STEVENS, Circuit Judge.
Hansen’s application for a patent on a process for assembling pieces of turkey meat into a boneless roast was originally
rejected as obvious in view of the prior work of Maas.
2Armour persuaded the Patent Examiner that Maas’ disclosure related to red meat which had not passed through rigor mortis, and that Hansen’s discovery that essentially the same process would work with post-rigor poultry meat was patentable. The primary issue here is whether the Examiner’s second or his first, appraisal of the invention was correct.
We think an artisan who was familiar with the Maas process, without really knowing why it worked, would nevertheless have tried to use it on poultry meat if he was seeking a comparable result; we also think it would have been obvious to an analyst who understood why the process worked that it would also be successful with poultry meat. The question is whether incorrect or misunderstood hypotheses in the art at the time of Hansen’s discovery made it nonobvious. We do not have the benefit of the Examiner’s judgment on this precise issue because Armour apparently persuaded him that an incorrect hypothesis was valid. The district court found that Armour then knew, or should have known, better. An understanding of how the issue arose requires a fairly complete statement of the facts.
I.
In May and June of 1964, Armour introduced a new boneless turkey roast into selected test markets. The product had a natural texture, did not lose its flavor or moisture during cooking, and could be sliced easily without crumbling or coming apart. It was a marked improvement over earlier attempts to combine chunks of meat with string or with adhesives such as gelatin or wheat gluten. It received prompt and widespread consumer acceptance.
Equally promptly it was copied by Swift. In August of 1964 a Swift research team thawed out an Armour roast, tore it apart, and identified the process by which it had been put together. The basic process can be simply described.
When chunks of meat are agitated in a salt solution a tacky exudate forms on the surface of the chunks; when they are pressed together and cooked, adjacent pieces are cemented to one another almost as though they originally comprised one large piece. Results are affected by the temperature during agitation, the extent of agitation, the amount of salt in the solution, and the character of the meat; however, the key to the process is the use of protein within the meat itself as the agent which cements the pieces together. Swift had not found this key until it analyzed Armour’s product in 1964.
Armour applied for the patent in suit
on March 16, 1964; the patent issued on November 15, 1966. This infringement action followed. After a long trial the district court, 320 F.Supp. 433, held the patent invalid; the court found that Hansen’s invention had been anticipated by the work of Maas at Oscar Mayer
***, and by the work of Dr. Baker at Cornell University
, that the process would have been obvious to one of ordinary skill in the art at the time the alleged invention was made
, and
that Armour had breached its obligation of candor to the Patent Office.
Armour appeals, urging that the findings are clearly erroneous. It argues that Maas’ work on poultry to the extent that it preceded Hansen’s discovery had been either “concealed” or “abandoned” within the meaning of § 102(g)
, that Dr. Baker’s work related to an emulsion composed, in part, of ground poultry meat, unlike the chunks bound together by Maas and Hansen; that Swift’s failures demonstrated that Armour’s success was anything but obvious; and that its disclosures to the Patent Office were accurate if appraised in a correct temporal context and its nondisclosures were inadvertent and insignificant.
We agree with Armour that it is imperative that prior art, as well as Patent Office proceedings, be analyzed without reliance on after-acquired knowledge. As so analyzed, however, we are satisfied that the district court correctly concluded that this patent is invalid. The defects in Armour’s prosecution of the patent application overcome the normal presumption of validity. Moreover, Hansen’s probable knowledge • — which we treat as actual knowledge in view of every inventor’s presumed familiarity with the totality of relevant prior art
— of work which Maas had done at Oscar Mayer before Hansen left that company and joined Armour in 1962 and which was the subject of a patent application by Maas adequately explains the contrast between Armour’s success and Swift’s failure.
We first analyze the state of the prior art as of September 24, 1962, when Hansen’s conception was disclosed to Armour’s Patent Department.
We then
contrast Armour’s representations to the Patent Office on May 26, 1966, with what Armour actually knew, or as the district court found, should have known, on that date. It was Armour’s filing on that date which persuaded the Examiner to reverse his original rejection of the Hansen application as obvious in view of the work of Maas.
II.
Several hours after an animal dies, its muscles pass through rigor mortis. The proteins actin and myosin, which are normally separate in living and pre-rig- or tissue, combine to form actomyosin in post-rigor tissue. The binding agent in the tacky exudate which agitation brings to the surface of the meat chunks in the Hansen or Maas process is either myosin, actomyosin, a combination of both, or even possibly something else. Swift contends that it is actomyosin; Armour claims that it is myosin, or at least at the time of Hansen’s invention, the “art” taught that it was.
Under Armour’s theory of the case, successful binding was only possible if adequate amounts of myosin could be extracted by the agitation process. Under this theory, the fact that the amount of available myosin would decline drastically when meat passed through rigor mortis made it necessary to use pre-rigor meat in the Maas process. According to Armour, the teaching of Maas applied only to pre-rigor red meat. The Hansen discovery was that poultry — unlike red meat — could be successfully processed
after
rigor mortis had taken place because enough myosin survived rigor to achieve an adequate bind. Alternatively, the theory of the discovery was that perhaps, contrary to the teaching of the art, it was not myosin after all that effected the binding but something else.
On either hypothesis, Hansen made the surprising discovery that the Maas process would work on post-rigor poultry.
Armour’s alternative theories make it proper to summarize the state of the art at two levels. First, putting to one side an attempt to understand exactly
why
the process worked, we consider
what
an ordinary artisan, fully informed about what had worked in the past, might reasonably expect to work in the future. Second, we consider whether a student of protein theory would have considered the use of Hansen’s process on post-rig- or poultry meat an obvious extrapolation from Maas’ work on red meat in the light of other disclosures prior to September 24, 1962.
A.
Hansen worked in Oscar Mayer’s Research Department from 1957 until April of 1962. During that period Maas was in charge of Oscar Mayer’s pilot plant. Hansen, who worked in the laboratory, had access to the pilot plant, but visited it only occasionally. In October of 1958 he prepared a confidential research report on the use of pre-rigor meat as a water stabilizing agent in sausage.
The report discussed the effect of rigor mortis on the two main proteins of muscle, myosin and actin, the salt solubility of the proteins, the extractability of the proteins, and the increase in viscosity associated with the formation of actomyosin. About two years later he published an article on the emulsion formation in finely comminuted sausage
in which he pointed out that the salt soluble proteins, myosin and aetomyosin, appear to form a stabilizing membrane which add fat and water binding qualities to sausage emulsions.
In 1959 Maas first applied for a patent on a process for mechanically working pieces or chunks of meat to form a tacky exudate which could be used to cement abutting meat surfaces together. That application was abandoned and replaced by a continuation in-part which ultimately issued as the Maas patent on February 5, 1963. The specifications' in that patent described how the process worked on various kinds of cured meat and fresh meat. The term “meat” was not expressly defined, the specifications merely stating that the “invention is generally applicable or useful in connection with various types of meat or sources of meat,
e. g.,
pork, beef, veal, mutton, lamb, venison, etc.”
One example in the specifications described the procedure for making a product resembling a fresh pork chop.
Maas testified unequivocally that he used post-rig- or beef in the experiments described in the specifications.
The patent does not mention poultry meat. However, in December of 1960 and January of 1961 Maas did successfully use his process to make boneless turkey roasts. Apparently Oscar Mayer, a company engaged almost exclusively in the manufacture and sale of red meat products, never marketed this product.
On one occasion Hansen ordered some “mock pork chops” from the Oscar Mayer pilot plant.
What Hansen did with this product is not entirely clear. One of his duties at Oscar Mayer was to serve as a member of a taste panel; in that capacity he sampled products which had been made by the Maas process. Although the district court’s finding that “Hansen had participated in Maas’ work” may be over-stated, the evidence clearly supports the finding that he “had requested and received products made in accordance with Maas’ principles,” and it was certainly reasonable for the district court to infer that Hansen “had evaluated such products.”
At the time Hansen left Oscar Mayer in 1962 he clearly had knowledge of relevant art which did not become generally known until the Maas patent issued the following year.
Before Hansen went to work for Armour, Armour had had no more success in developing a sliceable turkey loaf than Swift. As Hansen testified on deposition, which testimony he acknowledged as accurate at trial, when he came to Armour, Colburn (one of his co-inventors) was trying to develop such a product “using the traditional approaches then current in the industry— gelatin, wheat gluten, et cetera.”
Hansen shared a laboratory bench with Colburn and in due course “asked him if he had ever considered using the natural indigenous proteins in the turkey himself.”
Hansen explained how functional proteins were extracted from the meat in sausage technology to give a “continuous eoagulum in the wiener after cooking.”
According to laboratory notes which Colburn apparently made several weeks later, on August 7, 1962, he agitated scraps of turkey meat with salt in a mixer until a “paste consistency” was achieved.
The next experiment recorded in Armour’s laboratory notes apparently was performed successfully on September 21, 1962.
In the meantime, on September 11 and 12, Hansen’s immediate superior, Rogers, had attended a conference on new poultry products at Cornell University conducted by Dr. Baker.
At that time Rogers sampled Dr. Baker’s frankfurters and bologna products made with poultry instead of red meat and received Dr. Baker’s explanation of how these products were made. Simply stated, he had studied the manufacture of such products from red meat and had applied the same procedures to poultry meat.
In his deposition testimony introduced at trial, Dr. Baker described the “old-time hot dog maker” who tested the quality of an emulsion by whether it would “stick to his' hand for a length of time; it should be tacky.” Dr. Baker did not find much difference between the tacky exudate which resulted from beating poultry meat instead of red meat.
Rogers testified that his exposure to Dr. Baker’s products contributed nothing to the work of Hansen and Colburn which he was supervising. There is no evidence that he mentioned the Cornell work to anyone at Armour; indeed, he apparently overlooked it completely when assisting with the prosecution of the patent application because it was not cited to the Patent Office as relevant prior art (although several red meat sausage references were cited).
Thus, on September 24, 1962, the state of the art encompassed: (1) knowledge that ground red meat used in sausage-making contains salt soluble proteins which could be extracted to achieve a binding effect; (2) knowledge that the “tacky exudate” with which the old-time sausage makers were familiar would also result when similar methods were applied to poultry meat; and (3) the findings by Maas that this tacky exudate could also be used to combine chunks of red meat together to form a large sliceable roast.
On the basis of that knowledge of the art, Maas and Hansen both took the next step and successfully applied the Maas process to chunks of poultry meat. The procedure which Dr. Baker had followed in developing his chicken frankfurter and bologna products indicates that if he had been trying to develop a boneless turkey roast, and if he had known of Maas’s work, he would have extrapolated from red meat to poultry meat. The attempts to use other adhesives such as gelatin had been employed first with red meat products and later with poultry.
We are therefore persuaded that if the level of skill in the art in September of 1962 included nothing more than knowledge of what processes had achieved successful binding of red meat and poultry meat, without any real understanding of why the processes were successful, the Hansen discovery was nevertheless obvious at the time the invention was made. We attribute his success at Armour, as contrasted with Swift’s failures, to his greater familiarity with the state of the art rather than to the originality of his conception.
B.
The foregoing discussion is an oversimplification of the obviousness issue because the level of skill in the art included a great deal of information about characteristics of proteins such as myosin and actomyosin. Armour persuaded the Examiner that the art plainly taught that the phenomenon which made the Maas process succeed could only be present in pre-rigor red meat. In view of the well recognized differences (1) between red meat and poultry and (2) between pre-rigor and post-rigor meat, Hansen testified that the success which he and Colburn achieved with post-rigor poultry came “much to our surprise.”
Our review of the various technical references in the record persuades us that
Hansen’s surprise, even if genuine, was not justified. We attribute his surprise to the fact that he may not have been aware of aspects of the art which the law presumes he should have known. We consider the red-meat/poultry and pre-rigor/post-rigor distinctions separately.
Unquestionably, there are numerous differences between red meat and poultry. Those differences, however, are not pertinent to the analysis of the relevant characteristics of the proteins that act as a binding agent in sausage, or in products made by Maas’ or Hansen’s process. On the other hand, the analysis of the behavior of actin, myosin, and actomyosin, and the salt solubility, extract-ability, and adhesive character of these proteins in several technical references in the record, is not at all dependent on whether the experiments being described used red meat or poultry.
Indeed, Hansen himself, when pressed to explain unverified comments about red meat in the patent specifications, relied on general “protein chemistry”
or “muscle literature,”
rather than just red meat art. Finally, we note that if Hansen had been informed about the work of Dr. Baker at Cornell, that practical experience would have supported the conclusion that basic principles of protein chemistry are equally applicable to the muscles of mammals and birds.
Armour’s argument that the distinction between pre-rigor and post-rigor muscle made Hansen’s discovery nonobvious is supported by Armour’s interpretation of the Maas patent, and, most strongly, by an earlier Oscar Mayer patent (“Turner and Olson”).
Armour contends that the references to “fresh meat” in the Maas patent should be construed as though they meant “pre-rigor” meat. The text of the patent itself does not support this construction. Indeed, the difference between pre-rigor and post-rigor beef is not even mentioned in Maas’ patent; if such a difference were critical to the successful use of the process, a reader of the patent would certainly expect it to be mentioned. The term “fresh meat” is more reasonably read as the housewife or the industry would understand the term,
i. e.,
meat which is neither frozen nor cured, and not so old as to be spoiled. Finally, as we have already noted, Maas himself unambiguously testified that he had worked with post-rigor beef. Thus, with the Maas patent included in the prior art, the art clearly taught that an adequate tacky exudate could be obtained on post-rigor chunks of meat.
Armour’s most persuasive argument is based on the teachings of Turner and Olson who had found that pre-rigor red meat had better binding qualities than post-rigor meat.
Their patent de
scribes a method of quick-freezing meat to arrest the rigor mortis process and enable the sausage maker to take advantage of myosin’s binding qualities before it combines with actin and forms actomyosin. Turner and Olson can certainly be read, as Armour urges, as teaching that myosin, rather than actomyosin, is the significant binding agent in meat, and therefore that the use of pre-rigor meat is essential if successful binding is to be achieved.
But the argument proves too much and ignores other art which necessarily limits the import of Turner and Olson. That patent merely teaches that
better
binding can be obtained with pre-rigor meat; it does not obsolete the know-how of the old-time hot dog maker, who obviously worked with post-rigor meat. Whether he achieved a sufficient degree of “sticky effect” from the actomyosin, from the myosin remaining after rigor mortis had set in, or even from some other inherent quality of the meat, his experience demonstrated that Turner and Olson’s use of pre-rigor meat was not
necessary
to achieve binding.
Moreover, Turner and Olson were binding substantial percentages of water and fat into sausage, a problem quite different from, though related to, the problem of sticking chunks of meat together. Perhaps, as the Turner and Olson patent implies, even the chunks could be more effectively cemented if pre-rigor meat was used in the Maas process. But since Maas itself teaches that the process works successfully on post-rigor meat, Hansen’s discovery that Turner and Olson had incorrectly implied that only pre-rigor meat contains an adequate binding agent would have been obvious to an artisan presumptively familiar with Maas.
Furthermore, even before Maas, the technical references taught that actomyosin as well as myosin might be an effective binding agent. Hansen’s own research in 1958 implied that actomyosin was more adhesive than myosin.
But
whichever was better, we are persuaded that the view that either could act as an effective binding agent is consistent with the literature as a whole — at least to the extent that it has been called to our attention and that we are able to grasp its relevant significance.
Accordingly, although we recognize that Turner and Olson does imply that the Maas process might not work on post-rigor meat, the art as a whole, particularly when Maas itself is included in the body of knowledge presumptively available on September 24, 1962, made Hansen’s invention obvious.
III.
In analyzing the obviousness issue, we have not accorded the Hansen patent the customary presumption of validity. Although the district court found other omissions significant, we think two defects in Armour’s prosecution of the patent application destroy the normal presumption. They are (1) the omission of any reference to the work on poultry sausage by Dr. Baker at Cornell;
and (2) the use by Armour of affidavits which contained inaccuracies of which Armour had, or certainly should have had, knowledge at the time the affidavits were presented to the Examiner.
Armour explains its omission of any reference to the Cornell work on the ground that Rogers’ knowledge is not imputable to Hansen or his co-inventors, or to Armour’s attorneys, and on the further ground that the chicken sausage work was not pertinent prior art. We categorically reject both explanations. Rogers employed Hansen and both of his co-inventors, supervised their work, and worked closely with the lawyer who prosecuted the application. Whether or not he actually mentioned Dr. Baker’s work to any of those individuals, his knowledge must unquestionably be treated as the knowledge of Armour, the assignee of the invention. With respect to the relevance of Dr. Baker’s work, and specifically his written descriptions of it, Armour’s citation of several red meat sausage references forecloses its argument that poultry sausage references were irrelevant.
The misleading use of the affidavits of Professors Dawson and Forsythe is less plain but no less serious. Those affidavits were filed to support Armour’s thesis that the prior art taught “that
the use of pre-rigor meat is essential,”
and to persuade the Examiner that he was in error in believing that “it is clearly evident that the meat materials employed in the Maas reference have passed through the rigor mortis stage. ,”
In his affidavit, Professor Dawson stated that he had prepared poultry products using the Hansen process and considered the results “unobvious.” He stated that a similar phenomenon had occurred in connection with pre-rigor red meat but “to my knowledge
has not been tried
or believed practical in connection with aged, or post-rigor mortis red meat.”
(Emphasis added.) That affidavit was executed on May 16, 1966.
Although Dr. Dawson no doubt believed his statement to be true, since he had no experience with the Maas process or with the work then under way in Armour’s laboratory, Armour is charged with knowledge of the statement’s inaccuracy.
Wilcox, the head of Armour’s sausage development section, testified that in 1963 and 1964 Armour had attempted to bind together pieces of post-rigor red meat by agitating them in the presence of salt.
Tests conducted in July of 1964 were unsuccessful, but the failure was not attributed to the use of post-rigor meat,
and the suggestions for future test runs did not include any recommendation to use pre-rigor meat.
By July of 1965, Armour’s work with post-rigor meat had progressed sufficiently for Wilcox to apply for a patent on what may be described as an improvement or modification of the Maas process by fraying surfaces of meat chunks of a particular size to obtain better binding effect.
Thus, regardless of when the Wilcox process became commercially acceptable, it is perfectly clear that serious attempts to bind chunks of post-rigor red meat had been made by Armour at least two years before Professor Dawson’s affidavit was filed in the Patent Office.
Armour offered three excuses for its failure to inform the Examiner about the work of Wilcox: First, that it was subsequent to the date of Hansen’s invention and the purpose of the affidavits was to establish the state of the art at that time; second, that Wilcox had not yet demonstrated that his attempts would succeed;
and third, that the Wilcox development was actually dis
closed to the Patent Office in a patent application which was eventually allowed by the same Examiner who processed Hansen’s application.
The fact that Armour’s filings were properly directed to establishing the state of the art in 1962 and 1963 cannot excuse the making of a representation which was inaccurate when it was made. In 1966 Armour knew that it was attempting to bind chunks of post-rigor meat using inherent protein as an adhesive and already had had sufficient success in the laboratory to justify applying for a patent. That work not only demonstrated inaccuracy in Dawson’s statement that protein binding of post-rigor red meat “has not been tried,” and further that Armour is chargeable with knowledge of that inaccuracy; it also tended to undermine the basic theory of the Hansen application. That theory rested on the premise that the art so clearly demonstrated the impossibility of binding post-rigor red meat that no knowledgeable person would even try such a thing. We recognize that the Wilcox process involved the fraying of the surfaces, but that fact does not make the affidavit any less inaccurate; moreover, as early as September 24, 1962, Hansen had referred to “protein on the abraded meat surfaces” in his disclosure to the Armour Patent Department.
The fact that Armour’s attempts to use the Wilcox process did not achieve commercial success in 1964 provides no justification for a representation that no such process had even been tried. It assumes, as does Armour’s first excuse, that the affidavit had been prefaced with a warning that the representation described the affiant’s understanding of the state of the art as it had existed three or four years earlier. Even though relevant inquiry was directed at an earlier date, the affidavit purported to advise the Examiner of a contemporaneous fact.
The suggestion that the Examiner was in fact correctly informed because he is the individual who ultimately allowed the Wilcox patent to issue is not persuasive. The Wilcox notice of allowance was not signed until November of 1968, more than two years after the Hansen patent issued.
Moreover, we find nothing in the portion of the Wilcox file wrapper which is in the record to direct the Examiner’s attention to the significance of the fact that Wilcox worked with post-rigor meat; that fact does not appear to have been mentioned in the Wilcox application. In any event, we think that it is unfair to a busy Examiner, no matter how diligent and well informed he may be, to assume that he retains details of every pending file in his mind when he is reviewing a particular application. Especially since Patent Office proceedings are ex parte, the applicant has the burden of presenting the Examiner with a complete and accurate record to support the allowance of letters patent. Armour failed to discharge that burden.
In our opinion the patent which issued on the Hansen application is not entitled to a presumption of validity. Without that presumption, we are convinced that the district court did not err in deciding that the claimed invention was obvious in view of the state of the art, particularly the work of Maas, as it existed in September of 1962. Accordingly, the judgment is affirmed.