This is a patent infringement action in which American Original complained that the defendants had infringed its patent for hydraulic evisceration of mollusks. Defendants admitted infringement but asserted several grounds on which they claimed the patent was invalid. The district court found that the patent was valid, and the defendants appeal. We affirm.

* To begin with, it is helpful to understand the relationship of the litigants. The plaintiff, American Original Corporation, is primarily engaged in harvesting, processing and selling clams and other seafood. Defendant H. Allen Smith, Inc. also is engaged in the processing of clams, although the extent of its operations is limited to shucking and debellying. Much of the product is sold to defendant Sea Watch Limited for further processing. H. Allen Smith, Sea Watch Limited, and Jenkins Food Corporation, the third corporate defendant and a processor and canner of vegetables, are under common ownership. Individual defendants Walter Quimby and Lawrence A. DeMarino are former officials of American Original who went to work for one or more of the defendants. The parties stipulated that Jenkins hired Quimby in order to allow Jenkins to enter the clam processing business. Quimby subsequently recommended that Jenkins hire DeMarino to supervise the clam operations. Quimby and DeMarino admittedly patterned the clam eviscerating equipment that they built for H. Allen Smith after that in use by American Original, their former employer.

The patent-in-suit involves a process to separate the bellies from the tongues of clams and other mollusks. The tongue means the balance of the clam except the belly. The belly portion not only causes an undesirable taste, but also discolors the remaining meat during processing. Traditionally, the debellying or eviscerating operation was carried out by hand, an expensive and time consuming process. The plaintiff's late president, John Marvin, who died during the case, was issued U.S. Patent No. 4,148,112 (Marvin '112) for a process known as the "Hydraulic Evisceration of Mollusks." The patented process completes the debellying process without the need for human handling. This is the patent which the plaintiff claims is infringed.

The events leading up to the development of Marvin '112 will be discussed later, but an explanation of the patent itself is in order now. A schematic flow drawing of the process appears as Figure 1 appended to this opinion. Water or some other fluid, at a high velocity, enters the apparatus at 2. This high velocity substance creates a vacuum in pipe 3. One at a time, shucked clams or other mollusks are introduced at 3 and sucked up the pipe. As the clam meets the high velocity flowing fluid at 1, the shearing force of the fluid on the clam eviscerates the mollusk. The separated components of the mollusk are carried by the flowing substance through pipe 5 and are eventually sorted through the use of a screening apparatus.

At trial, as here, the defendants asserted basically three defenses to the charge of infringement. First, and principally, they assert that the Marvin '112 patent is invalid because it was anticipated by prior art. 35 U.S.C. Sec. 102(b). Second, they maintain that even if the patent was not anticipated, Marvin obtained it fraudulently. Finally, they claim that the claims and specifications of Marvin '112 are fatally indefinite under 35 U.S.C. Sec. 112.

The prior art at issue concerns U.S. Patent No. 3,688,344 for an Impact Clam Extractor, issued in 1972 to Harold C. Carlson (Carlson '344). It is alleged that claims 1, 3, 4 and 5 of Marvin '112 read on licensed versions of the Carlson debelliers which were installed at the plaintiff's Grasonville, Maryland plant. There is another installation, but the parties agree Grasonville is typical. An aspect of this prior art is certain experiments run on these machines by the plaintiff's employees at Marvin's direction.

Carlson testified that he devised his patent after observing gulls drop clams onto a hard surface in order to break open the shells. He noted that the bellies of these dropped clams were almost separated from the rest of the clam meat. He verified his observation of the gulls by throwing shucked clams against a brick wall with the same result. Utilizing an emergency bilge pump as a propellant system, Carlson devised an apparatus that would propel shucked clams against a metal screen. Clam bodies were left on the screen, but the bellies passed through it. Carlson eventually was issued a patent for a commercial version of the device. A schematic of the Carlson patent appears as fig. 2, also appended. A stream of fluid is introduced at point 30. The clams are fed, one by one, into the funnel at point 28. They meet the fluid at elbow joint 24 and are then carried by the flow until they impact on a rotating metal disk 54. The disk provides the impacting surface which Carlson had earlier observed.

American Original, under license of the Carlson patent, built several of the machines for its Grasonville plant. The machines were not exactly like the schematic shown in fig. 2, having a tee joint instead of an elbow at the point clams were introduced into the water flow. After installation of the first machine at Grasonville, Marvin, owner and chief officer of American Original, discovered that the impacting process tended to mutilate and otherwise damage a significant portion of the clam tongues. Reducing the water pressure lessened the damage but it also limited the effectiveness of the debellying. Marvin then tried changing the tee joint to a Y joint, suspecting that the clam tongues were being damaged when they hit the water stream. This failed to alleviate the damage problem, and it had the additional undesirable effect of reducing the effectiveness of the debellying.

In January 1976, Marvin began to suspect that not all of the debellying was taking place at the point where the clams impacted with the metal plate. In August 1976, Marvin requested Grasonville plant manager Richard Adams to experiment with completely removing the impact apparatus. Instead, the clams would be fed directly from the water pipe into a cushioning water tank. Adams observed that these changes not only eliminated the damage problem, but that 90% of the bellies were still being separated from the tongues without the damage to the meat when the impact surface was used. (Still later tests reveal that almost 100% of the clams would be debellied with increased pressure.) Subsequent experiments confirmed these results, and application was eventually made for the patent-in-suit. Believing that Marvin's changes in the system constituted a new invention, American Original terminated the licensing agreement with Carlson in October 1976 and began commercial operation with the modified system at about the same time.

Marvin's patent application would not be valid if the claims were anticipated by prior art. 35 U.S.C. Sec. 102(b). The defendants' basis for claiming that Marvin '112 is anticipated is the allegation that claims 1, 3, 4 and 51 of the Marvin patent application read on certain aspects of the Grasonville installation of Carlson '334.

Defendants contend that Marvin's use of the word comprising in the patent's claims makes the claims invalid in light of the prior art. They maintain that comprising is a term of expansive meaning which would allow the claims to include silently a step where the clams impacted against a hard surface. Defendants contrast the meaning of comprising with the word consisting, which has acquired a limiting meaning in the construction of patents. Parmelee Pharmaceutical Co. v. Zink, 285 F.2d 465, 469 (8th Cir.1961); 2 Chisum, Patents Sec. 8.06 (1982). Defendants assert that Marvin should have used the term consisting if he meant to limit his claims to a process without an impacting step.

Defendants further argue that in ascertaining the true scope of the patent's claims we should not take into account the content of the patent's specifications. It is axiomatic in patent law that the scope of a patent is to be measured by its claims. 35 U.S.C. Sec. 112; Smith v. Snow, 294 U.S. 1, 11, 55 S. Ct. 279, 283, 79 L. Ed. 721 (1934); Nichols v. Minnesota Mining and Manufacturing Co., 109 F.2d 162, 165 (4th Cir.1940); 4 Deller's Walker on Patents Sec. 241 (1965). Nevertheless, it is recognized that the specifications which accompany the claims are of use in ascertaining the extent of the invention. The Supreme Court has observed:

While the claims of a patent limit the invention, and specifications cannot be used to expand the patent monopoly ... [italics omitted] it is fundamental that claims are to be construed in the light of the specifications and both are to be read with a view to ascertaining the invention...

United States v. Adams, 383 U.S. 39, 48-49, 86 S. Ct. 708, 712, 15 L. Ed. 2d 572 (1966); see 4 Deller's, supra, at 111. Similarly, the Ninth Circuit has observed, "It is also to be borne in mind that the specification may not be used to enlarge a claim, but can be used to limit a claim." Illinois Tool Works, Inc. v. Brunsing, 389 F.2d 38, 40 (9th Cir.1968).2

In the instant case, the Marvin '112 patent's specifications refer to the Carlson patent, stating:

One example of such an attempt at mechanical debellying is shown in Carlson, U.S.Pat. No. 3,688,344. In Carlson, whole clams are propelled by fluid stream against an impact surface whereby the impact allegedly results in separation of the belly from the remainder of the clam meat. The violent impact of the Carlson apparatus, however, is not particularly efficient and results in unnecessary structural damage to the remaining clam meat.

Defendants argue, however, that even though this specification limits rather than expands the scope of the patent, we must ignore it in interpreting the claims. Not only are we told to ignore this statement, but we are asked to read the Marvin claims to permit the opposite meaning. That is, we are asked to conclude that the Marvin claims can be read to include an impact step.

We are not unaware that patent draftsmen will often prepare broad claims so as to cover as much subject matter as possible. In Nichols v. Minnesota Mining and Manufacturing Co., 109 F.2d 162, 165 (4th Cir.1940), we quoted the D.C.Circuit, saying:

Over and over again, we have ruled that claims must be given their broadest interpretation which they reasonably will support, and that a party will not be permitted to narrow his claims to suit the exigencies of a particular situation.... The party who states his claims before the Patent Office in broad language is not in a position, when thrown into interference, to read limitations into them.

quoting In re Levy, 2 F.2d 939, 940 (D.C.Cir.1924). This case is simply not the situation contemplated in the quotation from Nichols just above. The claims in the Marvin patent do not appear to us, nor did they appear to the district court, to be ambiguous. The Marvin claims call, in plain terms, for eviscerating the clam bellies in a "fluid shearing zone" by a "high velocity stream of fluid creating forces to shear the bellies from said whole clam bodies." There is nothing in the claims that would lead one to believe that an impact step could or should be included in the patented process. Everything elsewhere in the patent, including the specifications, illustrations, descriptions, and discussions of the prior art, indicates that an impact step is to be avoided. Should we accept the defendants' argument, we would have to ignore not only the claims in the patent but the other information as well, and give too strained a meaning to the word comprising. We decline to take such steps to interpret what we find to be unambiguous claims in the Marvin patent showing hydraulic evisceration rather than evisceration by impact.

Having determined that the Marvin '112 claims do not include an impact step, we then must determine whether the installation at Grasonville can be said to anticipate silently evisceration by shearing. This court has observed about the anticipation defense generally:

Anticipation is a technical defense which must meet strict standards: "Unless all the same elements are found in exactly the same situation and united in the same way to perform the identical function" in a single prior art reference "there is no anticipation."

Tights, Inc. v. Acme-McCray Corp., 541 F.2d 1047, 1056 (4th Cir.), cert. denied, 429 U.S. 980, 97 S. Ct. 493, 50 L. Ed. 2d 589 (1976), quoting, Ceramic Tilers Supply, Inc. v. Tile Council of America, Inc., 378 F.2d 283, 284-85 (9th Cir.1967).

An opinion on the question of silent or accidental anticipation is the Supreme Court's decision in General Electric Co. v. Jewel Incandescent Lamp Co., 326 U.S. 242, 66 S. Ct. 81, 90 L. Ed. 43 (1945). There, the prior art demonstrated a process of frosting the inside of electric light bulbs in order to diffuse light. Id. at 246, 66 S.Ct. at 82. The patent-in-suit recognized that earlier advances had the additional benefit of strengthening the light bulb. Id. at 244, 66 S.Ct. at 82. The application claimed patentability for putting the previous discoveries with their known effects inside the bulb where the frosting would not collect dirt. The Court held that this did not amount to a new discovery, and said:

Where there has been use of an article or where the method of its manufacture is known, more than a new advantage of the product must be discovered in order to claim invention ... [italics omitted] It is not invention to perceive that the product which others had discovered had qualities they failed to detect.

Id. at 249, 66 S.Ct. at 84.

Similarly, this court in American Stainless Steel Co. v. Rustless Iron Corp., 71 F.2d 404 (4th Cir.1934), considered the validity of a patent for a steel alloy that was easily machinable. Some years earlier, General Electric had produced the same alloy for commercial use. The later patentee argued that the prior use of the alloy was "merely accidental" and thus not anticipatory. Id. at 409. We held the prior use invalidated the patent, but stated the law:

The rule undoubtedly is that a prior use, in order to negative novelty in a patent, must be something more than an accidental or casual one. It need not have been persisted in, and the thing produced need not have gone into commercial use; but at least the discovery must have gotten beyond the experimental stage and have become complete, and the inventor must have understood and appreciated that it was capable of producing the results sought to be accomplished.

Id. at 407.

The statements from General Electric and American Stainless, however, must be compared with cases such as Tilghman v. Proctor, 102 U.S. 707, 26 L. Ed. 279 (1881). In Tilghman the Court considered the validity of a patent for manufacturing certain fat acids and glycerine. Id. at 709. At least two previous processes had the accidental effect of similarly producing some of the same products. Id. at 710-11. The Court, nevertheless, upheld the validity of the patent-in-suit, saying:

[The earlier processes] revealed no process for the manufacture of fat acids. If the acids were accidentally and unwittingly produced, whilst the operators were in pursuit of other and different results, without exciting attention and without its even being known what was done or how it had been done, it would be absurd to say that this was an anticipation of Tilghman's discovery.

Id. at 711-12. Accord Eibel Process Co. v. Minnesota & Ontario Paper Co., 261 U.S. 45, 43 S. Ct. 322, 67 L. Ed. 523, see esp. p. 58-60, 43 S.Ct. at p. 326 (1922); McKee v. Graton & Knight Co., 87 F.2d 262, 264 (4th Cir.1937).

In applying the foregoing principles to the instant case, we think we should keep in mind the intent and the claims of the inventors. Carlson, in developing his invention and obtaining his patent, was clearly attempting to eviscerate clams by their impact against a metal surface. This is apparent by reading the patent's descriptions and claims. It is supported by the explanation of how Carlson developed his invention, and by his testimony that he believed impact was necessary to obtain the desired result. The title of the Carlson patent, "Impact Clam Extractor," is a good synonym for that invention. By contrast, Marvin's intention was to eviscerate clams by exposing them to a shearing hydraulic force. Such a force was only incidentally part of the Carlson invention and the Grasonville debelliers.

The defendants argue that incidental shearing is analogous to the alloy in American Stainless and the frosting of the light bulb in General Electric. Nevertheless, we think the rationale of Tilghman, Eibel and McKee applies here. Like the fat acids of the previous processes in Tilghman, the shearing evisceration here was "accidentally and unwittingly produced." And, like the hole in the picker for the shuttle loom in McKee, the "prior accidental production of the same thing, where character and function are not recognized [the incidental shearing] does not anticipate." "[T]he inventor must have understood and appreciated that [his discovery] ... was capable of producing the results sought to be accomplished." American Stainless at p. 407. Here, Carlson neither understood nor appreciated that substantially all the shearing was by hydraulic forces rather than by impact. Carlson sought to use the high velocity fluid as a vehicle to hurl the clams against the metal wheel to eviscerate by impact. Marvin, on the other hand, taught evisceration by hydraulic shearing, an entirely different theory.

As much or even greater support for the proposition that Marvin's was a patentable discovery is supplied by the Supreme Court's decision in Lawther v. Hamilton, 124 U.S. 1, 8 S. Ct. 342, 31 L. Ed. 325 (1888), relied on by the district court, the Seventh Circuit's opinion in Hardinge Bros. v. Marr Oil Heat Mach. Corp., 27 F.2d 779 (7th Cir.1928), and the Second Circuit's opinion in Aerovox Corp. v. Concourse Electric Co., 65 F.2d 386 (2d Cir.1933).

In Lawther the Court considered the validity of a patent which, while based on prior art, omitted one step from the earlier process. The patent concerned extracting oil from seed such as flax and consisted of omitting the step of crushing the seed between "muller stones." The Court held the patent to be valid, and said:

If as ... [the court below] says, and we think rightly says, the omission of the muller stones is a real improvement in the process of obtaining the oil from the flaxseed, if it produces better oil and better oil cakes, and it is new, and was not used before, why is that not a patentable discovery?"

124 U.S. at p. 6, 8 S.Ct. at p. 344. In Hardinge Bros., the validity of a patent was upheld against a contention that one of the principal differences between it and a similar previous patent was that it had eliminated the cover on an oil burner which was the subject of the patent. The court decided against that argument, stating that "[t]he omission of an element of a combination, without retaining its function, is not invention. Yet, if the element does not contribute to the success of the combination, but prevents or defeats it, the omission of such an element may be invention. The latter is what King did, although not all that he did, over Becker." Aerovox concerned a condenser used in radio receiving sets. Previous devices had been made up of the same parts with the space between the cartridge and the walls of the box container filled with pitch or wax. The invention principally was to take out the pitch or wax, which reduced current leakage with "success immediate and great." The court sustained the validity of the patent, saying that "when an art has for long gone on the assumption that it is necessary to retain an element ineffective and uncertain in operation, its omission ought not to be retained within the compass of the routineer."

In the case before us, there is no doubt that removal of the impact step was a vast improvement on the Carlson installation at Grasonville. Damage to the clam tongues was significantly reduced but evisceration was maintained. Plaintiff claims, without contradiction, that the Marvin device saved it $500,000 per year. We think that the Marvin discovery is a real improvement in the process of eviscerating clams, Lawther at p. 6, 8 S.Ct. at p. 344; that it removes an element which prevents much of the success of the previous Carlson patent, Hardinge Bros., at p. 781; and that the omission of the impact wheel should not be rated within the compass of the routine, Aerovox at p. 388. These three cases are very nearly on point with the fact situation at hand, and they are persuasive to us.

Defendants' final anticipation argument is that the Adams experiments in August 1976 amounted to a prior use of the technology. At this point, we mention the language of 35 U.S.C. Sec. 102(b) which provides that no patent is available if:

the invention was patented or described in a printed publication in this or a foreign country or in a public use or on sale in this country, more than one year prior to the date of the application for patent in the United States...

We do not think the Marvin experiments amount to a public use under the statute. The parties stipulated that Adams reconnected the impacting apparatus after his experiments and before resuming commercial operation of the Grasonville debelliers. Commercial operations with the modified equipment did not begin until October 17, 1976, less than a year before Marvin filed his patent application on September 28, 1977. Adams' experiments at Marvin's direction were, as the district court found, part of Marvin's "act of invention." They were not prior art.

The defendants also assert that the Marvin patent application failed to disclose fully prior art and thus the patent was obtained fraudulently. Precision Instrument Manufacturing Co. v. Automotive Maintenance Machinery Co., 324 U.S. 806, 818, 65 S. Ct. 993, 999, 89 L. Ed. 1381 (1945). An infringer may avoid liability on the ground that a patent was obtained fraudulently. Walker Process Equipment, Inc. v. Food Machinery & Chemical Corp., 382 U.S. 172, 176, 86 S. Ct. 347, 349, 15 L. Ed. 2d 247 (1965). Nevertheless, the defendant asserting a fraud defense in a patent action has a heavy burden in establishing it. The evidence must be "clear, convincing, and satisfactory." United States v. American Bell Telephone Co., 167 U.S. 244, 262, 17 S. Ct. 808, 819, 42 L. Ed. 144 (1897).

There is no claim that the Marvin application failed to disclose the Carlson patent or its perceived method for debellying clams. Rather, the defendants assert that the debelliers which were installed at the Grasonville plant were separate instances of prior art from the Carlson patent and thus had to be disclosed separately. Furthermore, defendants contend that the Marvin tests were also prior art that had to be disclosed. Finally, the defendants contend that Marvin knew that the Carlson patent actually worked by hydraulic shearing instead of impacting, and it thus was fraudulent for Marvin's application to characterize the Carlson patent as being based on impacting.

The district court found no basis for any of the fraud claims, and we agree. The court determined that the Grasonville debelliers were cumulative prior art and not a separate invention which needed disclosure apart from the Carlson patent which was disclosed. Furthermore, as discussed above, the Marvin experiments were not prior art because they were part of the inventing process. Finally, we do not think that Marvin's discussion of the Carlson patent can be fraudulent when the discussion is based on the explicit claims and specifications of the Carlson patent. The finding of the district court that there was no fraud is supported by the record and it is not clearly erroneous. FRCP 52(a).

Finally, the defendants assert that the Marvin claims are "fatally indefinite" under 25 U.S.C. Sec. 112. The district court found that the claims complied with Sec. 112 which requires the specification to be in such terms as to enable any person skilled in the art to which it pertains to make and use the device. It based its opinion on the expert testimony of a plaintiff's witness and the testimony of the fabricator of the infringing device or devices of the defendant. We think its holding is supported by the record and is not clearly erroneous. FRCP 52(a).

NOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLENOTE: OPINION CONTAINS TABLE OR OTHER DATA THAT IS NOT VIEWABLE