patent specimen

Harry D. Van Campen, ‘Explosive Compound’ US Patent No. 288,516 (13 Nov 1883).

Courtesy of Hagley Museum and Library.

Figure 3.3 Detergent compound patent specimen

Edward Henderson, ‘Detergent Compound’ US Patent No. 259,389 (13 June 1882).

Courtesy of Hagley Museum and Library.

Initially, the public were alerted to the existence of specimens by patentees in their patents, usually in the description but sometimes in the claims. For example, Robert Bartholow’s 1865 patent for an improved oil for paint stated: ‘I do hereby declare that the following is a full and exact description … being had to the accompanying specimen’.¹⁸ Similarly, the 1845 patent for a new and improved dye made from spent madder called ‘carasene’ stated that: ‘A specimen of the spent madder from which the carasene is made accompanies the specification in a packet marked “spent madder”, and of the dye-stuff, after it is prepared, in another packet marked “carasene”’.¹⁹ From 1880, notification that a specimen had been lodged as part of a patent application shifted from the body of the patent to the header material, where it accompanied information about the patent, including the name of inventor and assignee, the name of the invention, when the application was filed, and the date when the patent was granted (see Figure 3.4). From 1880 to 1905, the Patent Office also included information about whether a specimen had been deposited with a patent – ‘specimen’ or ‘no specimen’ – in the monthly summary of patented inventions that was published in the Patent Office’s Official Gazette (see Figure 3.5).²⁰

Figure 3.4 Lilly Patent

Josiah Lilly, ‘Composition for the Production of Ferrous Carbonate’ US Patent No. 876,366 (14 Jan 1908).

Courtesy of the United States Patent and Trademark Office.

Figure 3.5 USPTO Official Gazette summary of Lilly patent

US Patent Office, Josiah Lilly, ‘Composition for the Production of Ferrous Carbonate’ US Patent No. 876,366 (14 Jan 1908), patents granted January 14, 1908 (1908) 132 Official Gazette of the United States Patent Office 255.

Courtesy of the United States Patent and Trademark Office.

Historically, there has been very little consideration given to chemical specimens and the role they play in patent law. To the extent that they have been discussed, it was usually as an afterthought when patent models were being considered. As with chemical patents generally, chemical specimens were overshadowed by mechanical inventions and their models. While it is unclear how frequently specimens were used and what role they played in the examination process, there is evidence to suggest that in the later part of the nineteenth century, specimens were commonly used when chemical compounds were patented (even when the use of patent models declined). As Walker said in his 1883 patent law treatise, the Commissioner ‘does at least call for at least a specimen of the composition, put up in proper form to be preserved, unless that composition is in its nature perishable’.²¹ Specimens were also important enough in 1880 for the Patent Office to include information about whether a specimen had been deposited in the header information of each chemical patent and in the Official Gazette.²² While it is difficult to obtain exact figures, an examination of the chemical patents granted across the later part of the nineteenth century shows that specimens were widely used. It also seems that specimens continued to be used in the early part of the twentieth century. As Hugo Mock wrote in 1911: ‘It is a somewhat curious feature of Patent Office practice that for some years the only department of the Patent Office requiring what may be termed models of inventions, is that relating to novel chemical products, as specimens of new chemical products are generally requested by the Patent Office examiners. It is many years since models have either been requested or accepted in other lines of invention.’²³

Chemical specimens performed a number of different roles in patent law. Like patent models, specimens functioned as proof of the existence of the invention.²⁴ In the same way in which the Patent Office used the model requirement as a way of deterring applicants from applying for improbable inventions such as perpetual motion machines, the specimen requirement was also seen as a de facto workability requirement for chemical compositions.²⁵ For example, in an action relating to the patentability of artificial alizarine (a red dye originally obtained from the root of the madder plant) that was produced from the chemical compound anthracene (that was derived from coal tar), it was argued that the absence of a specimen showed that the claimed artificial alizarine was not a patentable composition of matter. While there was some indication that workability was one of the reasons why specimens were required in the United States, it was not a primary reason. This was in contrast to the position in the United Kingdom where from 1907 applicants for chemical patents were required to deposit samples of their invention at the UK Patent Office. This was to prevent foreign (German) patentees from lodging incomplete or faulty applications, or applications that did not properly disclose how to make the patented invention.²⁶

Importantly, because the specimens deposited at the US Patent Office had the same proportions of elements as the chemical compound that had been made in the laboratory, the specimen not only functioned to define the invention: the specimen was the invention. By requiring applicants to place their chemical inventions in glass bottles, paper sachets, and metal cans, the chemical specimen helped to reinforce the mental or semiotic representation of the chemical substance as an individual bounded object. This also allowed chemical compositions to be seen as bounded ‘wholes’ (which helped to answer the question of what it means to speak of closed chemical subject matter given that chemical compounds as liquids, powders, amorphous solids, and gases have no inherent shape or form). In turn, by limiting protection to the specific substance contained in the specimen bottle, paper sachet, or metal can, the specimen requirement also reinforced the idea that protection was not available for abstract classes or groups of inventions. That is, it helped to individualise the subject matter.

Interestingly, the ability for chemical samples to individualise inventions was used by the German Patent Office in the 1880s to deal with a problem that had arisen in relation to chemical patents in Germany. While the German Patent Law of 1877 excluded chemical substances from the scope of patentability, it did allow chemical inventions to be patented insofar as they ‘concerned “a particular process” (ein beslimmles Verfahren) for the manufacture of such substances’. While the law had been designed to exclude product patent protection for chemical compounds, nonetheless applicants attempted to use the fact that the law did not define what was meant by a particular process to circumvent the exclusion by indirectly claiming large classes of compounds.²⁷ This was particularly the case with azo dyes (which was a large class of synthetic organic dyes). The reason for this was that the dye industry used a specific process called a coupling reaction or Griess’s method to produce new synthetic compounds that involved ‘the pairwise combination of diazo compounds with aromatic amines or phenols to form azo compounds, i.e., compounds with a double nitrogen (azo) group uniting aromatic rings’.²⁸ The making of azo dyes was ‘an endless combination game. The number of possible combinations was estimated at more than 100 million’.²⁹

Building on the fact that the coupling reaction used by the dye industry could be construed as a particular process, patent applications for processes for the preparation of chemical substances frequently claimed the protection of groups or classes of compounds that had been produced by this process. To maximise protection, patents in the field of azo dyes also often claimed classes of analogues, homologues, and isomers of those compounds. As a result, patents for chemical inventions would often claim hundreds of individual substances. The practice of claiming classes or groups of compounds (rather than individual compounds) not only created administrative problems for the German Patent Office, who had to examine and classify the applications, by effectively allowing de facto product protection it also undermined German patent policy that had sought to exclude chemical products from the scope of patentability. In 1887 the German Patent Office responded to this problem by issuing a regulation that required patent applicants to provide samples of the substances for which they claimed protection.³⁰ As the new regulation stated:

In view of the fact that patent applications for processes for the preparation of chemical substances frequently claim the protection of entire groups of bodies without convincingly demonstrating the technical utility of the individual members of these groups, samples of the substances presented will be kept in the Patent Office so that, for instance, if they are important pieces of evidence in the course of any controversy and for the appraisal of new patent applications, in accordance with the wishes of the industrialists involved.³¹

While an exception was made in the case of explosive substances, the new regulation provided that for ‘those patent applications which relate to new methods of presentation of chemical substances, samples of these substances, as well as those intermediaries necessary for the production of substances which are still unknown, should be attached in two copies’. Specifically, applicants had to submit samples with a mass of approximately 8–10 grams in glass flasks of approximately 30 mm diameter and 80 mm height with glass stoppers, sealed with the seal of the patent examiner, and with a description matching the contents provided. By limiting protection to individual physical samples, the German Patent Office found a way of excluding applications for classes of chemical compounds. As in the United States, the intangible chemical property was treated as if it was coextensive with the chemical compound in the glass flask.³²

The treatment of intangible chemical property in the United States as if it was coextensive with the material chemical compound was effectively a taxonomic-like decision that limited chemical subject matter to the level of species.³³ One of the consequences of this was that the intangible chemical property did not extend upwards to include more abstract classifications such as classes or families of chemical compounds, nor to what a more generous abstract (equity-based) legal reading might have provided. It also meant that the intangible property did not extend to equivalent or similar inventions, to abstractions of the invention, nor to other inventions in a species-genus like relationship. Rather, patented chemical compounds were treated as if they were interchangeable with the individual material chemical compound. In this sense, patent law mirrored nineteenth-century organic chemistry where the ‘most fundamental category of chemical practice – that is, of experimentation and classification – was that of a chemical substance’.³⁴

The legal focus on specific compounds rather than on more abstract classes of subject matter helped US patent law to deal with the empirical nature of chemical subject matter. It also allowed the law to deal with the fickleness of chemical compounds, with the fact that slight changes in the ingredients in a compound or the conditions under which the ingredients were combined could fundamentally change the nature of the resulting compound. This was done by ensuring, for example, that if an inventor disclosed a new chemical compound (sulphanilyl cyanamide) that was made up of five elements: carbon (42.6%), hydrogen (3.5%), nitrogen (16.3%), sulphur, (16.3%), and oxygen (21.3%) with the empirical formula C₇H₇N₃SO₂, they would not be entitled to claim ‘a compound consisting of carbon, hydrogen, nitrogen, sulphur and oxygen’ because of the rule that ‘an inventor is not entitled to claims broad enough to cover subsequent independent discoveries of others’.³⁵ While in this case the inventor may have discovered one compound consisting of these five elements, they ‘would not be entitled to a patent covering every other compound which might therefore be discovered having those same elements in its make up’.³⁶ For similar reasons, an inventor was not entitled to claim a range of the various elements such as a ‘compound consisting of carbon 15 to 45 %, hydrogen 2 to 6 %, nitrogen 10 to 25 %, sulphur 15 to 39 % and oxygen 10 to 40%’,³⁷ the reason being that these proportions would extend to cover a number of different chemical compounds with very different structures and properties, including taurine (H₂NCH₂CH₂SO₃H), acetylthiourea (CH₃CONHCSNH₂), and nitrothioene (NO₂C₄H₃S).

The treatment of the intangible chemical property as if it was coextensive with the material chemical compound also meant that the intangible chemical property did not extend downwards (or internally) to the hidden microworld of the chemical compound. Rather protection was limited to the surface of the compound. In this sense, patent law followed the practice within science where the inability to explain what went on below the surface meant that when chemists interpreted or represented the results of their experiments or when they described and talked about what they had created, they ‘were restricted to an operational level of macroscopic objects’.³⁸ To do this, patent law embraced the agnosticism that allowed chemists to work with and create new chemical compounds without having to commit to a particular way of thinking about what went on below the surface in chemical compounds.³⁹ In so doing it allowed patent law to deal with the fact that organic chemists were unable to explain the reasons why things had happened and why it was, for example, that mixing compound A with compound B produced compound X.

Based on the idea that the ‘law requires no further certainty than science can afford’,⁴⁰ patent law was indifferent to what happened below the surface of the compound, to the reasons why chemical change occurred, and to the workings that occurred when ingredients were combined. It did not matter, for example, that the ‘method by which the ingredient perform[ed] its function in the combination’ may have been ‘entirely undiscernible’,⁴¹ nor that ‘the inventor was … in total ignorance of the scientific nature of what he employed and what he did’.⁴² Instead, all that mattered was that the patentee had created something that could be called a composition of matter, that they had ensured that the composition was in a format that allowed third parties to replicate the invention, and that they had met the other criteria for patentability.⁴³ For example, in response to an argument that a patent for a ‘spirit varnish’ that gave leather a bronze finish similar to a French metallic bronze finish was invalid because the specification did not explain why ‘the articles are so compounded as to produce a chemical change’, Justice Sheplex said it was ‘not essential that the inventor should have been sufficiently understood or accurately stated the philosophy of a process which he had invented and reduced to practical use’.⁴⁴ So long as the specification allowed for the invention to be used, this was enough. It did not matter that the reasons why something occurred were unknown; what was important was that the end result was achieved by following the directions in the patent.⁴⁵

While the law’s response to the fact that chemists were unable to explain what was happening beneath the surface was typically framed as one of indifference, the law’s pragmatic acceptance of the end products of the research process was not neutral. This was because although the law may have been indifferent to the reasons why chemical change occurred, this did not mean that the science that it accepted was. Like a Trojan horse, the veil of legal indifference not only allowed patent law to accommodate and protect organic chemical compounds, it also allowed certain types of scientific thinking to be inculcated within the law. Although science could not explain why and how chemical change occurred, it did embody theories (models, formulas, etc.) about what was going on, along with the language, rules, and techniques to identify and describe what was created. This was particularly the case in relation to the acceptance of chemical formula as a way of representing chemical compounds. This is because the writing of a chemical formula ‘is not innocent. It is ideology laden. It carries, besides its face value, another message; in this case, the modern reunification of the theoretical and the experimental’.⁴⁶ As we will see, the adoption of structural formula in the later part of the nineteenth century profoundly changed the way that patent law interacted with chemical subject matter.

Informed Subject Matter

While the decision to treat intangible chemical property as a singular bounded object that was coextensive with the material chemical compound played an important role in allowing nineteenth-century American patent law to accommodate some of the idiosyncrasies of chemical subject matter, it only tells part of the story. The reason for this is that chemical subject matter was not merely coextensive with the physical chemical compound; it was also an ‘informed material’ that carried particular ways of thinking about the world with it. The idea of informed material is based on Whitehead’s argument that material entities, including chemical compounds and molecules, are not bounded, discrete objects.⁴⁷ Rather, material entities extend into other entities while folding elements of other entities inside them. From this perspective, chemical compounds ‘should not be seen as discrete objects, but as constituted in their relations to complex informational and material environments’.⁴⁸ Importantly, while chemical compounds exist in an informational and material environment, this environment is not simply external to the object. Rather, this environment enters into the constitution of the chemical compound itself.⁴⁹ From this perspective, the research process operates to build information into the chemical compound. In this sense, to say that a material entity such as a chemical compound is informed or ‘rich in information’ is to say that the material embodies information.

The result of this is that even when chemical substances were treated as closed, fixed objects, they still embodied information that connected compounds to other compounds, that told something of the compound’s past (and whether it was novel and non-obvious) and, in some cases, its potential future (utility). Patent law relied upon the informed nature of chemical subject matter to accommodate the idiosyncrasies of nineteenth-century organic chemistry in a number of ways, perhaps the most important being in the identification of the subject matter (both during the application process and also after protection was granted). Specifically, patent law relied upon the fact that as informed chemical compounds carry their context with them, this meant (at least in theory) that compounds could be traced as they circulated beyond the reach and control of their creators.

The question of how the ephemeral and malleable intangible interest is to be described and identified has long troubled intellectual property law. These problems were amplified with chemical subject matter because chemical reactions (along with ‘atoms’ and molecular structures) were invisible processes that could not be seen, touched, or otherwise observed.⁵⁰ As Judge Lacombe said in the 1897 decision of Matheson v. Campbell, while the ‘observation as the eye can give to the machine at rest and in action … will be ordinarily sufficient to determine its classification’, it was ‘far different … with a chemical compound’. The reason for this was that ‘[n]o mere observation by the eye, supplemented even by taste and touch, can go very far towards a solution of the problem. The same mysterious forces through whose action and reaction the compound was produced must be availed of to disintegrate and disrupt before there can be any assurance of what it is that we have before us’.⁵¹ The problems that arose in ‘defining something invisible’,⁵² which Edward Thomas saw as the root cause of many of the problems that arose with chemical subject matter, were compounded by fact that chemical subject matter had no inherent external shape or form.

As Robinson said, the particular nature of chemical compounds led ‘to radical differences in the rules by which the identity of these’ compounds is determined.⁵³ To understand the reasons why these rules were so different, it is important to appreciate that the ingredients, along with the way they were combined, were critical to the nature of the resulting chemical compound; a slight change in proportions or the conditions in which the ingredients were combined could lead to a very different compound. In this sense the ingredients determined the nature and make-up of the compound. At the same time, however, chemical compounds differed from other combinations in that once they were united, the ‘ingredients or elemental forces’ ‘often become individually undiscernible by human sense’.⁵⁴ This is because chemical compounds do ‘not carry any characteristics peculiar to the process used in the manufacture by which the latter could be identified and by which infringement could be established directly’.⁵⁵ In other words, a chemical ‘composition of matter is a complete and independent means, having an existence distinct from that of the substances of which it is composed, and the processes by which it is created’.⁵⁶ One of the consequences of this is that the ‘character of a composition of matter cannot … be determined from an examination of its elements alone, nor of the method by which they have been combined. It must be judged also by its own intrinsic attributes’.⁵⁷ The upshot of this was that products had to be identified independently of the process by which they were made. As Lewers said, ‘Every patent for a product or composition of matter must identify it, so that it can be recognised aside from the description of the process for making it, or else nothing can be held to infringe the patent which is not made by that process’.⁵⁸

The fact that a chemical compound was integrally connected to the ingredients and how they were combined and, at the same time, independent and distinct from those ingredients meant that description was a two-stage process (which was mirrored in the requirement that patentees needed to deposit specimens of both the ingredients and the resulting compound as part of the application process). This meant that it was necessary to describe the ingredients and how they were combined to create the compound in question. At the same time, it was also necessary for patentees to describe the resulting chemical compound so that it could be identified as an entity in its own right.

The first stage of the process of describing a chemical compound was to detail the ingredients and how they were combined to create the compound in question. The fickle nature of chemical compounds meant that it was important that the ingredients that were used to create a composition were set out clearly and precisely.⁵⁹ This was because while replacing ‘an iron bar in place of a wooden one and serving the same purpose’ did ‘not change the identity of the machine’, with chemical inventions even a very small change either in the ingredients used or in the way that the ingredients were combined could lead to a very different compound.⁶⁰

As Rossman explained in his 1932 treatise on chemical patent law, a ‘patent for a composition of matter must name or describe and give the exact proportions of the ingredients used in addition to describing how the ingredients are mixed or combined to give the desired result. Inasmuch as the discovery of a new substance by means of chemical combination is empirical, and results from experiment, the law requires that the description in a patent for such a discovery should be specially clear and distinct’.⁶¹ Unlike the case with mechanical compounds, where the law was willing to allow the patentee to leave the proportions to be determined at a later stage (so long as the elements and their characteristics were known), with chemical compounds, where the end result was often dependent on the precise proportions used, the law not only required the patentee to provide details about the ingredients and how they were to be combined, it also required patentees to provide details about the specific proportions needed to produce the compound.⁶²

In order to describe ingredients with the requisite degree of precision, patentees tended to describe the ingredients in terms of proportions or ratios, rather than in terms of a fixed specific weight (such as ‘rosin, three hundred pounds, Kentucky cement seventy-five pound’ etc.⁶³). As Ruby said, chemical compound were ‘defined and claimed in terms of the chemical elements that are ultimately composed, by bloc-or graphic formula, parts-by-weight, or percentages compositions, and hence in terms of invariant ratios of the amounts of the chemical elements present in chemical combination in the true chemical compounds’.⁶⁴ In some cases (particularly in earlier patents), the proportions were simply listed in the patent, such as in US Patent Number 45,518 for ‘Improved Composition for Crayons’, which said that for ordinary school uses on a blackboard, the crayon should be made up using the ‘formula’: ‘kaolin 48 parts, calclined plaster-of-paris (gypsum) 16 parts, water 35 parts, and white glue 1 part’.⁶⁵ More commonly, the proportions were captured in empirical or rational formula that specified the elements and their proportions in the patented compound. Thus, for example, US Patent Number 775,978 claimed a new product, ‘a fragrant oil having the [empirical] formula C₁₅H₂₆O’.⁶⁶ Or, US Patent Number 719,720, which claimed a new indigo-reducing agent that had ‘a chemical composition corresponding to the [rational] formula Na₂S₂0₄ + 2H₂O’.⁶⁷ Here, the chemical formula, which provided a succinct guide or blueprint to the elements in a compound, denoted its composition and helped to distinguish it from other compounds. The fact that the ingredients were listed in proportions or ratios rather than in fixed quantities reinforced the idea that a chemical compound was an individual that had no inherent size or (external) shape and whose parameters were determined by the proportion of its parts.

The fickle nature of chemical compounds also meant that it was important that the patentee provide clear and precise instructions as to how the ingredients were combined.⁶⁸ Typically, the descriptions of how the ingredients were to be blended read like cooking recipes. For example, an early patent for artificial alizarine (dye) began by stating: ‘We take one part, by weight of anthracine, two and half parts, by weight, of biochromate of [potash], potassa, and ten or fifteen parts, by weight, of concentrated acetic acid, and we heat these substance together in a vessel either of glass or clay to about 100° centigrade to 120° centigrade, till nearly all of the bichromate of potash is dissolved and the liquid has acquired a deep green colour’.⁶⁹

As well as describing the ingredients and how they were combined in order to create the compound, it was also necessary for patentees to describe the resulting chemical compound as an entity in its own right. Importantly, this had to be done in such a way that the compound could be identified independently of the ingredients that it was made up of. For the most part, patent law followed the practice within chemistry when it came to describing and identifying chemical compounds.⁷⁰ While the techniques used to describe chemical subject matter changed considerable across the nineteenth century,⁷¹ one thing that remained constant was that the identification of chemical substances ‘always took place on the macroscopic level of substances’.⁷²

At the beginning of the century, plant and animal materials were identified and classified according to their natural origin, the mode of extraction or preparation, observable properties, and practical uses.⁷³ Using these criteria, early chemists were able to distinguish and characterise a substantial number of simple chemical substances such as sea salt, soda, magnesia, potash, along with certain elements (such as sulphur). ‘These were set apart from each other by qualitative distinctions: texture, color, odor, taste, the source from which the material was obtained, or the use to which it was put.’⁷⁴ Early chemical patents followed this practice and described chemical subject matter using qualitative criteria.

As the century progressed, patentees increasingly relied on more quantitative ‘chemical’ criteria to describe their chemical compounds.⁷⁵ Thus, patentees began to describe their chemical substances in terms of physical constants and measurable chemical properties such as how the compound reacted when it was combined with other compounds (or reagents), along with the all-important melting and boiling points (which were ‘the best tools available for the identification of substance’).⁷⁶ For example, the traditional practice of describing sugars by their sweet taste and their ability to support fermentation was slowly supplemented by descriptions that used empirical formula, melting points, and hydrazine tests (which examined how the compound reacted with phenylhydrazine).⁷⁷ While the Patent Office did not formally insist on a particular ‘style or phraseology’, patentees were advised ‘to use the conventional chemical nomenclature wherever possible for the sake of definiteness and clarity’.⁷⁸ As we will see in Chapter 4, patentees followed this advice and adopted the naming practices developed by chemists to identify chemical compounds.⁷⁹

The result of this was that as well as including the name and chemical formulae of the patented compound, it was also common for compounds to be described in terms of how they looked, smelt, or tasted, along with their melting and/or boiling point, and a description of what happened when particular chemical reagents were applied to the compound. To situate and contextualise the subject matter in relation to similar types of subject matter, patentees also often included a history of the invention, along with what amounted to a ‘brief essay on the materials or the steps usable and a brief mention of the difficulties and failures of the prior art, and to include detailed examples of practising the invention’.⁸⁰ For example, US Patent Number 400,086, which was for a pharmaceutical product known commercially as ‘phenacetine’ and chemically as ‘mono-acetyl-paramido-phenetol’, stated:

The product herein described, which has the following characteristics: it crystallizes in white leaves, melting at 135° centigrade; not colouring on addition of acids or alkalies; is little soluble in cold water, more or so in hot water; easy soluble in alcohol, ether, chloroform, or benzole; is without taste; and has the general composition C₁₀H₁₃O₂N. ⁸¹

In a similar manner, Hoffman and Weinbergs’s 1886 patent for ‘a new coloring matter’, which they called napthol-black, was described as producing ‘on the fiber in an acidulated bath dark-blue shades’ and as being ‘very soluble in water, insoluble in spirit, dissolves in strong sulphuric acid with green color’.⁸²

While patent law followed the practice within chemistry of describing chemical compounds in terms of how they smelt, looked, and tasted, the temperature that they boiled or melted at, and how they responded or reacted when combined with other compounds, patent law added a subtle twist. This was because rather than merely providing a description of a compound’s witnessable properties, the description of a compound in a chemical patent also included specific instructions that explained how the identity of the compound was to be determined.⁸³ That is, the descriptions included the experimental tests that were to be used to identify the patented compounds.⁸⁴ As Judge Lacombe said, these ‘tests of identity’ were ‘devised by those skilled in the art and science of chemistry, which, in their opinion, as experts, will reveal the secrets of the composition to make the answer to the question positive enough to support the judgement of a court’.⁸⁵ In this sense, patentees included both the experiments and the results of those experiments in their patents as ways of identifying their patented chemical compounds.

Like many things in relation to chemical subject matter, the process of identification and description used in patent law was an inherently empirical process.⁸⁶ As in organic chemistry, the process of identifying and classifying chemical compounds in patent law was ‘tied to processing and interpreting the experimental marks of the invisible object, and to the application of sign systems, culturally impregnated with meaning, in that endeavour’.⁸⁷ In the absence of any other way of identifying a chemical substance, the alleged infringing or anticipating compound was ‘tried’ to determine whether it met the experimental criteria set out in the patent: did it ‘melt at 135° centigrade’, was it ‘easily soluble in alcohol, ether, chloroform, or benzole’ and ‘without taste’?⁸⁸ Patentees used the presence or absence of these identifying traces or ‘marks of identification’⁸⁹ as litmus tests for determining whether a compound was the same as or different to the patented compound. If a claimant could show that the product ‘answers all the tests of the patent, and other well-known test not therein named’, this was taken as proof of similarity.⁹⁰ As the court said in Matheson v. Campbell, (an action in relation to a patent for a black azo dye), the ‘proofs show satisfactorily … that the defendants’ colouring matter possesses the same peculiar characteristics of the patented article’.⁹¹ This was sufficient ‘to establish the chemical identity of the defendants colouring matter with the complainants by the evidence of the results produced by each in the experimental tests’.⁹²

One of the consequences of the decision to use experimental tests to identify chemical compounds was that it increased the role that science played within the legal process. Another consequence of the decision to use experimental tests to identify chemical compounds was that it further embedded the chemist as expert within patent law.⁹³ This was because chemists not only devised the tests that were used to identify chemical compounds, they were also called on to undertake those tests and to interpret and explain the results to the courts.⁹⁴ While experts were commonly used in patent law to assist the courts in reaching decisions, the nature of chemical subject matter meant that the chemical expert or ‘patent chemist’⁹⁵ took on a particular importance. While with a mechanical device like a drill or a typewriting machine, ‘the deliverances of the experts are mere aids to the comprehension of the structure. If there be disputes among them as to how various parts are correlated and how they act, a judge must examine the device and decide for himself as to which is correct’. However, the situation was different with chemical patents. The reason for this was that there were ‘things which the independent senses cannot appreciate, which cannot be seen or felt or heard … the reactions of bodies into some chemical union or disunion, are matters in which a court must perforce depend upon the assertions of someone who has made a profound study of the matter’.⁹⁶ As a result, when dealing with chemical patents, the court had to ‘wait till someone skilled in the intricacies of that science and art appears to lead the way though its labyrinth of terms and symbols’.⁹⁷ Or, as the court said in Matheson v. Campbell, the determination of the issues raised in relation to the patentability of naphthol-black colour compound was made ‘more difficult by reason of the mass of expert testimony concerning chemical characteristics and laboratory processes, which the court cannot verify by inspection or experiment’.⁹⁸

The reliance on technical chemical information within the patent process led to calls for scientific experts to be embodied more formally within the legal process, rather than merely as witnesses (who largely operated as temporary visitors).⁹⁹ This included calls for the establishment of ‘technical juries’ consisting of experts engaged in the industry or science to which the action relates,¹⁰⁰ an unofficial and independent patent court which would be ‘equipped to deal quickly, economically and wisely with patents, especially in the highly technical chemical field’,¹⁰¹ and the appointment of ‘technical referees to assist the court to pass judgement’ on technical chemical patents.¹⁰²

The increased reliance on chemical information within the patent process also added impetus to the calls being made for a chemical laboratory to be established within the Patent Office to allow examiners to test the validity of applications.¹⁰³ The growing number of chemical applications was said to ‘render it highly desirable, and indeed indispensable, that the examiners should have at hand the means of arriving at correct and definite decisions’.¹⁰⁴ As a research engineer said in 1918, a ‘complete fireproof laboratory in the Patent Office Building for making physical, chemical, mechanical and electrical qualitative tests’ was needed ‘so that any questions arising in the course of the prosecutions as to the actual performance of processes and mechanism could be answered by authoritative tests’.¹⁰⁵ To this end, it was proposed to ‘have room fitted up as a laboratory, and that the Commissioner be authorized to procure the requisite apparatus at an expense not exceeding $800’.¹⁰⁶ The provision of experimental facilities in the Patent Office, which was intended to provide a legal-space in which chemical compounds could be tested and witnessed,¹⁰⁷ would ‘enable the examiner to verify or disprove alleged ingredients and results of applicants for patents for materials, processes and compounds’:¹⁰⁸ particularly in relation to whether an application met the requirement of adequacy of disclosure.¹⁰⁹

The repeated calls for the establishment of a chemical laboratory in the Patent Office were reinforced by the absence of a compulsory working requirement in US patent law which allowed foreign companies to take out product patents without requiring them to manufacture the products in the United States.¹¹⁰ The lack of a working requirement was particularly problematic in relation to chemical compounds given that the only way of knowing whether a patent did what it claimed to do was to follow the instructions in the patent and to replicate the invention. The lack of a working requirement (which would have acted as de facto proof that the written patent properly disclosed how to make the patented compound) created the potential problem that patents could be granted for chemical compounds that did not meet the disclosure requirement. This became a concern in the early part of the twentieth century when complaints were made that many of the chemical patents that had been granted to German companies were ‘faulty’ in so far as the patented chemical compounds could not be made following the instructions in the patents. The fact that attempts to replicate inventions had been made by scientists of high standing led to the conclusion that the patents contained deliberate misrepresentations and that the ‘literature on chemistry was clogged with such deceit’.¹¹¹ While it was possible for faulty patents to be invalidated after they had been granted, this was seen as a ‘tedious process, necessitating a great amount of laboratory work and expense and loss of time in litigation’.¹¹²

Despite repeated calls by successive Commissioners of Patents for the establishment of a laboratory within the Patent Office, concerns about ventilation, explosion, and fire (the latter being particularly important in light of the fires that had occurred at the Patent Office in 1836 and 1877 and the fact that the call for the establishment of a purpose-built laboratory in the Patent Office came at a time when the ‘practice of chemistry had never been more dangerous’¹¹³), combined with concerns about the cost of building and maintaining a laboratory meant that an in-house laboratory was never built. Other suggestions, including that the Patent Office could make greater use of existing government laboratories such as the Bureau of Standards¹¹⁴ or that the onus of proof be placed on the applicant to show the ‘correctness of the specification’ ¹¹⁵ were also rejected: the former because the Bureau was already overworked and the high cost, the later because it thought to be onerous on applicants of ‘small means’.

Decoupling Chemical Subject Matter from Its Material Form

While the ability of chemists to describe and identify chemical compounds improved greatly across the nineteenth century, many problems remained. In part this was because despite improvements in the accuracy of chemical analysis, there were still many problems including errors in collection, sampling, and measurement¹¹⁶ caused by things such as inaccuracy in chemist’s manipulations, accidental contaminations of samples, or parts of samples being lost when they were moved to new vessels or when they were weighed.¹¹⁷ As a result, when experiments were repeated, they often yielded different results. Given that ‘a relatively small error in the percentage composition could significantly affect the formula assigned to the compound’,¹¹⁸ these errors undermined the accuracy and thus the effectiveness of chemical formulae, which ‘remained unstable for much longer than is usually recognised’.¹¹⁹

Given that the study of chemical compounds often produced ambiguous outcomes, this undermined the effectiveness and accuracy of the empirical formula that were developed from this empirical information. These problems were compounded by the fact that while patentees used melting and boiling points to identify their compounds, nonetheless the ‘problem of how to obtain reliable, comparable boiling points that would function as useful markers of identity and purity proved persistent’ across much of the nineteenth century.¹²⁰

Another related problem facing nineteenth century chemistry and by default patent law was that absolute chemical purity was an unobtained ideal: local samples of chemical species were rarely identical with other samples in every single aspect – there were ‘varieties of indigo, potash, steel purchased from merchants; varieties of vitriolic acid, nitric air, spirit of wine prepared in the laboratory’¹²¹ – samples often differed in colour, smell, taste, consistency and properties.¹²² While the purity of chemical elements improved greatly across the century, this only served to create a new problem; namely, that chemical records were ‘often difficult to interpret because the names of substances … remained constant while their purity has undergone changes of various magnitudes’.¹²³ This undermined the effectiveness of patents as a source of reliable technical information.

The ability for patent documentation to identify and recreate the patented compound was also undermined by the fact that in many cases important aspects of ‘chemical knowledge did not reside in formula and structure, but rather in laboratory reasoning, the process by which chemists connected the minutiae of laboratory work with major advances in chemistry’.¹²⁴ This was particularly the case in relation to the creation of synthetic chemicals which was dependent ‘on practical experience that was developed, learnt and taught in a very particular place: the institutional chemical laboratory’.¹²⁵

These problems were compounded by the fact that in some ways organic chemistry was a victim of its own success. Improvements in chemical knowledge consistently forced chemists to ‘differentiate between compounds that were previously considered to be identical, and to recognise as mixtures materials hitherto thought of as pure’.¹²⁶ The constant revision had flow-on effects as it called into question and undermined existing practices which only served to complicate things further.

The ever-expanding number of chemical compounds also presented problems for patent law. Unlike the position with inorganic substances where the relatively small number of substances meant that ‘a statement of a novel compound’s constituent elements along with an identifying characteristic was enough to identify the compound’,¹²⁷ the huge number of organic compounds meant that more complex modes of description were needed. The rapid increase in the number of organic compounds not only exacerbated and highlighted the taxonomic and nomenclatural uncertainty that existed, it also made it much more difficult to navigate the prior art (which was essential for determining whether a would-be patent was novel). These problems were reinforced by the fact that it often took some time for the scientific community to reach agreement about a compound’s formula. Until this happened, a compound could be represented by a number of different formula. Water, for example, was represented at times as 2H+O, H₂O, and HO. In 1854, there were 11 different formula for acetic acid, which increased to 19 by 1861.¹²⁸

While the ability to name, describe, and identify new compounds was essential to the ongoing success of organic chemistry, it took second place to the creation of new compounds. As a result, there was often a lag between scientific innovations leading to new compounds and the development of the taxonomic tools needed to describe these innovations. This can be seen, for example, in relation to Graebe and Liebermann’s 1868 preparation of artificial alizarin (a red dye for cotton and a red pigment in painting) which is often seen as ‘the first time a chemist had succeeded in producing a particular target molecule by synthesis’. Prior to this, (natural) alizarin had been sourced from plants and insects. While the creation of the artificial synthetic dye marked a major advance in organic chemistry, the same cannot be said for the way that the invention was described in the patent. This is reflected in the fact that because Graebe and Liebermann were unable to demonstrate either the purity or the chemical identity of their synthetic alizarin, which ‘could be used in the same way as various madder compounds’, they were limited to identifying their product by the ‘yellow flocks of alizarin’ that was produced in the compound when a particular process was followed.¹²⁹

The upshot of this was that although there were many improvements in experimental practices across the nineteenth century, a number of problems undermined the accuracy and effectiveness of the way compounds were described and identified both within organic chemistry and by default in patent law. Here, the law was faced with a number of options. One possibility, which was not considered, would have been to deny chemical compounds patent protection on the basis that they did not meet the basic requirements of patentability. Another possibility, which was also not adopted, would have been simply to accept the best efforts of chemists to describe chemical subject matter, forcing examiners, patentees and interested third parties to make do with the descriptions that science could offer.¹³⁰ A third option, which was adopted, was to modify and adapt the nomenclatural and taxonomic practices used within chemistry to suit legal ends. As a result, the process of describing and identifying chemical subject matter in patent law became a scientific-legal hybrid. While this took many forms, I focus here on patent law’s use of physical specimens as a means of identifying patented chemical compounds. I look at the efforts by the Patent Office to ensure that the scientific prior art was legible to a legal audience in the next chapter.

Chemical Specimens

While specimens did not have a direct bearing on the patentability of chemical compounds¹³¹ nonetheless they played an important role in improving the accuracy and effectiveness of the way compounds were described and identified. In this sense it may appear that chemical specimens operated like patent models to evidence and showcase the invention. While chemical specimens were exhibited alongside patent models in the Patent Office Museum¹³² and at industrial exhibitions,¹³³ they were very different. This was because while the ‘materiality of the model’ may have ‘provided the basic medium in which inventions were revealed, scrutinized and compared’,¹³⁴ the materiality of chemical specimens performed a different role. The reason for this was that while patent models (along with some biological specimens) could be evaluated on their face,¹³⁵ chemical specimens were mute. As Lloyd Van Doren explained in an article written in the Journal of Chemical Education that introduced chemical students to patent procedure, the key difference related to the fact that ‘a mechanical patent has to do with something which is tangible, for example, a machine’. It was ‘quite possible for a court to look at the drawings or perhaps even at a model of the machine and be able to satisfy itself that’ the machine described in the patent would operate. ‘In short, something tangible will be presented’. In contrast, in the case of a ‘chemical patent which deals, for example, with a process for making anthraquinone, the court is not able from a drawing or even from a demonstration in the court room to visualise directly the operativeness of that process’.¹³⁶

One of the notable things about chemical specimens was that there was little to see.¹³⁷ Other than the specimen number, the names of the patentee (or assignee) and the invention, and the date the compound was patented, a dark glass bottle or a sealed paper sachet revealed little about the intangible chemical property that was hidden inside (see Figures 3.1–3.3). As a clerk of the US Patent Office said when reporting back on his visit to the 1851 Great Exhibition in London, while the machinery displays presented the spectator with ‘much to attract his observation and occupy his thoughts’, the specimens of chemical and pharmaceutical products provided ‘little that was interesting’.¹³⁸ The mute nature of chemical specimens meant that chemical proof was something that had to be mediated through the expertise of the chemist.

In so far as chemical samples were objects that either revealed or had the potential to reveal the ‘traces of the invisible objects of inquiry’, they were there to be tested (or at least potentially tested). This was reflected in the language of successive patent statutes which required applicants to provide specimens of ingredients and of the composition of matter ‘sufficient in quantity for the purpose of experiment’.¹³⁹ While the patent legislation required applicants to deposit samples to allow the application to be tested, as successive Patent Commissioners complained, the absence of an in-house laboratory in the Patent Office meant that it was not possible to examine chemical applications properly. As the Commissioner of Patents Benton J. Hall said, the lack of laboratory facilities in the Patent Office meant that there was ‘no means of testing such specimens that have been provided, although obviously within the meaning of the law’.¹⁴⁰ Despite this, specimens were still important.

The reason for this was that by linking the description of the patented invention to the physical chemical specimen, and by ensuring that sufficient materials were available ‘to allow experiments to be undertaken that revealed the essential features of the invention’,¹⁴¹ specimens ensured that the accuracy of the written description could be tested during the application process if needed. In doing so, chemical specimens provided ‘greater accuracy and completeness in the description of patented inventions’.¹⁴² This was because once a patent had passed through the examination process and received the official imprimatur of the Patent Office, it could be presumed that the written description was accurate: either because the description was clear on its face, or because when the Commissioner asked for a specimen to be deposited, the written description corresponded with the material specimen (if it didn’t, the patent would not have been granted). While applicants sometimes provided affidavits that attested to the qualities of the invention, the Patent Office was reluctant to rely upon this information because while the assertions may have appeared to be reliable, as the Commissioner of Patents said, ‘in the absence of some means of testing the truth of the facts claimed, it is impossible for the Office to determine with what degree of certainty which should exist whether the invention is novel and useful and should be covered by a patent’.¹⁴³

In this context it did not matter whether or not the patented compound had in fact been tested (the absence of an in-house laboratory within the Patent Office meant that this was rarely the case). It also did not matter if there were problems in the way chemical compounds were described. This is because the legal fiction of the chemical specimen allowed the patent system to operate on the basis that the description was accurate. By ensuring that any potential problems there might have been in identifying or replicating a patented compound were ‘resolved’, chemical specimens allowed third parties to place trust in the written descriptions of the patented chemical compound.

At first blush it may seem that chemical specimens operated like biological specimens in so far as they provided an objective standard against which the written description could be evaluated. While in some ways they were similar, they differed in a number of ways. With biological subject matter the deposited specimen, the name, and the description all worked in tandem to define the invention.¹⁴⁴ In contrast, with chemical inventions, the specimen did not operate in conjunction with the written description (and the name) to define the invention: the specimen was the thing that was being described: it was the invention. This had important consequences for the way patent law interacted with chemical compounds. Because the written description and the chemical specimen were the same thing, and because it was possible to replicate the specimen from the written description, after a patent was granted the written description of the chemical invention could be uncoupled from its material form (unlike the type specimen in biology which is permanently tied to the name and the description). Thus, while during the grant process chemical specimens played an important role in building trust in the accuracy of the written description, after a patent was granted the specimens were no longer needed. Post-grant, the written description not only provided third parties with all that they needed to know about the patented chemical compound (both to identify the compound and to recreate it), they also did so in an easy-to-use and comparatively uncomplicated way. Uncoupling the tangible chemical specimen in this way allowed people interacting with chemical patents to focus on the written description in the patent documentation, rather than having to go through the timely and arduous process of testing the specimen. Post-grant, it was no longer necessary to refer back to the specimen at all. In so doing, it allowed chemical (paper) patents to circulate as immutable mobiles: as closed, fixed, and trustworthy scientific objects.

The shift away from the material specimen towards the written description that occurred after grant was reinforced by the fact that once a patent was granted, the findings of the Commissioner in relation to specimens could not be questioned or challenged. This can be seen, for example, in the 1874 decision of Tarr v. Folsom, which concerned a challenge to a reissued patent for an antifouling paint that was said to have ‘launched the first industrial revolution in North America, that was commercial fishing’.¹⁴⁵ The patent was challenged on the basis that the reissued patent ‘described substantially different inventions from any described and shown in the original patent … or in the samples filed in the patent office in illustration thereof’.¹⁴⁶ The Circuit Court of New York rejected this argument and held that as the specification ‘clearly describes the composition of matter and all the ingredients and proportions, in language perfectly intelligible to those skilled in the art, it would not be invalidated by the failure to deposit in the patent office a sample of one of the ingredients’.¹⁴⁷ Importantly, the court was not willing to reopen the question of the accuracy of the specimen and its relationship to the written description and the intangible chemical property. As Judge Shepley said, the requirement to deposit a specimen was obligatory before the granting of the patent, where it ‘was for the commissioner to decide, before granting a patent, whether it is complied with. If he does so decide, and grants the letters-patent, that cannot be subsequently impeached by evidence tending to show a want of compliance with the law as to giving notice, or paying fees, or performing the other acts, or performing the other acts required before the patent is granted’.¹⁴⁸ That is, once the Commissioner had accepted that the written description corresponded to the deposited specimen, their decision could not be reopened or challenged.

By simultaneously black-boxing the chemical specimen and by decoupling the physical specimen from the written description, it was possible to focus on the paper form of the invention in the patent documentation. The focus on the written two-dimensional form of the invention was reinforced by the absence of a workability requirement which would have required patentees to show a material instantiation of the invention. The focus on paper-based inventions meant that in an infringement action or where the novelty of a patent was challenged, the written specification was treated as if it encapsulated the invention (or at least provided instructions for how the invention could be identified); it was the alleged infringing or anticipating compound, rather than the physical specimen, that was tested to see whether it complied with the descriptive tests set out in the patent documentation. In this sense, chemical specimens not only helped patent law to deal with any problems that might have arisen in the way chemical compounds were described, they also allowed the patent system to circumvent some of the problems that arose when dealing with empirical inventions more generally. In this sense, the legal fiction of the chemical specimen allowed the patent system to deal with chemical inventions in the much the same way as it interacted with mechanical inventions. While the process was not complete, it also played an important role in decoupling chemical subject matter from its material physical form.