Full opinion text
OPINION MAZZONE, District Judge. STATEMENT OF THE CASE This is an action for patent infringement between the makers of two popular aerosol anti-perspirant products. The plaintiff, Carter-Wallace, Inc. (Carter-Wallace), is a Delaware corporation with its principal place of business in New York. Defendant, The Gillette Company (Gillette), is a Delaware corporation with its principal place of business in Massachusetts. Carter-Wallace is the owner of United States Letters Patent No. 3,968,203, issued on July 6,1976. The patent application was initially filed on October 1,1965, and refiled in 1969. The patent is entitled “Aerosol Astringent Composition” and covers Carter-Wallace’s aerosol anti-perspirant, “ARRID EXTRA DRY.” A copy of the patent is attached to this opinion as Appendix A. Carter-Wallace charges that Gillette has willfully infringed Claims 1, 2 and 5 of the patent by making, using and selling fifteen different aerosol anti-perspirant packages embodying the product invention set forth in the above claims. The most commonly known products charged as infringing are marketed by Gillette under the name “RIGHT GUARD” anti-perspirant. Carter-Wallace seeks damages for past infringement, an injunction against future infringement, and attorney’s fees under 35 U.S.C. § 285. Gillette denies that it has infringed the patent in suit, and claims that the patent is invalid both for lack of novelty under 35 U.S.C. § 102(a), (b), (e) or (g), and obviousness under 35 U.S.C. § 103. The defendant also claims that the patent is invalid under 35 U.S.C. § 112 because: (1) the claims are vague and indefinite and do not claim the invention in sufficiently specific terms, (2) the patent fails to disclose the best mode of practicing the invention in that it does not disclose a valve for use in the claimed invention, (3) the patent claims more than was invented by including in the claims the container and the valve as well as the composition. Finally, the defendant claims that patent is invalid or unenforceable because the applicants withheld from the Patent Office certain prior art and information as to the relevance of prior art. Gillette also seeks an award of attorney’s fees under 35 U.S.C. § 285. Briefly, then, Carter-Wallace’s theory of the case is that two inventors, George Spitzer and Lloyd Osipow, invented a novel aerosol anti-perspirant composition which satisfied a long felt consumer need. The claimed invention led to the development of ARRID EXTRA DRY and eventually to the issuance of the patent in suit. Carter-Wallace claims Gillette’s RIGHT GUARD and other products are nothing more than thinly disguised copies of its ARRID formulations, and defendant is therefore liable for infringement. While Gillette asserts a number of defenses, the thrust of its case is that the Spitzer-Osipow discovery lacks the necessary features of a patentable invention, in that it simply contains ingredients known in the art performing their usual functions. At best, Gillette asserts, the inventors stumbled upon a particular ratio of these known ingredients which produced surprisingly good results. Even these results, however, Gillette claims were obvious in view of the teachings of the prior art. Finally, Gillette argues that, if the patent in suit were valid, it would be so only because the particular ingredients and numerical limits claimed in the patent were critical. Because its challenged formulations include certain slightly different ingredients, Gillette claims it cannot be held liable for infringement. The case was tried to the court without a jury. The record consists of the trial testimony, various depositions, and numerous other exhibits. Pursuant to Fed.R.Civ.P. 52(a), the following constitute our findings of fact and conclusions of law. BACKGROUND OF THE CASE Aerosol products have become common in the United States since World War II. They are generally sold in the form of valved containers in which a product to be dispensed and a liquefied propellant are held under pressure. When the valve is opened, the propellant and the product are expelled as a spray. The propellant evaporates in the atmosphere so that only the product reaches the area being sprayed. The amount of product deposited on the surface, and the spray pattern, depend upon many variables including the type of valve used, the amount and type of propellant, the nature and relative concentration of ingredients in the product, and the length of time the valve remains open. The convenience of aerosol packages has led to their use in dispensing insecticides, cosmetics, paints and, more recently, deodorants and anti-perspirants. Perspiration odor is not unpleasant in itself, rather, the decomposition by bacteria of organic matter contained in perspiration causes malodor. This unpleasant odor can be eliminated by either reducing underarm perspiration or preventing decomposition of the organic matter. Deodorants. destroy the bacteria or simply mask the unpleasant odor which it causes. Anti-perspirants, on the other hand, check both wetness and malodor by actually reducing perspiration. A perfume or fragrance ordinarily included in an anti-perspirant composition masks whatever odor is caused by the remaining perspiration, which is never completely eliminated. Anti-perspirants in the form of creams and lotions first became popular during and immediately after World War II. The active ingredient, usually aluminum chlorohydrate, was present in such products in the form of an aqueous solution. In order to be effective, aluminum chlorohydrate must be dissolved in water, producing aluminum ions. However, aluminum chlorohydrate salt, when dissolved in water, produces an acidic solution which is corrosive to ordinary metals. The salt is also hygroscopic, that is, it has the tendency to absorb water from the atmosphere. As a result, initial efforts to develop an aerosol anti-perspirant were hampered by serious problems such as can corrosion and valve clogging. Clogging results when residual alcohol in the valve orifice evaporates, leaving behind aluminum salt which absorbs water from the atmosphere and recrystallizes. In 1961, Gillette introduced RIGHT GUARD aerosol deodorant. RIGHT GUARD did not contain aluminum chlorohydrate. It prevented malodor from perspiration but did not reduce perspiration itself. RIGHT GUARD was a solution-type product like the early insecticides and contained no solid powder. It became a great commercial success, proving that the public would accept an aerosol underarm product even though a mildly unpleasant sensation was caused by the cooling effect of the spray. However, while aerosol deodorants were acceptable, they did not provide lasting protection against perspiration and body odor, and did cause irritation in the armpit, especially after shaving. Because it had long been recognized that the public preferred anti-perspirants to deodorants, the search by cosmetics manufacturers for an aerosol anti-perspirant intensified after the introduction of RIGHT GUARD aerosol deodorant. An aerosol anti-perspirant would be extremely successful, it was felt, because it combined the effectiveness of anti-perspirant creams and lotions with the ease and convenience of application of the aerosol. To avoid the corrosion problem, researchers looked for an anhydrous, i.e., water-free liquid medium in which to dissolve the aluminum salt. Special aluminum compounds designed to be soluble in alcohol were developed. While these formulations reduced corrosion, other difficulties, such as valve clogging, prevented any product from achieving widespread consumer acceptance. In addition to aerosol solutions, researchers experimented with other types of aerosol systems, in which small solid particles of aluminum chlorohydrate were suspended or dispersed in a liquid inside the aerosol package. Aerosol packages containing aluminum chlorohydrate particles suspended in a propellant and stabilized with small quantities of conventional suspending agents produced a spray of dry aluminum chlorohydrate powder. These systems effectively avoided can corrosion but experienced other problems, including settlement, valve clogging, billowing or dusting, and product inhalation. In addition, these early dry powder sprays were no more effective in stopping perspiration than aerosol deodorants. During this period, both Carter-Wallace and Gillette attempted, without success, to develop an effective, commercially acceptable aerosol anti-perspirant. Both believed that the breakthrough, when it came, would involve the solution technology described above. Against this background, George Spitzer, in 1964, began to take an active interest in the growing underarm products market. Since Gillette’s main defense is that the claimed invention was obvious, our analysis must begin with a review of the relevant prior art references. Our task is to determine what the prior art references would have revealed to a person of ordinary skill in the art at the time the claimed invention was made. PRIOR ART We find that the relevant art encompasses both anti-perspirant chemistry and the general field of aerosol technology, and includes patents as well as commercial and scholarly research. We set out below those references which we regard as most pertinent, and note that other references were also considered, but further discussion is sacrificed to the need to limit this opinion to a reasonable length. DeGiacomo Article In an article published in September, 1956 in a publication entitled. Drug and Cosmetic Industry, DeGiacomo described the efforts being made to apply aerosol technology as a means of dispensing finely divided powders. Some of these efforts, he indicated, involved the use of wetting agents to disperse the solid particles in a liquid medium. By contrast, DeGiacomo proposed the use of a suspending medium, that is, a liquid capable of suspending the powder without attacking the valve or other components of the aerosol package. DeGiacomo concluded that a purified grade of isopropyl myristate, an oily ester used widely throughout the cosmetics industry, had all the properties necessary to act as an ideal suspending agent for powdered products. For example, it was soluble in the conventional propellants, greaseless in character, capable of suspending powder formulations, and did not adversely affect the valve or other components. He further indicated that used in high concentrations in certain products such as liquid make-up, isopropyl myristate enhanced the functional properties of the finished product. “This higher concentration,” he noted, “creates a material with a higher residual nature allowing the solid ingredients to adhere in a given area of application.” Defendant’s Ex. 25-5, at 329 (emphasis added). Geary Patent The Geary patent, No. 3,088,874, issued May 7, 1963, describes a powder aerosol system capable of dispensing a composition containing at least 15 percent by weight active powder from a conventional aerosol package. The patent states that previously, increasing the amount of active powder beyond 10 percent resulted in serious problems of valve clogging, agglomeration of powder inside the can, and dusting or billowing. Geary teaches that, if a liquid surfactant, a bulking agent, a non-polar liquid, and a propellant are combined in a certain critical ratio by volume, any active powder having a diameter of at least 15 microns can be dispensed in quantities of at least 15 percent, and preferably at least 25 percent. The organic liquid, according to Geary, can be any non-polar organic liquid which is miscible in all proportions with the surfactant and propellant, reduces the cohesive force of the active powder particles, and is an electric insulator. Examples given include propyl myristate, butyl stearate, pentyl laurate, and others. The propellant can be any conventional propellant used in aerosol containers. The surfactant can be any nonionic, anionic, cationic, or amphylatic surfactant, including the sorbitan derivatives. The shape and size of the container and valve are not critical. The patent states that any conventional aerosol package can be used. Geary claims that the range of active powders which can be dispensed from an aerosol system manufactured in accordance with the teaching of the invention is practically unlimited, the only limitation being the maximum particle size of the powder. A typical anti-perspirant formulation appears in the patent as Example 5, and contains 25 percent by weight aluminum chlorohydrol and 8 percent by weight non-polar oil. Finally, Geary states that the presence of conventional amounts of other ingredients, such as perfume, does not affect the powder aerosol system. Yakubik Article In July, 1964, John Yakubik published an article in Drug and Cosmetic Industry entitled “Formulation of Aerosols.” The article describes the general function and composition of aerosol products, as well as the major types of aerosol systems, i.e., suspension, dispersion, solution, and separate-phase systems. Yakubik indicates that one of the advantages of suspension-type aerosols is their stability, that is, the ease with which chemical reactions among the ingredients can be kept to a minimum. He states: As a matter of fact, it is not unusual to encounter the same degree of stability in an anhydrous suspension aerosol that one expects of the active material in the dry state. Defendant’s Ex. 25-7, at 38. Two of the greatest problems with suspension formulations discussed by Yakubik are valve clogging and leakage. He suggests two ways of dealing with these problems: (1) reducing the particle size of the active material; and (2) using a non-volatile liquid diluent, such as mineral oil or alcohol. One of the advantages of using a diluent, Yakubik claims, is that it acts as a carrier for the active ingredient. In the case of high potency material the diluent increases adherence to the skin. In addition, he states that the diluent may act as an auxiliary dispersing agent and valve lubricant. Yakubik teaches that mineral oil, isopropyl myristate, isopropyl palmitate, alcohol, and certain other materials are all excellent diluents. Regarding the proper amount of such ingredients in the aerosol package, he states: The concentration of non-volatile vehicles in the total aerosol is a matter of personal preference. High concentrations of nonvolatile vehicles give wetter sprays. Alcohol in 10 per cent or higher concentrations gives a cooling sensation to the spray on the skin; oils may give an oily feel on the skin especially when the powder concentration is low. Id. Thiel Patent The Thiel patent, No. 2,169,095, granted February 9, 1965, claims an improvement generally applicable to self-propelling, powder-dispensing compositions consisting of a finely divided powder suspended in a liquefied propellant and a liquid, non-ionic surface active agent, i.e., an oil. Although not specifically concerned with aerosol anti-perspirant compositions, Thiel states that the active component of the formulation may be an anti-perspirant such as aluminum chlorohydrate. Briefly, the improvement consists of adding a higher alkyl halide quaternary salt, such as pyridine, to the composition. Thiel acknowledges that one problem in the preparation of compositions consisting of a finely divided powder suspended in a liquid is the tendency of the powder to agglomerate and form deposits on the walls of the container. Thiel believed the problem was due to a mutual repulsion of individual particles. In a prior patent, No. 3,014,844, it was revealed that the agglomeration problem could be somewhat reduced by adjusting the specific gravity of the liquid components of the composition. In the improvement patent, Thiel claims that adding a higher alkyl quaternary salt in an amount between .01 and .1 by weight further reduces the tendency toward surface caking, a problem not addressed in the prior art. The patent states that a liquid, non-ionic surface active agent should be present in a concentration between .1 and 20 percent by weight, preferably between about .25 and 1 percent. The finely divided powder can range from .01 to 20 percent by weight, preferably between .1 and 3 percent. Thiel provides a dozen detailed examples of formulations which were represented to be suitable, with varying proportions of the active agent powder, the non-volatile liquid and the pressurized propellants. The ratio of solid powder to non-volatile oil in these examples ranges from a low of 0.7 to 1 in Example 5 (excess oil) to a high of a little over 1.5 to 1 in Example 11 (excess powder). Shulton Patent Shulton British patent No. 987,301, published March 24, 1965, describes the invention of an aerosol composition consisting of an active ingredient, a liquefied gaseous propellant, and colloidal silica in an amount sufficient to render the entire composition a substantially homogeneous gel. The stated purpose of the invention is to provide a stabilized, self-propelling aerosol composition which is of uniform and homogeneous consistency while in the container. Among the problems Shulton purports to alleviate are clogging, caking, billowing, and, in the case of anti-perspirants, can corrosion. With respect to the corrosion problem, Shulton states: Aqueous solutions of the well-known aluminum hydroxychloride (chlorhydrol) are of the ... corrosive type and attack conventional metal containers producing undesirable reaction products and reducing the effectiveness of the aluminum salt. Accordingly, the application of aluminum hydroxychloride to the skin as a very finely divided powder, particularly as an aerosol, would be highly desirable since this salt will readily dissolve in perspiration and become as effective an anti-perspirant as if it had been applied to the skin as an aqueous solution. Defendant’s Ex. 25-3, at 3. The primary teaching of Shulton, at least insofar as relates to the instant case, is that a finely divided aluminum anti-perspirant salt could be effectively dispensed as an aerosol, without significant problems of valve clogging, caking, billowing, or can corrosion. This result was obtained by combining the aluminum salt and a liquid propellant with a gelling agent, or colloidal silica, such as Cab-O-Sil, in an amount ranging between 0.1 and 10 percent by weight. The patent notes that the higher the quantity of finely divided powder, the less colloidal silica is required to produce a satisfactory result with a given liquefied propellant. Shulton also teaches that in addition to the essential components of the powder aerosol composition, the product may contain a so-called powder lubricant, such as zinc stearate or an oily ester; a perfume in alcohol; and a skin conditioner or lubricant, such as isopropyl myristate. These components, the patent states, are included to facilitate or supplement the functioning of the active ingredient when applied to the spray surface, or to facilitate the proper functioning of the valve in the pressure tight container. The following is an example of a typical aerosol anti-perspirant formulation prepared in accordance with the teaching of Shulton: Percent by Ingredient Weight Aluminum Hydroxychloride 6.0 Powder Colloidal Silica 2.7 (Cab-O-Sil) Isopropyl Myristate 1.0 Powder Lubricant 0.8 (Brij 96) Hexachlorophene 0.2 Perfume 0.5 Specially Denatured 2.8 Alcohol Propellant 86.0 Boyle Patent The Boyle patent, No. 3,218,263, granted November 16, 1965, purports to solve the problems, such as valve clogging, associated with the dispensing of dry powder from an aerosol package. The key to the invention is the inclusion in the aerosol package of a silica-containing thickening agent, such as Cab-O-Sil or bentonite, in a concentration ranging from 4 to 55 percent by weight, which, when combined with a liquefied propellant, forms a homogeneous gelled composition inside the aerosol package. Boyle claims that, among the many advantages of the patented compositions is the ability to dispense a “substantially liquid-free spray.” Defendant’s Ex. 25-9, at col. 4. The patent expressly states that the claimed invention may be used in the preparation of aerosol anti-perspirant formulations. Specialties Article An article published in December, 1964 in a publication entitled Specialties provides a brief but excellent summary of the state of the art with respect to aerosol anti-perspirant technology at the time. The article describes the problems of product breakdown, corrosion and valve clogging associated with the introduction of water into an aerosol package containing aluminum chlorohydrate salt. It states that the problem of clogging can be minimized by the addition of a relatively high proportion of a humectant, such as glycol, but that excessive amounts prevent the anti-perspirant from drying on the body, leaving a tacky film. The article indicates that the problems of solubility, hydrolysis, and crystallization are not present when the aluminum chlorohydrate is utilized in the dry state. It states: The aluminum salt can be extended with a suitable talcum powder and the propellants, which of course are in the liquid state inside the container, act as the carrier. A valve suitable for dispensing powders must be used. Since water is excluded from the contents, .no corrosion occurs inside the containers. Plain aluminum or tinplate containers may be used. Defendant’s Ex. 25-1, at 18. The article provides the following example of an aerosol anti-perspirant powder formulation: Percent Ingredient by Weight Talc 9.0 Aluminum Chlorhydroxide 1.0 Complex (Impalpable) Isopropyl Myristate 0.3 Perfume 0.3 Propellants 89.4 Id. Goldberg Patent The Goldberg patent, No. 3,288,681 was issued on November 29, 1966, effective May 1, 1962. The patent relates to the development of an anhydrous aluminum anti-perspirant powder aerosol composition. It teaches that dispersion technology can be used to manufacture such a composition containing the following ingredients: (1) a finely divided aluminum compound, such as aluminum chlorohydrate, in an amount between 4 and 10 percent by weight; (2) an ester, such as isopropyl myristate-palmitate, in an amount ranging from 2 to 5 percent by weight, to facilitate dispersion of the powder; (3) commercial alcohol, in an amount between 2 and 4 percent by weight; and (4) a liquefied, gaseous propellant such as dichlorodifluoromethane. Other, nonessential ingredients include a non-irritating germicide and a perfume. Goldberg claims that the patented invention makes it possible to produce an antiperspirant aerosol powder spray. The success of the invention, the patent states, requires strict adherence to what Goldberg describes as the critical limits of the essential components and the required processing conditions. Too great an amount of aluminum powder, he states “is too dusty when sprayed and clogs the aerosol valve; and a smaller amount is relatively ineffective as an antiperspirant.” Defendant’s Ex. 25 — 4, at col. 1. Too much oil on the other hand, produces “a powder film which is too wet” while too little “provides little observable benefit.” Id., at col. 2. CALM An aerosol anti-perspirant powder was being sold in the United States as early as 1964 by Alberto-Culver under the name “CALM.” CALM contained dry powder aluminum chlorohydrate, talcum powder, an oil, a propellant, and several additives. The ratio of aluminum powder to oil in the CALM formulation was approximately one-to-one, i.e., 2.7 percent aluminum chlorohydrate versus 2.475 percent oil by weight. Tests performed by cosmetics manufacturers as well as Hilltop Research, Inc., acknowledged as the industry leader in anti-perspirant efficacy testing, revealed that CALM was no more effective in reducing perspiration than Gillette’s popular RIGHT GUARD aerosol deodorant, achieving an average sweat reduction of 6 percent. See n. 3 above. While there was some contrary evidence in the form of testimony from Gus Kass, a former vice president of Alberto-Culver, that in-house “T-shirt tests” showed that CALM was effective, the more persuasive evidence indicates that CALM, as originally formulated, was ineffective as an anti-perspirant, achieved little commercial success, and was eventually removed from the market in that form. Virzi Patent The Virzi patent, No. 3,472,928, issued October 14, 1965, but effective as of September 23, 1965; relates to a liquid antiperspirant composition which can be sold as an aerosol without causing undue can corrosion. “So far as is known,” the patent states, “an effective amount of astringent has not yet been incorporated into a liquid cosmetic composition suitable for packaging in a metal container or can and having an acceptable shelf life.” Defendant’s Ex. 16-8, at col. 2. Virzi teaches that the corrosion problem can be eliminated by dissolving the astringent in an anhydrous, volatile liquid solvent, preferably ethanol alcohol. The solvent must be sufficiently anhydrous to prevent an appreciable dissociation of the astringent salt, yet sufficiently volatile to vaporize within a short period of time after application, leaving the astringent on the skin ready to be activated by moisture in the atmosphere and on the body. The patent also describes a method of preparing the claimed invention. A typical formulation appears below: Percent by Weight Ingredient (propellant excluded) Aluminum Hydroxychloride 5 Lanolin 2 Propylene Glycol 2 Hexachlorophene 0.4 Felton Aphrosia 0.1 Ethanol 90.5 Id., at col. 3. In the above example, the lanolin functions as an emollient and gives soothing qualities to the composition. Propylene Glycol is used as a humectant; hexachlorophene is a bacteriostat; and felton aphrosia is a perfume. The patent states that any of the conventional propellants may be used. The evidence in the case indicates that the “Virzi approach,” i.e., an anhydrous solution in which the aluminum salt is dissolved in a volatile liquid such as alcohol, dominated the thinking of those skilled in the art as to how to solve the aerosol antiperspirant dilemma during the early and mid-1960’s. For example, plaintiff’s Ex. 85, the so-called “Prussin Art Collection,” includes some 25 references in the area of anti-perspirants and aerosol anti-perspirants. To the extent they address the problem, these references unanimously teach that an alcohol or hydro-alcohol based solution was the best method of formulating an effective aerosol anti-perspirant composition. THE TEACHINGS OF THE PRIOR ART A person of ordinary skill in the anti-perspirant industry in the mid-1960’s was aware of the efforts being made to develop an effective, commercially acceptable, aerosol anti-perspirant. He knew that most industry experts believed the key to the problem lay in solution technology generally, and anhydrous solutions in particular. However, being familiar with the literature in the field of aerosol technology, he would (or should) have read and understood the references described above, and would have known that other approaches to the problem had been suggested. One such approach involved the dispersion or suspension of finely divided aluminum chlorohydrate particles in a liquefied propellant or other suitable suspending medium. This use of aluminum powder in the “dry state” effectively eliminated the corrosion problem, but was associated with other problems, such as agglomeration or caking, valve clogging, and billowing or dusting. Perhaps most significantly, the skilled person would have known that early formulations of powder aerosols failed to achieve significant consumer acceptance, largely because they were ineffective anti-perspirants, i.e., they failed to achieve a significant degree of sweat reduction. This ineffectiveness may have been due to the failure of the anti-perspirant composition, which was dispensed as a dry powder, to adequately adhere to the underarm in sufficient quantities to effectively inhibit perspiration. Consulting the cited references, a person of ordinary skill would have learned that the inclusion of relatively small quantities of a suspending agent, particularly colloidal silicas such as Cab-O-Sil or bentonite, in the aerosol formulation would help disperse the aluminum chlorohydrate particles and reduce agglomeration and valve clogging. He would also have learned that the use of small amounts of oil, such as isopropyl myristate, would facilitate dispersion, lubricate the valve and prevent clogging, and impart certain desirable cosmetic properties to the skin. In greater concentrations, the oil would facilitate adherence of the anti-perspirant powder to the skin and reduce billowing, thereby making it possible to increase the concentration of aluminum ehlorohydrate in the formulation. He would have learned from Yakubik that the relative concentrations of powder and oil in aerosol compositions is largely a matter of personal preference and that, in general, the use of oil in higher concentrations would produce a wet, oily spray. He would have been aware that the use of oil in concentrations as high as 20 percent by weight had been contemplated by Thiel, and the use of isopropyl myristate in unspecified quantities as the primary suspending medium had been recommended by Yakubik. On the other hand, he would have been aware of the general feeling among industry experts that large quantities of oil in an anti-perspirant formulation would inhibit activation of the aluminum chlorohydrate, thereby reducing anti-perspirant efficacy. Against this backdrop of the prior art and teachings of the prior art, we turn to the work of Spitzer and his colleague, Osipow. DEVELOPMENT OF THE SPITZER FORMULATION George Spitzer graduated from St. John’s University in 1935 with a degree in Pharmacy. After finishing college, Spitzer devoted his career exclusively to the development of new products. Among the products he developed were a laxative in the form of a jam, a series of dietetic health foods, and an aerosol shave cream sold in the United States by Carter-Wallace under the name “RISE.” Most of the laboratory research in connection with the RISE project was performed by a research and development organization by the name of Foster D. Snell Laboratories (Snell). In late 1964, Spitzer instructed Snell to develop an aerosol anti-perspirant. The most important property of the product he wanted was efficacy. As Spitzer indicated in his deposition: I wanted it to be effective, that it would stop perspiration as good as any product on the market, like Arrid Cream, or Arrid Roll-On or any other antiperspirant presently being sold. So effectiveness was first. Defendant’s Ex. 21, at 7. The person in charge of the project at Snell was Lloyd Osipow. Osipow graduated from Columbia University with a degree in chemistry in 1939. He began working for Snell in 1946, eventually becoming a Research Group Director in the areas of cosmetics, toiletries, and detergents. Osipow had been involved in the development of aerosol products since joining Snell, and had a limited role in the RISE shaving cream project. He had also worked for Spitzer on an aerosol product containing talc. Osipow was thoroughly familiar with the technical literature in the field of aerosol products and kept up to date while working on the latest Spitzer project. As group director, Osipow supervised a number of other chemists who performed the actual experiments and analyses. Osipow described the problems encountered in the search for an acceptable formulation. These included coldness on the skin; dustiness and mistiness; failure to stick to skin; corrosion of metal containers; and clogging or malfunction of the valve. Osipow did not consider that he had solved the coldness problem, and stated that the problem of dustiness or mistiness has never been completely solved. He avoided the corrosion and clogging problems by packaging the anti-perspirant powder in an anhydrous liquid suspension, and used a colloidal silica as a suspending agent. He frequently discussed the progress of the work with Spitzer, and gave him numerous sample formulations which Spitzer tested personally. After working on the project for about a year, Spitzer and Osipow came up with a product which they both agreed was effective and relatively problem-free. The product consisted of a composition containing less than 10 percent by weight aluminum chlorohydrate powder suspended in an essentially anhydrous solution of a hydrophobic, non-volatile oil and a conventional propellant. Provided that the amount of hydrophobic oil was at least as great as the amount of salt in the package, they found that the suspension of aluminum chlorohydrate in oil could be released through the valve and deposited on the underarm as a wet spray, where it would be activated as an anti-perspirant by moisture in the air and on the body. Neither Spitzer nor Osipow quite understood how an aerosol anti-perspirant containing the amount of oil contemplated by their formulation could function effectively. As Spitzer testified: Q. How did the moisture on the skin or in the air get to the aluminum chlorohydrate if it were dispersed in this anhydrous oil? A. I don’t know. It seemed to get on the skin. And the moisture seemed to do its work. It seemed to get to the aluminum chlorohydrate and make it active. Defendant’s Ex. 21, at 23. Nor could Spitzer explain what effect, if any, a change in the amount of oil would have on the efficacy of the product: Q. Does the moisture go through the film of non-volatile, non-hygroscopic liquid just as rapidly no matter how thick it is? A. I can’t answer that, but I can say that the amount that we finally selected, it apparently did. In other words, the ratio that we selected finally for our product of aluminum chlorohydrate and isopropyl myristate, apparently the moisture was sufficient to go through that film and cause anti-perspirant action. Id. at 69. Q. Would it work just as well if it were oily, even though the customer didn’t like it? A. I don’t know that. Id. at 15. Both Spitzer and Osipow acknowledged that changing the amount of oil, or varying the ratio of oil and aluminum chlorohydrate powder would affect the wetness of the spray. According to Spitzer: The wetness of the spray is due to the amount of non-hygroscopic, non-volatile material, liquid that’s in there. The oil. That’s what makes it come out as a wet material, carrying the aluminum chlorohydrate. And we wanted enough in there so as it kept it dispersed until it was emitted, and not too much so that it came out too wet or too oily. So that was trying to find an optimum amount between the two. Id. at 30. Both inventors conceded, however, that the change in wetness corresponding to a change in the amount of oil was a gradual one, and that the so-called “optimum amount” was largely a matter of personal preference. Spitzer testified: Q. Were the relative amounts of aluminum chlorohydrate and of the non-volatile, non-hygroscopic liquid important? A. It was to us, yes, at that time. Q. Why? A. Well, we didn’t want the product to be overly powdery. We didn’t want it to be overly oily. Q. Why didn’t you want it to be overly oily? A. Because it would feel unpleasant on the skin. Q. It’s a matter of popularity then? A. A matter of the consumer finding it not very desirable. Id., at 14-15. As to the required amount of aluminum chlorohydrate powder in their formulation, the inventors stated they wanted to include enough to provide effective anti-perspirant activity, but not so much that the product would cause excessive billowing or dusting when sprayed. Spitzer testified: Q. If you have the non-volatile, non-hygroscopic liquid there in sufficient amount, does it make any difference how much aluminum chlorohydrate is present? A. Yes. Q. Why? A. Well, it shouldn’t have any more than what the patent says. It shouldn’t have any more than would cause excessive dusting. Too great an amount would cause excessive dusting when sprayed against the skin in general. Generally the astringent is in an amount .2 to 10 percent. Id., at 39-40. Q. Why did you want it not to be powdery? A. Well, if it was powdery, again the consumer would not like it. It would be rolling off the under armpit. It would fly all over the air. It would be inhaled and breathed in the room, and I think it would make it very unpopular for the consumer. Q. You decided that you wanted it not to be too oily and not to be too powdery just on the basis of your own experience with aerosol — with anti-perspirants? A. I think so, yes. Id., at 15. Neither Spitzer nor Osipow explained how the amount of aluminum powder specified in the invention and, later, in the patent was arrived at, or the effect a variation in the amount of powder either within or outside the specified limits would have on the efficacy of the composition or its relafive dustiness. In fact, neither inventor could recall how the numerical limits of aluminum powder in the formulation were determined. Spitzer testified: Q. ... do you still get dustiness if you have say 10 per cent by weight of aluminum chlorohydrate, even though you have an equal amount of non-hygroscopic liquid present? A. I don’t know if we ever did that at all. We didn’t do any experiments other than those that would make a commercial product. Q. How did you know that 10 per cent was the upper limit? A. I guess we must have tried it out in the laboratory to see what the lower limit is and what the upper limit is. Q. And you did find then, that if you exceeded the upper limit, you had dustiness, is that right? A. I guess that’s right, Mr. Furlong. Id., at 40-41. Osipow testified: Q. Are you the one that came up with these numbers, these ratios? A. I do not recall that. Q. Do you recall whether anyone else came up with them? A. I don’t recall. Q. Were they based on any of the work you had actually done? A. I am sure they are based in part at least on the work that we have done. Q. Would it be correct to say that they were educated guesses based on the work that you have done? A. I don’t know. I would have to review the work. Q. Have you seen any thing in the work which you have been reading for the last day or so which helps you to answer that question. A. No, I have not. Defendant’s Ex. 17, at 164-66. This testimony of Spitzer and Osipow suggests that their discovery was not the result of inventive genius, but rather patient experimentation and trial-and-error. The significance of this will be discussed in greater detail below. For the moment, we note only that those experiments resulted in a relatively problem-free aerosol formula which proved remarkably effective in reducing perspiration. DEVELOPMENT OF ARRID EXTRA DRY Anthony Parisse was hired as a research chemist by Carter-Wallace in 1965 and instructed to develop an aerosol anti-perspirant. Parisse experimented with many different aerosol systems but, like most of the experts in the field at that time, was primarily interested in solution technology, i.e., aerosol compositions in which the active aluminum chlorohydrate powder was dissolved in either anhydrous alcohol or a mixture of water and alcohol. Late in 1965, Parisse became aware of the so-called “Spitzer technology.” Parisse was initially skeptical about the product, for two reasons. First, he knew that to be effective, aluminum chlorohydrate had to be dissolved in water. Parisse was concerned that the presence in the formulation of significant amounts of oil would prevent the aluminum salt from dissolving, thereby inhibiting anti-perspirant efficacy. Second, Parisse was aware that several dry powder formulations, including the CALM product discussed earlier, had proven relatively ineffective as anti-perspirants and had not achieved significant commercial success. He assumed that increasing the concentration of hydrophobic oil in what was essentially a powder formulation could only result in a less effective end product. Nevertheless, at the direction of Carter-Wallace’s management, Parisse had the Spitzer formulation tested on several occasions, and found it a remarkably effective aerosol anti-perspirant. Carter-Wallace immediately embarked on a crash program to develop a commercial product utilizing the Spitzer technology. The result was ARRID EXTRA DRY (ARRID). ARRID was introduced into the retail market in 1967, and gained immediate consumer acceptance. Sales in 1968 were approximately $20 million, as ARRID captured approximately 10 percent of the underarm products market. By 1972, annual sales of ARRID climbed to nearly $65 million, and ARRID’s market share reached approximately 16 percent. Advertising support for ARRID during these early years was substantial. During the 5 year period from 1967 to 1972, Carter-Wallace spent nearly $50 million on advertising. This represented approximately one-fourth of the total of all advertising spent by the entire anti-perspirant/deodorant industry during that period. The evidence indicates that, while other new products have been similarly promoted in the past, ARRID was the most heavily advertised new underarm product the industry had seen in over twenty years. The ARRID products sold during this period were based on the Spitzer technology. While a number of different formulations were produced and sold, some 90 percent of the ARRID products sold as of June 4,1969 contained the following formulation: Percent by Ingredient Weight Aluminum Chlorhydrate 3.50 (impalpable) Isopropyl Myristate 6.08 Bentone 34 0.26 Perfume 0.16 Propellant 90.00 Defendant’s Exs. 5 and 15, at C-56. The evidence indicates that ARRID was an effective anti-perspirant, achieving an average sweat reduction of 35 percent. See n. 3 above. PROCEEDINGS IN THE PATENT OFFICE The initial or “parent” application for a patent on the Spitzer formulation, Serial No. 492,268, was filed October 1, 1965. On December 16, 1965, Spitzer and his financial partner, Marvin Small, sold all of their rights in the invention to Carter-Wallace, which thereafter handled the proceedings in the Patent Office. The initial application was completely rejected on the grounds of overbreadth, vagueness, and obviousness. In response, Carter-Wallace on June 30, 1967 submitted a series of amendments and a brief in which it attempted to distinguish the prior art references cited by the Examiner. The amended application was rejected by the Examiner on the same grounds. The Examiner also concluded that the Spitzer formulation was “fully met” by Shulton, which, he noted “discloses aerosol antiperspirant compositions which contain aluminum chlorhydroxide powder.” Defendant’s Ex. 15, at P-31. A subsequent amendment filed December 20, 1967, was similarly rejected as anticipated by Geary and obvious in view of Shulton. On June 3, 1968, counsel for Carter-Wallace met with the Examiner and discussed the Shulton and Geary patents at length. On June 11, Carter-Wallace filed a petition to revive the application which was allowed. In a series of supplemental briefs, patent counsel sought to distinguish Geary as involving an aerosol powder system and Shulton as involving both a powder system and an emulsion, whereas the Spitzer formulation was described as an anhydrous liquid suspension. In response to the Examiner’s request, a comparative study involving products prepared in accordance with Spitzer, Geary and Shulton was conducted and the results reported to the Examiner, together with actual samples of the respective products. The Examiner rejected the results of the comparison study and rejected Carter-Wallace’s supplemental amendments on the grounds previously stated. After this rejection, the Patent Office refused to enter subsequent amendments filed June 17 and July 31, 1969. In August, 1969, Carter-Wallace filed a new set of claims. Counsel acknowledged that the specifications of the new application, Serial No. 853,626, were identical to those of the parent application, and the new claims were directed to the same invention. The effect of the new filing, however, was that review of Carter-Wallace’s previously rejected application was continued. While the continued application was pending, the Virzi patent was approved. In a letter filed January 5, 1970, counsel for Carter-Wallace sought to distinguish Virzi on the ground that it taught the use of volatile alcohol as the carrier for the astringent, whereas the Spitzer formulation expressly cautioned against the use of volatile carriers, teaching instead the use of nonvolatile oils. In his next action, the Examiner on September 22,1970 rejected most of the claims of the continued application as unduly broad, and obvious over Geary. Notably, however, the Examiner did not object to the claims as unpatentable over Virzi, and withdrew his previous rejection based on Shulton. Presumably, the Examiner was persuaded that the absence of alcohol from the Spitzer formulation rendered these references— both of which taught the use of anhydrous ethyl alcohol — inapposite. The examiner did, however, cite Virzi and a number of other references “to show the state of the art and to make the instant record complete.” Id. at C-147-99. Thereafter a series of amended claims were rejected on July 23, 1971. While the application was pending before the Examiner, applicants’ corresponding application in the United Kingdom was published. The grant of a patent was opposed by several parties on the grounds that the subject matter was not new or alternatively that it was obvious, relying on numerous earlier patents and publications in support of their position. Some of these, such as the Yakubik article discussed above, were arguably more relevant than references cited by the Examiner in the United States. In a statement acknowledging their obligation to be completely frank and candid with the Patent Office Carter-Wallace in April, 1972 submitted a list of some thirty-eight references from the British opposition which, it indicated, the Examiner may have considered to be of some relevance. Several months later, Carter-Wallace followed up the earlier list by submitting copies of the references in two bound volumes. At the time, it stated: With regard to these additional references, it should be noted that these foregoing references are, at best, merely cumulative. None of these references is believed to be any more pertinent than the references now of record and such references certainly do not disclose or suggest applicants’ claimed invention. Defendant’s Ex. 15, at C-234. The continued application was finally rejected on January 18, 1973. In rejecting a further amendment filed after that date, the Examiner stated that the evidence relied upon by Carter-Wallace to distinguish Geary showed nothing more than a difference in degree as opposed to a difference in kind between the Geary composition and the Spitzer composition. Id. at C-308-09. On June 18, 1973, Carter-Wallace filed a notice of appeal from the final rejection of its application. After a hearing on February 4, 1976, a three member Board of Appeals on March 11, 1976 reversed the decision of the Examiner. The Board considered only Claims 1, 2, 3, 5 and 8 of the application, and stated the sole issue on appeal as the propriety of the Examiner’s rejection of the claims over Geary. The Board was persuaded by Carter-Wallace’s argument that the Spitzer formulation was not obvious over Geary, because the latter was a powder aerosol while the former was a liquid suspension. In addition, the Board noted that the numerical limits and ratios of ingredients in Geary were described as “critical” to that composition, and those limits and ratios were different from the ones taught by Spitzer. The Board disagreed with what it regarded as the Examiner’s implied conclusion that, since the same ingredients were present in both the Geary and Spitzer formulations, the same result would necessarily obtain when one combined those ingredients in different proportions, stating: The basic weakness in the Examiner’s position quite clearly is that although the same components are present in the system of Geary et al as are in the claimed system they are, nevertheless, not present in the same proportions. sjt * ‡ Jfc 5f! sf: Quite clearly, therefore, it cannot be said that the proportions of ingredients taught as critical by Geary et al to obtain a powder aerosol system make it prima facie obvious to use the claimed amounts of components and which when expelled from the aerosol bomb are effective in the form of a liquid suspension of the astringent powder; this is clearly a composition and result not suggested by the reference disclosure. Id. at C-414-15. However, the Board said nothing about the Examiner’s finding that the Spitzer formulation itself lacked criticality, i.e., that it produced a result different only in degree from the prior art. Finally, the Board was influenced by evidence of secondary considerations presented by Carter-Wallace in reaching the conclusion that the invention was not obvious, noting that: [T]he claimed package, an embodiment thereof being sold under the proprietary trade name of “Arrid Extra Dry” by appellants’ assignee (Carter-Wallace, Inc.), has achieved extreme success in the antiperspirant market giving rise to the manufacture and sale of many competing products of similar composition. Such can clearly be viewed as indicia of non-obviousness, the invention also satisfying a long-felt need and solving problems as to which many attempts had earlier been made. Id. at C-415 (citations omitted). Following a remand for consideration of certain patents referred to for the first time on appeal, the Patent Office issued a Notice of Allowance of the application on April 28, 1976. THE PATENT IN SUIT The patent in suit, United States Patent No. 3,968,203, as finally issued July 6, 1976, attached as Appendix A, relates to a novel astringent and anti-perspirant aerosol system, capable of dispensing a liquid spray of astringent from a conventional aerosol bomb. Prior to the invention, the patent states that there had not been a truly commercially acceptable aerosol anti-perspirant composition that could be packaged in a conventional metal aerosol container. The description points out that previously, anti-perspirants were effective in aqueous solution, but that aqueous solutions in metal aerosol containers caused a corrosion problem, and that crystallization of the solution in the valve orifice caused valve malfunction. It acknowledges prior aerosol packages containing the astringent in powder form, but asserts that they were unsatisfactory because they produced clouds of dust breathed by the user. The stated objects of the patent are to provide an improved anti-perspirant aerosol formulation which: (1) causes minimum internal container difficulties and improved results with respect to corrosion and valve malfunction; (2) is applied to the skin as a wet spray in the form of a liquid suspension of astringent material, the astringent material being activated by the moisture of the air and the body; (3) imparts a feeling of lubrication and reduces the likelihood of inhalation problems; and (4) minimizes valve clogging and leakage. The patent asserts that the above objectives can be realized by formulating an aerosol anti-perspirant composition consisting of an astringent material in fine powder form suspended in a liquid vehicle which is essentially anhydrous, the liquid vehicle containing as essential ingredients a liquefied propellant and a non-volatile, non-hygroscopic liquid. The patent specifically teaches that the use of a volatile component, such as alcohol, instead of a non-volatile component will not give suitable results. However, nothing in the patent expressly prohibits the inclusion of a small amount of alcohol not affecting the overall function and properties of the composition. The active anti-perspirant agent described in the patent is not limited to aluminum chlorohydrate but includes other salts of various metals said to be useful as astringents in powder form. However, the claims are restricted to aluminum chlorohydrate. The patent describes in scientific terms the non-volatile, non-hygroscopic liquid suitable for use in the composition, including among the examples organic esters such as isopropyl myristate. Any conventional propellant meeting certain general criteria specified in the patent can be used. Both the specifications and claims set forth the numerical limits of the essential ingredients and the reasons for those limits. These may be summarized as follows: (1) The astringent should be in an amount sufficient to impart the desired anti-perspirant activity when activated by the moisture in the air but should not be in too great an amount to cause excessive dusting when sprayed against the skin. In general, the astringent is in an amount from .2 to 10% and preferably .5 to 5% by weight of the total composition, and in an amount from .01 to 1.0 part by weight and preferably .05 to .3 part per part of the non-volatile, non-hygroscopic liquid. (2) The non-volatile, non-hygroscopic liquid should be in an amount sufficient whereby the astringent is deposited on the skin in the form of a wet spray but should not be in too high an amount wherein it would interfere with the anti-perspirant action of the astringent as well as causing the wet spray deposited on the skin to be of too oily a nature. In general, the non-volatile non-hygroscopic liquid is in an amount from 2 to 50% by weight of the total composition, preferably 5 to 25% by weight. (3) The propellant should be in an amount to provide the desired vapor pressure for dispensing the desired contents from the container without causing excessive dusting when the wet spray containing the dispensed astringent is applied against the skin. In general, the propellant is in an amount from 50 to 98% by weight of the total composition, preferably 75 to 95% weight. Appendix A, at col. 4 and Claims 1 and 2. The patent next states that “it has been found desirable” to incorporate in the aerosol composition of the invention materials having “grease imparting properties” in order to act as thickening or gelling agents for the non-volatile liquid and aid in preventing valve malfunction. Id. at col. 4. The expression “grease imparting agent” as used in the patent is merely the inventors’ lexicon for a suspending agent. Examples of such materials recited in the patent include hydrophobic clays such as bentonite, colloidal silicas such as Cab-O-Sil, and grease forming soaps such as aluminum stearate. As used in the patent in suit, “bentonite” is a generic term which, to a person skilled in the art would include modified forms such as Bentone 34. The patent states that, in general, these agents should be present in an amount “up to 1% by weight of the total composition and preferably 0.05 to 2% by weight.” Id. at col. 4-5 and Claim 5. The patent indicates that small amounts of certain other ingredients, such as perfume and a bacteriostat, may be included in the composition if desired. Finally, the patent provides six “working examples” of compositions intended to illustrate the claimed invention. See Appendix A, at col. 5. THE DEVELOPMENT OF RIGHT GUARD Prior to the introduction of ARRID, Gillette was also working feverishly to develop a commercially viable aerosol anti-perspirant. It also subscribed to the view that solution technology would produce favorable results, and directed its research efforts toward reducing or eliminating the problems, such as can corrosion and valve clogging, which until that time, had prevented the development of a successful solution-based product. During this period Reheis Chemical Company, a leading producer of anti-perspirant chemicals, spent approximately $1 million developing and promoting an alcohol-soluble aluminum salt which it sold under the tradename “Rehydrol.” The evidence indicates that Reheis received a preliminary order for Rehydrol from Gillette early in 1963, but that Gillette later cancelled the order. Early in 1967, Gillette obtained and tested samples of ARRID used in a test market offering. Its analysis of the product left Gillette skeptical about the product’s efficacy, as reflected in the following excerpt from a “Competitive Products Review” prepared for management in March, 1967: The antiperspirant effectiveness has not been determined, but the amount of aluminum salt found in the product should contribute some anti-perspirant action, unless their emollient and/or emulsifier system is interfering with activity as sometimes happens. Plaintiff’s Ex. 48. By the spring of 1968, the commercial success of ARRID was widely known, and top management at Gillette directed the Toiletries Division to develop a competitive product utilizing suspension technology. A number of formulations were prepared, some of which were acknowledged by Gillette to be “virtually identical (except for fragrance)” to the ARRID formulation. Plaintiff’s Ex. 59. In its haste to get a competitive product on the market, Gillette avoided extensive product testing. As a result, the first RIGHT GUARD anti-perspirant formulations were available by August of 1968. RIGHT GUARD anti-perspirant was quickly accepted by the public. In 1969, sales of RIGHT GUARD were approximately $20 million. By 1972, annual sales were nearly $38 million, and RIGHT GUARD captured approximately 8 percent of the underarm products market. In 1969, Gillette also introduced “SOFT’N DRI,” a product similar to both ARRID and RIGHT GUARD designed especially for women. Between 1969 and 1972, SOFT’N DRI achieved annual sales of approximately $26 million and captured a 6 percent share of the market. Together, Gillette’s two suspension-type aerosol anti-perspirants enjoyed an aggregate market-share of approximately 14.6 percent in 1972. In the meantime, sales of RIGHT GUARD deodorant dropped from a high point of 25.6 percent market share to an 11.4 percent share in 1972. The Gillette formulations alleged to infringe the patent in suit are set forth as numbers 1, 4-8, 11, 12, 14-17, 19, 20 and 21 in Plaintiff’s Ex. 98A. The evidence indicates that these formulations are strikingly similar to those sold by Carter-Wallace, and the examples provided in the patent in suit. All of the accused Gillette formulations contain aluminum chlorohydrate astringent in fine powder form. The astringent is in an inactive state when dry but is corrosive to the metal of containers and has anti-perspirant properties only when dissolved in water. Each of the formulas includes an essentially anhydrous liquid vehicle in which the astringent powder is suspended and dispersed. The vehicle in each of the compositions consists essentially of a liquefied propellant which is gaseous at room temperature and atmospheric pressure and non-reactive with the other components in the container, and a non-volatile, non-hygroscopic, non-polar organic liquid which meets the technical specifications of the liquid claimed in the patent in suit, see n. 20 above. In each of the accused products, the astringent powder and non-volatile, non-hygroscopic liquid are dispensed together from the container in the form of a liquid suspension upon the opening of the valve so as to be applied to the skin in the form of a wet spray. In each of the accused products the astringent is present in amounts from 0.2 to 10 percent by weight of the total composition, the non-volatile, non-hygroscopic liquid is present in an amount from 2 to 50 percent by weight of the total composition, and a propellant is present in an amount from 50 to 98 percent by weight of the total composition. In each of the accused products there is at least as much non-volatile, non-hygroscopic liquid as there is astringent. The Gillette formulations numbered 1, 4-8, and 11 in Plaintiff’s Ex. 98A