This appeal concerns the validity of United States patent No. 3,067,816, which relates to a process for the production of ceramic suction box covers used in papermaking machines. Plaintiff-appellant, The Carborundum Company (Carborundum) sued defendant-ap-pellee, Wilbanks, Inc. (Wilbanks) for patent infringement, injunctive relief and damages. Wilbanks denied infringement on the grounds that the patent in suit was invalid under 35 U.S.C. §§ 101 , 102, 103 and 112, and counterclaimed for a declaration of invalidity under 28 U.S.C. § 2201 and injunctive relief against further suits or threats of suit against customers of defendant.

The district court found the patent invalid under 35 U.S.C. § 103 as obvious to a person having ordinary skill in the art. In addition, it found the terms of the specification required by 35 U.S.C. §§ 111 , 112 were stated too indefinitely to enable a person skilled in the art to make and use the invention. In this appeal, Carborundum assigns error to those conclusions of law and the factual findings upon which they are based.

The district court had jurisdiction under 28 U.S.C. § 1338 (a). Our jurisdiction rests upon 28 U.S.C. § 1291 . We affirm.

(a) Factual Background of the Case

As is common in patent cases, a relatively complex mechanical process must be explained before we move to a discussion of the legal questions raised by the appellant. The numbered diagram on page 45 has been provided to serve as an aid in describing the nature and function of the alleged invention in the context of the entire papermaking process.

1. The Fourdrinier Papermaking Machine

The Fourdrinier machine, as it is called by those in the papermaking industry, consists of five major parts: (1) a headbox (Fig. 1, No. 11); (2) an endless wire screen (Fig. 1, No. 12); (3) two rollers, one of which is mechanically driven (Fig. 1, No. 13); (4) draining apparatus consisting of table rolls (Fig. 1, No. 14); suction boxes (Fig. 1, No. 15); and (5) drying equipment consisting of numerous rollers (Fig. 1, No. 16). When in operation, the wire screen rotates continuously around the rollers and thus transports the raw material over the various components that transform it into finished paper.

The raw material used in the paper-making process is chemically-treated wood pulp. In the initial phase of the process, a quantity of wood pulp, called a slurry, is mixed with water and fed from the headbox onto the rapidly moving endless wire screen. The screen provides support for the wood fibers and, through a sideward shaking motion, helps form them into a thin matted sheet as the water is drained from the slurry.

Next, the draining process is accomplished by the table rolls (Fig. 1, No. 14) and the suction boxes, which are located directly below and which support the rapidly moving wire. (Fig. 1, No. 15) The suction boxes are encased in metal and divided into numerous parallelogram-shaped segments. (Fig. 2, No. 24) The cover of each segmented suction box is perforated (Fig. 2, 3, 4, No. 23) to permit the downward passage of air, which is induced by tubing connected to an evacuation pump.

Standing by itself, the suction box is a relatively simple two-part device. (See generally Fig. 2, 3, 4) It consists of a cover (Fig. 3, No. 20), that is attached by a bolt and flange (Fig. 3, No. 36) to a base member. (Fig. 3, No. 21) As is evident from a glance at the overall schematic diagram (Fig. 1), the suction box covers are in constant contact with the rapidly moving wire screen. The result of the friction caused by this constant rubbing is a high rate of wire wear.

2. The Problem of Wire Wear and the Industry Task Force Studies

Wire replacement is probably the most costly aspect of using the Fourdrinier papermaking process. The endless belt of woven mesh wire is usually made of bronze and has a length of 75 to 137 feet and a width of 155 to 264 inches. Wire life ranges between three and six days and replacement cost varies between $3000 and $5000. These figures do not include lost profits caused by decreased production during the installation of new wire screens.

At the turn of the century, the high cost of wire replacement was accepted as one of the realities of using the Four-drinier process, as almost all paper producers did. At that time, machines were operated at a rate of speed in the neighborhood of 500 wire-feet per minute. However, as the result of an almost geometric increase in demand for raw paper, it was necessary to increase the speed of the entire production process. Thus, by the early 1950’s, machine speeds of approximately 2000 wire-feet per minute were not uncommon. One result of the fourfold increase in the speed of the machines was a disproportionately higher rate of wire wear through abrasion caused by the constant contact between the wire and the stationary parts of the machine.

In an effort to discover a method of reducing their rapidly rising wire replacement costs, members of the industry, acting through the Canadian Pulp and Paper Association, formed a task force to study the entire papermaking process. The studies confirmed speculation that the major sources of wire wear were dragload, the presence of grit in the pulp slurry and the increased speed of the mesh wire. All of these factors were found to accentuate the abrasive contact between the bronze mesh wire and the suction box covers, which were found to account for the largest exposed surface area over which the wire screen passed. The task force concluded that the most feasible approach to the solution of the wire wear problem would be to find a substitute material for suction box covers, which previously had been made of end-of-grain maple.

A second task force made extensive tests on specific materials such as tef-lon, nylon and rubber. The conclusions drawn from the results of these tests were considered in detail by Judge Kil-kenny in deciding the legal issue of patent validity. They are discussed in section b. infra.

3. The Application for the Gould Patent

In 1958, Mr. William Gould, a recent college graduate employed by Carborundum as a sales engineer, heard of the suction box-wire wear problem and suggested the use of a hard ceramic substance called “KT” Silicon Carbide as an alternative to end-of-grain maple. Favorable results were obtained from tests run by the Ontario Paper Company. An application for patent was filed in January 1960 and the patent was granted in December 1962.

The file wrapper (C.T. 5, 6) of Patent No. 3,067,816 — the “Gould Patent” as it was called in the trial court- — consists of two pages, the first of which is introductory in nature reciting in greater detail substantially the same history of the papermaking industry and description of the Fourdrinier process as outlined in section (a) (l)-(3), supra. The inventor then generally describes his invention and numerous materials that might be suitable to construct it. As to the physical specifications, he states: “I prefer to use an impermeable ceramic having a density of approximately 80% or more of its theoretical density and a hardness of seven or more on Moh’s scale.” (C.T. 5)

He also states: “[t]he dynamic bearing surfaces should be smooth and are, therefore diamond ground, if necessary.” However, there is no technical specification of required smoothness. The introductory passage ends with a short recitation of the favorable results obtained using silicon carbide suction box covers.

The inventor then makes eight specific claims defining the scope of his patent.* The common theme running through the first five claims is a suction box cover that is described as having the following physical properties: “a screen contracting surface portion that is dense, hard and wear-resistant that will not retain grit or other foreign particles * * * being composed of an impermeable ceramic material having a hardness of 7 on Moh’s scale.” (C.T. 6) Claim number five adds the following information: “ * * * an impermeable ceramic material of the character of self-bonded silicon carbide, and having density of at least 80% of its theoretical density. * * *” (C.T. 6)

4. The Alleged Infringement by Wilbanks

Since August 1964 defendant-appellee Wilbanks has engaged in the manufacture of highly polished solid ceramic segments of aluminum oxide. The hardness of the aluminum oxide used by defendant is approximately 9 on the Moh scale and has a surface finish “of 4-7 rms in Microinehes.” (C.T. 758) (See Note 5, post)

(b) The Trial Court’s Findings of Fact and Conclusions of Law

As the district court pointed out in its opinion there was little doubt that Wil-banks was an infringer if the Gould patent was valid. (C.T. 766) Thus the central issues before the trial court were (1) whether the subject matter of the patent was obvious under 35 U.S.C. § 103 , and (2) whether the patent specification was sufficiently explicit to meet the requirements of 35 U.S.C. §§ 111 , 112. Those statutory provisions read in relevant part as follows:

“§ 103 Conditions for patentability; nonobvious subject matter.

A patent may not be obtained * * * if * * * the subject matter sought to be patented and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art to which said subject mat-. ter pertains. * * *”

“§ 112 Specification

The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention. * * *»

1. The Prior Art

In determining whether the subject matter was obvious and thus unpatentable, it was necessary for the district court to evaluate the state of the prior art in the papermaking industry. Specifically, it was necessary for him to evaluate the quality and quantity of technical knowledge in the industry concerning the wire wear-suction box problem. In reaching his conclusion he placed substantial reliance on the previously mentioned task force studies.

Contrary to appellant’s claim that the results of these experiments unequivocally pointed to the use of “soft” materials as an alternative to end-of-grain maple, he found that there also was evi dence supporting the use of hard materials. He stated:

The known properties and uses of “KT” Silicon Carbide were also considered by the trial court in determining the obviousness of the invention. It was significant, Judge Kilkenny noted, that the abrasion-resistant physical characteristics of “KT” Silicon Carbide were well known in the industry as the result of the previously mentioned article published in 1956 by the inventor, Kenneth Taylor, who was also an employee of Carborundum.

Moreover, there was evidence that “KT” Silicon Carbide had been used in papermaking machines for purposes not substantially different from those mentioned in the Gould patent.

“It also appears that ‘KT’ Silicon Carbide was used in bearing members prior to the invention and in other wear resistant parts of papermaking machines.” (C.T. 763)

Finally there was evidence that silicon carbide deposits from the pulp slurry effected a reduction in wire wear after they had become imbedded in the soft end-of-grain maple suction box covers.

“ * * * wood suction covers caused less wear after grit composed to a great extent of silicon carbide from the pulp stones became imbedded in the surfaces thereby producing a smoother harder surface.” (Emphasis added) (S.T. 764)

On the basis of the preceding evidence, the court below found it “natural and obvious for Gould to suggest the use of ‘KT’ Silicon Carbide.” (C.T. 764) This compelled the legal conclusion:

“ * * * that the * * * Patent involved an obvious substitution of material and, consequently, [is] invalid.” (C.T. 765) (Emphasis added) ******

“In Hotchkiss v. Greenwood, 52 U.S. 248 [ 13 L.Ed. 683 ] (1850), the Court held that the mere substitution of materials was not patentable.” (C.T. 761)

2. The Specification Issue

In turning to the second defense of Wilbanks — that the statutory requirement of specification had not been met —the trial court found the descriptive words “impermeable ceramic” and the technical specifications of hardness, density and surface finish to be too indefinite “to enable any person skilled in the art * * * to make and use the [invention].” ( 35 U.S.C. § 112 , in greater detail, supra.)

First, it was noted that at least one expert defined “ceramic” in such terms as to include glass. Since glass with a density of 80% and a hardness of 7 Mohs existed at the time of the application for the Gould patent, it was concluded that the Gould patent encompassed U.S. Patent No. 128,469, relating to glass suction box covers.

Second, trial court thought “a highly critical eye” (C.T. 760) should be cast on the hardness and density specifications since “[t]he testimony is practically undisputed that a density of 80% is insufficient to give the mirror-like surface required for a successful cover.” (C.T. 760) Indeed, evidence had been introduced that flame-sprayed aluminum oxide with a'density of 88% could not be polished to the required smoothness. (Id).

Third, after observing that a mirror-like surface was “absolutely critical to success” (C.T. 761), the trial court noted that the Gould application contained no technical specification whatever relating to the required surface finish. There seemed to be little excuse for the absence of such a specification since an established standard of measurement exists for measuring surface smoothness.

It was concluded from the foregoing analysis that the physical specifications could not be considered critical and thus

“* * * it could well be said that the plaintiff has attempted to block off a whole area of scientific development by drawing over extended boundaries around his alleged invention. Brenner v. Manson, 383 U.S. 519 [ 86 S.Ct. 1033 , 16 L.Ed.2d 69 ] (1966).” (C.T. 761)

Two legal issues are raised in this appeal. First, were the factual findings of the trial court clearly erroneous? Second, were the legal standards governing obviousness and specification correctly applied? We think the record supports a negative answer to the first question and an affirmative answer to the second.

(d) The Findings of the Trial Court Were Not Clearly Erroneous

We reject appellant’s contention that the trial court’s findings of fact were clearly erroneous under Rule 52(a), Fed.R.Civ.P. As is evident from our previous discussion (section b., supra), each finding of the trial judge finds support in expert testimony or documentary evidence. Conflicts in the evidence were resolved against the appellant on the basis of the trial judge’s first hand evaluation of the various witnesses’ credibility. This function is reserved for the trier of fact and, we think, must be respected, particularly in a case such as this involving as it does a mass of highly technical factual information. Jacuzzi Bros. v. Berkeley Pump Co., 191 F.2d 632 (9th Cir. 1951).

(e) The Trial Court Correctly Applied the Legal Standards Governing

Obviousness

Unlike novelty and utility, which were prerequisites to patentability mentioned specifically in the Patent Act of 1793, the requirement of nonobviousness was of judicial origin, first appearing in the case of Hotchkiss v. Greenwood, 52 U.S. 248 , 13 L.Ed. 683 (1850). In 1952, Congress codified the Hotchkiss doctrine in section 103 of the patent revision act of that year. The recent case of Graham v. John Deere, 383 U.S. 1 , 86 S.Ct. 684 , 15 L.Ed.2d 545 (1966), represents the first extended discussion by the Supreme Court of the statutory requirement of nonobviousness.

In John Deere the Court discussed the historical underpinnings of Article I, § 8, cl. 8 of the Constitution and pointed to the early recognition by Thomas Jefferson of the need for a precise standard by which “ ‘the things which are worth to the public the embarrassment of an exclusive patent’ ” (383 U.S. at 10-11, 86 S.Ct. at 690) could be set apart from those other less useful devices that could not be thought to “ * * * promote * * * the Progress of * * * useful Arts” (Const. Art. I, § 8) sufficiently to merit the limited monopoly permitted by the Constitution. Mr. Justice Clark then moved to a discussion of the statutory provision enacted in 1952. He characterized the test of patent validity as one of law, based upon several factual inquiries. Specifically he stated:

The majority in John Deere also noted that certain secondary considerations such as “commercial success * * * [and] long felt but unsolved needs * * * might be utilized to give light to the circumstances surrounding the origin of the subject matter sought to be patented.” ( 383 U.S. 17 -18, 86 S.Ct. at 694)

We applied the standard enunciated in the John Deere case in Jeddeloh Bros. Sweed Mills, Inc. v. Coe Mfg. Co., 375 F.2d 85 (9th Cir. 1967) and Hensley Equip. Co. v. Esco Corp., 375 F.2d 432 (9th Cir. 1967). In deciding Jeddeloh, we noted:

We hold that the trial court’s in-depth analysis of the scope and content of the prior art (see (b) (1), supra) supported his conclusion that the device in issue was a mere “unpatentable substitution of material.” (C.T. 765) Moreover, we think the evidence that the device was commercially successful and met a long felt but unsolved need was correctly categorized as “sec ondary” under John Deere, supra (C.T. 764).

In view of our holding that the patent falls by reason of its obviousness, we need not reach the problem of the standards governing specifications. (Schriber-Schroth Co. v. Cleveland Trust Co., 305 U.S. 47, 57 , 59 S.Ct. 8 , 83 L.Ed. 34 (1938).)

The decision of the district court is affirmed.

. A few years before the completion of either of the task force studies, Kenneth Taylor, an employee of Carborundum, developed and patented the process for producing “KT” Silicon Carbide. In an article published in 1956, the inventor noted that “KT” Silicon Carbide had the property of “outstanding abrasion resistance.” (C.T. 763) . Only the first five claims are in issue in this case because claims 6, 7 and 8 all relate to suction box covers made with silicon carbide. Defendant Wil-banks’ covers were made with aluminum oxide. . Supra at pp. 6-7. . Supra at p. 10. . The standard of measurement is expressed in microinches (rms), . The trial court also noted that the application failed to specify which Molí scale — the standard or revised — was to be used in interpreting the Gould patent. However, the court was of the opinion that this fact was “[o]f less significance.” (C.T. 760) . We reject appellant’s contention that the trial judge mechanically adopted defendant’s proposed findings of fact and that, therefore, we should accord them less weight than the facts mentioned in the opinion. Compare C.T. 625-48 with C.T. 768-80. See also Ohlinger v. United States, 219 F.2d 310 (9th Cir. 1955). The trial judge accepted some proposed findings presented by each side, rejected some, and modified others. . It should be noted, however, that both sides introduced numerous depositions at the trial, which were taken from experts and other persons familiar with events surrounding the patent in issue. . In affirming a finding of patent invalidity, this court said in Jacuzzi: “ * * * it is contended that since the Patent Office and the Trial Court disagreed, we should find the facts de novo. The assumption of such authority by the appellate court would be an usurpation. However, we examine the facts to determine whether the findings of the Trial Judge are clearly erroneous under Rule 52 * * * and must be set aside. “If there is not firm adherence to such a rule, everything is cast adrift. The trial courts find the facts. If appellate courts exercise no self-restraint, then, after the primary facts are thus found, these same facts are found anew twice over, with varying results. Not only is there no finality, but the findings may change with the shifting personnel or on subsequent hearings. Not only finality, but stability is lost. All is confusion.” (Footnotes omitted) ( 191 F.2d at 634 ) Cf. Leishman v. General Motors Corp., 191 F.2d 522 (9th Cir. 1951) (Trial court findings of invalidity of reissue patent, affirmed, citing Rule 52(a), Fed. R.Civ.P.) . In Griffith Rubber Mills v. Hoffar, 313 F.2d 1 (9th Cir. 1963) we observed the following: