Circuit Court of Appeals, Sixth Circuit.
*831 Laurence A. Janney, of Chicago, Ill. (Emery, Booth, Janney & Varney, of Boston, Mass., and A. Trevor Jones, of Chicago, Ill., on the brief), for appellants.
Howard M. Cox, of Chicago, Ill., and A. C. Paul, of Minneapolis, Minn. (Cheever & Cox, of Chicago, Ill., on the brief), for appellee.
Before DENISON, MACK, and KNAPPEN, Circuit Judges.
DENISON, Circuit Judge.
Infringement suit, based upon three patents issued February 9, 1915, to Inwood and Lavenberg, numbered 1,128,144, 1,128,145, and 1,128,252, and covering respectively: An apparatus, called the "workholder," used in or in connection with a wirebound box blank machine; an organized box blank machine for applying binding wire to assembled box parts; and a method of making wirebound boxes. The first and second were based upon copending applications, filed in October, 1904, and the third upon a divisional application, cut out from the second. A fourth patent, now expired and not directly involved, was one issued to the same patentees, dated September 19, 1905, numbered 799,854 (reissued in 1907, No. 12,725), covering a specific form of box blank a product which might result from the use of the machine and method patents and might also be made without resort to this machine or method.
Much litigation touching these four patents was had, and which we will later discuss. It is now enough to say that what is certainly a simple and, as it appears to us, rather a clear issue has become obscured and, we venture to think, confused by the complications arising through an effort to consider all of these three or four patents at once; and the result has been that they have seemed to some extent mutually to destroy each other (as to the three now involved), and thus to leave without protection an advance step in the art which had great commercial acceptance and value. The organized machine involved and disclosed the use of the method of the third patent and, as well, the use of the subordinate and partial apparatus of the first. It seems to us the most logical method of approach, to consider first this machine patent on its merits, on the issues of validity and infringement, just as if the patentees had attempted no further and different protection; *832 and if that patent is found valid and infringed, then next to consider what effect, if any, the other patents should have upon the controversy.
As to the validity of the machine patent, the issue is clearly one of invention. Anticipation, as that word is used to differentiate from lack of invention, is not claimed. Before 1904, wirebound wooden boxes were common. They comprised top, bottom, two sides and two ends. To give endwise rigidity and strength, they were supported by such frame members as were desired. Support and strength around the box laterally were given by two or more series of cleats four in a series for top, bottom and sides. These cleats were beveled so that, set together, they made a square frame and, tacked together, they kept their form while the four sides were nailed on. Thereupon wires were run around the outside, opposite the cleats, and stapled through the sides to the cleats. This work was done manually, and all at once or at different times as needed.
As early as 1887 (Hamilton, 373,828), it was thought that there would be commercial value in the making and selling of a wirebound box blank which could be shipped in the flat form and set up by the user. Howenstine (1891 473,479) developed the use of wire and staples in a knock down box with cleats. As early as 1894 (Knutson and Uhri, 518,038), if not before, it was observed that an automatic machine for making such blanks in order to get mass production, was a desideratum.[1] Greenstreet (1897 579,574) had a similar idea, but used no cleats. Rosback provided elaborate machines for this purpose (1898 608,796 14 sheets of drawings); on the same day (608,917 16 sheets); and two weeks later (609,638 11 sheets); later in the same year (614,348 with Healy, 9 sheets); later in 1899 (step-mitering, 623,258 14 sheets); and (625,958 and 630,303 and 636,068 with 11 sheets). Uhri had tried again in 1898 (608,809 27 sheets, 71 figures). Without recounting, three other efforts at such machines were made before 1904. It is clear that all these efforts were failures, and no machines had been built on which wirebound box blanks could be made at a profit. The various patented machines were either never built or were tried and abandoned.
In June, 1904, Inwood and Lavenberg devised their workholder, their complete machine and their method. The machine and their box very soon went into general use, and in a relatively short time were adopted under license by many manufacturers in different parts of the country. There is no doubt that they created a new industry the wirebound box business. The issue of the patents was long delayed, and there was litigation as to ownership; but the record shows that in 1921 there were 70 licensed plants manufacturing an annual total of 25,000,000 of these boxes.[2] After the title was established, such infringements as had appeared had been stopped, by litigation or by adjustment, until defendant in this case began to make its machines; after some unsuccessful efforts, it built an effective machine, about 1922. Suit was promptly brought against a user, Gibbons; and, pending that suit, the present one was brought.
As so often happens with an invention which revolutionizes commercial practice, the advance step made by the patentees over the prior art was simple and, looked at in the light of subsequent experience and knowledge, its inventive character is plausibly challenged. The record shows that, at the critical time, the Rosback machines represented the highest development in the art; and with them comparison should be made. Rosback had a traveling endless belt or conveyer or pair of conveyer chains. They were fed forward over the surface of a table. Upon the conveyer, suitably spaced apart, were placed two strips of lumber, destined to form the two series of cleats for the four sides of the box. The strips were as long as the total length of the four cleats. Upon the cleats was placed the side material, which was in any form desired, but was preferably a continuous sheet of thin wood, like veneer. Above this side material, and over the cleat strips, was placed on each side a continuous wire. As the conveyer was advanced, the material passed under a wire stapling device, which placed staples straddling the wire, and drove them through the side material into the cleat. Suitable means upon the conveyer engaged the rear of this material and drove it forward along with the conveyer. When the preliminary stage was thus finished, and *833 the blank was rigid throughout its length, Rosback took it to another machine, which had properly spaced cutters and saws, and moved it laterally, so that, at the points intended for the corners of the box, the rigid cleat strip was cut through from below, resulting in the severing of the strip into four pieces and beveling these upon the lower corners, so that, when the blank was folded at the points of severance, the beveled ends of the cleats made a proper miter joint. The side material was also severed or cut partly through so that it would bend at the corners. Rosback employed a succession of cutters of diminishing size which produced step miters. The blank was then ready to be shipped. Inwood and Lavenberg discarded this plan, and went back to the manual art, to begin again the search for a satisfactory automatic result. They took the so-called performed cleats, which every carpenter had always used in making wooden boxes. They beveled the ends of these so that the final result would be a miter joint. They assembled these performed cleats usually eight, two rows of four each with beveled edges down and spaced apart a trifle, as desired, in an assembling or holding frame which they called a workholder, and on which also they placed in position the side material, which might be a continuous strip of veneer or might be four separate suitable sheets. The workholder thus laden was then placed upon the endless carrier, and, wire being supplied, it was fed under a stapling mechanism, and the bank delivered as a finished product. The assembled elements thus received treatment in a single automatic machine, while Rosback had submitted them after assembly to two successive machine processes in separate machines.
If reference were to be had only to the paper art, showing the machines above mentioned, we might, for the purpose of this opinion (although we are not prepared to do so otherwise), assume that the patentees' combination of the old material feeding, wire feeding and stapling mechanism with the old preformed cleats, would not involve invention; but in this case there is an undisputed history of the practical art which, in our judgment, makes that conclusion impossible.
About 1891, Henry Stephens began to try to make wirebound boxes. This was under a machine which Howenstine had built to make the Howenstine patented boxes. This machine was not successful. Later, he employed Uhri and some time was spent experimenting with one of the Uhri machines. After some four years of experimental and unsuccessful work by Stephens, never being able to make the boxes successfully, Mr. Healy came into the business. This was in 1894. Thereafter he was the president, manager, and chief stockholder in different companies which endeavored to develop such a machine. After the Uhri and the Knutson and Uhri machines had been experimented with and abandoned, Healy employed Rosback, an expert mechanic and inventor who had a shop in Chicago; his efforts were continued over several years and he built several successive machines. Eventually they were all discarded as failures and dismantled or thrown out in Rosback's yard. Healy then brought in another expert mechanic Flora. He also developed a machine which, in turn, was abandoned. The various machines built by Rosback and Flora were patented by the patents shown in the record and became the property of Healy's company. (Flora was 775,622 1904 see record, Gibbons Case.) All this work upon the Uhri, Rosback, and Flora machines had been with reference to the bent veneer type of box; the four sides were made of one continuous sheet of veneer, which would bend around the corners, and the bending might be aided by cutting the sheet part way through. This type of box was useful and was approved in many trials, but it could not be made at a profit and in competition with the ordinary wood box, unless it could be turned out by automatic machinery. All these machines over all these years were of the Rosback type that is, so far as they accomplished this purpose at all, they used a single rigid four-length cleat until the side material was attached, and then severed the cleats.
At about this time, in another effort to succeed, Rosback was again employed by Healy, resurrected some of the machines from his yard, brought them to the shop doing Healy's work, and spent some months with them. This also failed, the mitering and stapling machines were junked, and the blank machine dismantled. Up to this stage, in 1904, Mr. Healy, since he came in, had spent 10 years of his efforts and $62,000 of his money (as well as whatever capital others had furnished), and no machine had ever been able to run for a day in such a way as to make boxes that could be sold at *834 a profit. A few hundred boxes of this type had been made and sold, but they had been "tinkered by hand" after they came from the machine the machine product itself being too imperfect. Mr. Healy's total return had been $42.31; and the failure of all these machines and patents was demonstrated, unless vital changes and improvements could be made.
Early in 1904, Mr. Healy interested Inwood, who came to the plant and undertook to make the machine practical. Inwood associated Lavenberg with him. The contract relations between Healy and Inwood are not now important. Healy still believed that there was a great field open, and that he could control it through some of the claims of the existing patents, plus the necessary improvements. Inwood and Lavenberg devised certain betterments and, in May, 1904, placed a machine shop order for a new machine of the Rosback type. They still contemplated the single rigid cleat, to be severed by a second and separate operation. In this connection, they both testified and the testimony is not only undisputed but considerably confirmed that they observed that certain difficulties could be avoided by using the old preformed cleats, but they concluded that an attempt to do so would bring greater difficulties than benefits into the manufacture and that it was impracticable. They therefore adhered to the old form. A few weeks later, the idea that the performed cleats could be temporarily assembled and held in a former while the permanent fastenings were applied occurred to them. They developed this idea by drawings into a definite form. They became satisfied that it was practicable, and in June directed the machine shop to make changes in the existing order and to construct the necessary preliminary workholder. The machine was completed and delivered in October, 1904, was found to work successfully, and patents were at once applied for. The commercial growth then began. By 1908, the output was 2,000,000 boxes, 4,000,000 in 1909, and 6,000,000 in 1910.
Seldom does a record show such a demonstration of the practical value of what seems like a simple change. The preformed cleats were there, various workholders for temporary assembly were familiar to all mechanics, and these things had been at hand all this time; yet during ten years of effort by capitalists and inventors and the expenditure of large amounts of money, Stephens, Uhri, Rosback, and Flora,[3] and indeed for a time Inwood, had not seen what is now said to have been always obvious. Their conduct is far more convincing than the present opinion of some expert, who thinks that he would not have been so stupid.
It is not a new simile to say that the progress of an art is travel along a road. When Rosback, Uhri, et al., as road builders, saw that an automatic machine was the point they must reach, familiar preformed cleats, in connection with some temporary assembling means, presented the direct road to that terminal, and to take that road must have been the first thought of every traveler; but they all believed that this direct road was so obstructed as to make passage impracticable, and so they undertook a troublesome detour, which they never did succeed in opening and improving. The public would not travel, and the whole road remained unused. For 10 years the builders tried to develop their detour, to make it a good road, but failed. Then these patentees found an effective, though simple, means of overcoming the obstacle, opened the direct road, and the detour was abandoned. It would be a bold critic who would say that all the older builders were incompetent, because the way to clear the road had always been obvious.
The practical advantages of the change were that the cleats could be preformed in quantities and very cheaply; that short and otherwise waste pieces could be used; that the second operation for mitering the cleats, with all its time and expense, was avoided; that other union forms than miters could be used (as mortise and tenon); and that the inventors also avoided one of the chief troubles of the old machines, viz. that in step-mitering the ends from the solid cleat and in partially slitting the continuous sheet, the knives were continually cutting through the sheet in spots and cutting into the wires, injuring the product and dulling the knives. There was a saving of 75 per cent. in the cost of the lumber for the cleats, and this, with the savings in time, labor, and repairs, created a large and successful business out of what had been for 10 years a failing struggle. We are convinced that the advance step taken by these patentees *835 came well over into the field of patentability, as defined in the leading cases, too familiar for exhaustive citation, and as applied to a new association of old elements leading to a new result. See, e. g., Washburn & Moen Mfg. Co. v. Beat 'Em All Barbed-Wire Co. (Barbed Wire Case) 143 U.S. 283, 12 S. Ct. 443, 36 L. Ed. 154; Krementz v. S. Cottle Co. (Collar Button Case) 148 U.S. 560, 13 S. Ct. 719, 37 L. Ed. 558; Diamond Rubber Co. of New York v. Consolidated Rubber Tire Co. (Grant Tire Case) 220 U.S. 428, 31 S. Ct. 444, 55 L. Ed. 527; Minerals Separation v. Hyde (Flotation Case) 242 U.S. 261, 37 S. Ct. 82, 61 L. Ed. 286; Walker on Patents (6th Ed.) § 63.
What impresses us as the most serious challenge to patentability is the claim that, when the patent is tested by applying it to defendant's machine and giving it the construction necessary to make out an infringement, we see that all the patentees did was to take the Rosback machine and "add some extra pushers." By this is meant that the Rosback machine used two parallel conveyor chains; that each of these chains carried a projection or lug, which rested against the rear end of the long unsevered cleat, and so pushed the cleats and the attached side material through the machine; and that in defendant's final form of machine it has used these chains and these pushers, and in addition has only inserted in the chain, in selected locations, adjustable blocks, which project up between the ends of the preformed cleats, whereby each block serves to push the cleat which is immediately ahead of it.
If this statement about "adding some extra pushers" were correct, and were the whole truth, it would be unanswerable; but we think it neither exact nor complete. The pusher of Rosback was a pusher and nothing else; it had no substantial positioning function; the long cleat did not have to be positioned in any such secondary way; it was positioned with reference to the stapler by moving the chains. The defendant's machine uses on each side five of these devices which counsel now call pushers. They are all movable upon the chains for positioning purposes. The rear one, while it positions the rear cleat rightly with reference to the stapling which is to follow, may be said to have, directly or indirectly, a pushing function exercised upon all of the parts ahead of it. The next three of these devices have no necessary pushing function. The same pushing of all the cleats through the machine would follow if these three were of proper spacing size, but had no connection with the traveling chain; e. g., were dropped in from above. They are primarily positioners; such pushing function as they have is incidental to their positioning function upon the traveling chain and is not necessarily inherent; they are essential to the positioning, but the pushing could be done in a variety of ways. As the device is altered to adjust it to a box of wider or narrower sides, all these devices except the first are changed accordingly. They firmly hold the cleats, as against any backward tendency, in their proper place relation to each other. They pertain to and develop the thought of the inventors that the performed cleats can be assembled and held in position during the stapling operation. They would be an excrescence upon the Rosback thought of a single long cleat. When, in the Rosback machine, the four cleats later came into existence by severing, they were already positioned by the wire stapling, and the pushing operation had been finished. We can now see that at any time during their long efforts, Rosback et al. might have used the preformed cleats and assembled them upon the chains by these pushing, spacing, and holding blocks; but the art stood stationary all these years, while no one saw this solution. To put on "some more pushers," and utilize them as positioners of preformed cleats, was very simple after these inventors had pointed the way; but that is not persuasive, and, as we later point out, defendant has also added holdbacks, as well as pushers.
In reaching the conclusion of patentability, and giving reasonable force to the public acceptance and the creation of a new industry, we do not overlook the doubtless substantial part which another patent played in the commercial result. This was the so-called product patent of September, 1905, covering a blank for a box with overlapping edges, which could be constructed by this machine when including certain details. Until after the time when the machine invention was broadly complete and the order for its construction had been given, these patentees, like all who preceded them, had been planning for the manufacture of a box having for its side material a continuous sheet of veneer, which at the folded corners would turn at right angles with a bend, which might be accelerated by partial slitting; but later they observed *836 that the machine could be used to produce a box blank with four entirely separate sides,[4] and that when it was so used they could so space the sides and cleats that when the blank was folded the edges would overlap, with one edge flush with the adjacent side, according to the universal practice in building wooden boxes. The specification of the machine application shows that it might or might not produce this result.[5] At the same time came the application for a patent upon this specific product; that is, a box blank which would be built by this machine, provided it was arranged for separate sheets with overlapping edges instead of a continuous side. This box application went through without difficulty, and the patent was issued in 1905 and reissued in 1907, expiring in 1922. It was sustained in various litigations. Undoubtedly, for most purposes, this overlapping edge box was an improvement over the continuous side box. The public favor which has been mentioned was doubtless due, in some measure, to this overlapping edge feature. Non constant that the older form would not have been accepted on the market with favor. It is clear that the overlapping edge box could not have been a market success unless it had been the product of an automatic machine, which enabled it to be made cheaply. To give the credit for success to the preferred form of the product and to ignore the means of its manufacture, which means alone made the success possible, is an inference which we cannot reach. Perhaps the credit should be divided; enough remains to be persuasive of invention in the machine.
It has been suggested that the disclosure made by the method patent left no room for invention in the machine patent. This suggestion cannot be of force, since the patents were copending and one is not prior art as against the other, even if they had not both been issued on the same day. See cases infra. If by the suggestion it is meant that the primary thought was in the method and that little or no invention was needed to make a machine, once the method had been conceived, that may be true; but that cannot militate against the validity of the machine patent when the two are practically simultaneous, for in such case the concept involved in the method is an important and perhaps a vital part of the machine invention. The concept which underlies and precedes the machine may or may not involve an independently patentable method; in either event alike the machine is patentable.
Nor can the product patent invalidate the machine patent as double patenting or by any analogy to that theory. The monopolies granted by the two patents were not coextensive.[6] The machine patent pertains to assembling preformed cleats and holding them in proper relationship while side material and binding wire are attached and this, no matter whether the side material is a continuous sheet to be folded or slit, or both, or is a plurality of separate sheets or boards, whose edges overlap, or boards which do not overlap but leave either an edge contact corner or a slightly open corner, or is boards or sheets or slats which do not come near the corner but produce a crate; all these forms can be made on the machine; most of them have been, in substantial quantities. In other words, this machine patent pertains to cleat treatment. The product patent pertains to the selection and treatment of side material, and it covers and protects only one of the several forms of product of the machine. Plainly, the machine invention was the underlying thing, quasi generic; the product patent is a side issue, quasi specific. Even with patents relatively and completely generic and specific, the rule is established, at least in this circuit, that the generic patent earlier applied for will be neither invalidated nor limited by the specific patent later applied for and copending, even though the later application develops into the earlier issue. Much less will there be such avoidance or limitation, when the two patents are so far from coextensive that either one may be infringed without infringing the other.
To consider more generally and carefully the effect upon a later issued patent (as for the machine) of an earlier issued, simultaneously applied for, and copending patent (as for the product): Clearly if the monopolies granted in the two patents are identical, they would cover more than 17 years and there would be double patenting and the later *837 would be invalid. Even when they are not identical, but the small field of the earlier (or later) is included in the larger field of the later (or earlier), there is, as to this small field, a double grant and a monopoly extension practically, though not in form. Here are to be classified patents relatively generic and specific, as well as those machine, process and product patents which overlap, each on another, but in part only. It has been thought that in the cases thus classified the later patent is invalid, at least to the extent that the monopoly is extended. Language is to be found in Miller v. Eagle Mfg. Co., 151 U.S. 186, 14 S. Ct. 310, 38 L. Ed. 121, which may be claimed to support this conclusion, for, from the facts stated, it could be argued that the earlier patent was specific; the later, generic. But the court did not so regard them; and no such construction of Miller v. Eagle Mfg. Co. is now open to us. As early as the second Julien Case (C. C.) 41 F. 679, 685, Judge Coxe had pointed out that, though one patent may seem to have a broad claim and another a specific claim based on the same disclosure, yet analysis may show that there is really no such difference but the scope of each is the same; hence, one is invalid. This is the very principle developed and applied by Mr. Justice Jackson in Miller v. Eagle Mfg. Co. Accordingly it was held in the (New York) Thompson Co.-Ohio Co. Case, 71 F. 396 (C. C. A. 2. Wallace, C. J.), that where there are generic and specific inventions in the same structure, they may, at the option of the patentee, be secured by separate patents, and both are valid. Judge Wallace says (71 F. 405):
"Some observations in Miller v. Manufacturing Co. [Eagle] seem to have created some misapprehension of the scope of that decision on the part of the profession, but the principles enunciated in the opinion are so plainly stated that those observations, when considered in their application to the case before the court, ought not to be misconceived. The court decided in that case that the two patents to Wright were in fact for the same invention, and consequently the later patent was void."
In the Ohio case of the same title which reached this court and was reported in 80 F. 712, page 727, 728, the present Chief Justice, speaking for this court, said:
"The case upon which counsel for defendant chiefly relies to support his argument is Miller v. Manufacturing Co. [Eagle], 151 U.S. 186, 14 S. Ct. 310, 38 L. Ed. 121, where it was held that when two patents issued to the same patentee for the same invention the second patent was void for the reason that the new or later patent would prolong the monopoly beyond the period allowed by law. * * * In other words, the division of the original application into two patents was nothing more than an attempt to patent, as two separate inventions, the same device when discharging different functions. * * * Hence it follows from the propositions above quoted from the opinion in the case of Miller v. Manufacturing Co. that a patent for a generic invention is not avoided by the fact that a prior invention [patent] has been issued for a distinct improvement on that invention, provided always that the language of the application for the first patent and the circumstances of filing it are not such as to dedicate the generic invention to the public. The case of Miller v. Manufacturing Co., therefore, instead of sustaining the claim made for it, is distinctly an authority to the contrary."
On pages 726 and 727 of 80 F., Judge Taft had considered, on its merits, the argument that the later patent was invalid because it extended the monopoly, and disposed of it by saying "To our minds, this conclusion is reductio ad absurdum."
These Thompson-Ohio Cases have been frequently followed and applied, and we think the rule is now universally accepted that, as to typical and properly separable generic and specific patents, the incidental extension of the smaller monopoly does not affect the validity of either patent. In Century Electric Co. v. Westinghouse Electric & Mfg. Co. (C. C. A. 8) 191 F. 350, 353, Judge Sanborn says:
"His [the inventor's] is the option to secure all these inventions by a single patent, or by many patents, and the fact that he describes all of them in his application or specification for an earlier patent to secure one or more of them, does not invalidate a subsequent patent to him for those inventions there described but not claimed. * * * The sum of the whole matter is that while an earlier patent avoids a later patent to the same patentee for the invention claimed and secured by the former it does not invalidate a later patent to him for a distinct, different and separable invention whether generic or specific, whether an original machine or process, or both, or an improvement thereon which is not actually claimed * * * by the earlier patent."
This court further discussed the subject and applied the rule in Dayton Co. v. Westinghouse Co., 118 F. 562, 573; and Cleveland *838 Co. v. Detroit Co., 131 F. 853, 858. One of the latest applications is that made by the Second C. C. A. in Traitel Co. v. U. T. Hungerford Brass & Copper Co., 22 F.(2d) 259, 262. Judge Learned Hand, speaking of the effect of an earlier patent upon a later one of the same inventor and the resulting defense of double patenting, says:
"The defense is good only when the claims are the same, and the discussion in Miller v. Eagle Mfg. Co., supra, was really of this point. * * * It is, indeed, quite true that Calkins gets a protection for more than 17 years for the continuous edged strip. He got it during the existence of the original generic claims, and he gets it from the claims in suit. * * * In view of the fact that the date of issue of an application is beyond the control of the applicant, the chance of so extending the monopoly is disregarded, and it makes no difference which of the applications issues first."
So far as there is lack of perfect analogy between the case of the earlier and the later specific and generic patents, resulting in the extension of the smaller monopoly, and the case of the earlier and later machine and product patent, resulting, it is said, in an extension of one monopoly or the other, the distinction makes the defense of double patenting less plausible in the latter class of cases, because, in the former class, the monopoly extension is inevitable, while in the latter class it is only more or less probable in a commercial sense. If the machine, or process, as described and claimed, will turn out a particular product, which product also is new and patentable, and if the product is one which can be constructed by another process, or manually, or by another machine, the machine and product are separate inventions, separately patentable. It may well be that the machine or its equivalent is essential to make the product cheaply enough to be marketable, and thus, in a commercial sense, a later issued machine patent will extend the monopoly of the product patent; but this may or may not be so; and that it happens to be so in a particular case cannot invalidate an otherwise valid patent.
That there is no double patenting unless the claims are identical in substantial effect, see Walker (6th Ed.) § 225.
Mosler Safe Co. v. Mosler, Bahmann & Co., 127 U.S. 354, 8 S. Ct. 1148, 32 L. Ed. 182, is cited to the effect that an earlier issued product patent will invalidate a later patent to the same inventor for the process by which the product is made. The language on page 361 of 127 U. S., 8 S. Ct. 1151, is especially relied upon. The subject-matter of the patents was an angle bar bent for use in making safe frames. There were involved the patent for the formed angle bar, bent by the process described, and the patent for the process of bending the angle bar to produce the patented form. The Circuit Court (22 F. 901) had found that the product was not patentably new, and it was so obvious that, if this were true, there could be nothing patentably new in the process, that the Circuit Court dismissed the bill as to both patents, without specifically stating its reason for holding the process patent invalid. The Supreme Court approved and adopted the Circuit Court opinion, and then added justification for the result as to the process patent. It is not entirely clear upon which of the facts stated this result is intended to be placed. It is mentioned that the method was purely mechanical; but this ought not to be taken as the basis of the result, for, if so, it would be overruled in Expanded Co. v. Bradford, 214 U.S. 366, 29 S. Ct. 652, 53 L. Ed. 1034. Nor can it be thought that the result depends upon the fact that the process had been disclosed in the earlier filed product application, for then the case would be in conflict with the now unquestioned rule that among copending applications by the same inventor one is not prior art as against another.
The case must therefore, we think, depend upon its conclusion that the product and the method patents were not for two inventions, but for one. So construed, the opinion is consistent with the application of the settled principles which have been referred to; otherwise construed, it is not. Indeed, it is so clear upon analysis of the two patents, that the method could produce only the patented result and that the patented result could be produced only by the method, that the Patent Office, under the practice as now clarified, would not permit separate patents. The Mosler Case therefore is not pertinent where the use of the claimed machine may or may not result in the patented product, and where the claimed product can be made otherwise than by the patented machine. The machine will make the patented lap corner blank, or the unpatented open corner blank, continuous side blank or crate blank. The patented product can be made by this machine, or wholly by hand, or by the Rosback machines, or by hand feeding to a stapling machine.
As noted above, it may occasionally happen, as it did here, that the specific patent is so appealing in its specific features that it gets a practical market monopoly of the subject *839 matter, and that the generic patent of longer term will extend that practical monopoly; but if this is an evil sometimes it may be and sometimes not the remedy is with Congress, not with the courts. Clearly, during all the interval between the machine invention in 1904 and the issue of the patent thereon in 1915, the public had unrestricted right to use the machine, excepting to make the overlapping edge box, and, since the expiration of that product patent in 1922, the public has had the unrestricted right to make that box, by any machine or method except the patented one.
It seems to us entirely plain that the product patent must be immaterial as to the validity of the machine patent, unless the former was prior art or the latter was double patenting; and equally plain that neither condition exists.
Nor should there be any confusion or prejudice on account of the long delay of this machine patent in the Patent Office. After two years of diligent prosecution (in which the delay was mostly by the Patent Office), it went into interference with Brown, and this interference went through its due course before all the Patent Office tribunals. It is not suggested that there was any substantial delay for which the patentees were responsible. The final decision in their favor by the Court of Appeals of the District was in 1912. The further prosecution thereafter to issue was in ordinary form and was at least reasonably diligent if not unusually so. Here again, the delays resulting from interference or other adversary proceedings in the Patent Office, present an evil which can be remedied only by Congress. Patentees cannot be penalized for submitting to delay which they cannot avoid, nor are they under any obligation to be more diligent than the statute and rules require. U. S. v. American Bell Tel. Co., 167 U.S. 224, 17 S. Ct. 809, 42 L. Ed. 144.
We come now to the matter of infringement: We take as typical of the claims sued upon No. 25, which reads as follows:
"A machine for making box blanks comprising: Work-controlling means having cleat-positioning means to receive sectional cleats in parallel lines, and spacing means to space cleats endwise from each other in each line, preparatory to connecting said cleats in their spaced relation, said work-controlling means being arranged to receive side material to be secured to said cleats; fastener applying means for securing binding means to the side material and cleats controlled by the work-controlling means; means for relatively feeding the work-controlling means and fastener-applying mechanism for securing said binding means across the intervals between said spaced cleats to secure them together in their spaced relations to form a foldable blank."
The mechanical difference between the form shown in the patent drawings and defendant's third form, is that the patentees assembled their cleats and positioned them in the workholder, which was independent of the moving conveyer, and then put the holder on the conveyer to be moved along.[7] In substance this workholder consists on each side of four pockets, in each of which a cleat fits closely enough to be held in true position while the cleat and overlying side material are passed under the stapler. The defendant has united its workholder and conveyer into one structure. It does this by attaching to the conveyer chain, at selected spots, five upstanding and outstanding blocks. The conveyer chain runs between side channel walls, which correspond to the sides of the patented workholder. These blocks therefore divide that channel space into four pockets, each of which holds and positions its cleat, as it and the side material are advanced under the stapler. If we place side by side plaintiff's workholder upon its conveyer and defendant's conveyer with its four pockets thus formed, each just before passing under the stapler, we see no substantial distinction in form or in operation, and only one distinction which merits notice. This is, that in the patented machine the dividing partitions positively position the cleats in both directions, that is, each partition bears against the cleat behind it as well as the one ahead; while in the defendant's structure the positive contact of the positioner is only with the cleat ahead of it; there is no direct contact between this positioning block and the following cleat. This device is adapted for mortise and tenon joints and the cleats therefore have a side shoulder near the end, and the defendant's positioning block avoids the end of the cleat and reaches out to engage with this side shoulder. Thus there is between the extreme cleat ends a small free space, and hence possible slight forward motion of the cleat in its groove and away from its rear contact; there is practically a forward stop, though by less direct means. The positioning blocks carry upward lugs which act as spacers between the sheets of side material, and all four sheets must therefore move forward as a unit. *840 For the cleats to slip forward against what might be made the normal retarding friction on four sides would not be very probable. Further, and as a more direct stop, the defendant's machine carries, as the patent does, means for exerting a spring side pressure against the cleat just as it comes under the stapler. This, in different forms, is called the clincher or the detent. It forces that cleat to exact position laterally and, either by projecting in front of the cleat or by side frictional effect, forces it back against the rear positioning block, and to exact position longitudinally. Thus at the critical moment when the stapling is done, defendant's cleats are positioned as absolutely and as accurately as in the patented form. If the normal retarding friction does not hold the cleat back against the block as it may not against the intermittent motion such a front detent stop would seem essential. As far as we observe, it was provided in all forms of defendant's machines.
We find nothing either in the state of the art or in the specifications which requires a claim like 25 to be so limited to the use of a pocket which positively confines the cleat in all four directions as not to reach one which confines positively in three directions and sufficiently for all practical purposes in the fourth. In any fair degree of equivalency, the call for "spacing means to space the cleats endwise from each other" is met by the positioning means which defendant uses. This would be true even if the contemplated spacing were perfected only in each pocket successively, as it arrives at the only spot where perfect spacing is important. Though the patentees' advance step was such that its inventive character was plausibly challenged, yet after its patentability is determined, we see that it was the foundation of a large development in the art, and that it has a relatively pioneer status. Its claims are entitled to a fairly liberal scope.
As to the suggestion that the conveyer chain form of defendant's machine is so different from the patented form that it should not be contemplated as intended to be reached, we may cite a paragraph of the specification which is here quoted in the margin.[8] It would almost seem that the draftsman had prevision of the specific form which defendant's infringement was to take, many years later.
The method patent was issued upon a divisional application, filed in 1912, by separating it from one of the mechanical applications. The court below held it invalid for laches, applying what that court understood to be the rule of the Splitdorf Case, 264 U.S. 463, 44 S. Ct. 342, 68 L. Ed. 792. This was before we had by our Wagenhorst Case, 27 F.(2d) 27, adopted our interpretation of the Splitdorf decision. Under that interpretation, it is not seen how this delay could be sufficient reason for invalidating the patent. The method was fully disclosed by both mechanical applications. Method claims were added to the machine application within a few days after it was filed. The Patent Office said that these method claims covered only the operation of the machine and were therefore improper. Applicants then cancelled these claims. This was rather plainly a mistake, as the Patent Office later decided when it approved similar claims; but, mistake or not, this action by applicants was, of course, subject to reconsideration and disavowal by them at any time during the progress of the application, before other rights had intervened. It was an abandonment only until they should change their minds and reclaim. While in that situation, the interferences developed upon both applications. While these were pending it would have been technically possible to file a divisional method application, but this would have been a useless thing, as the decision in the existing interferences would decide also who was entitled to method claims, if any one was. After the interferences, the method claims were restored to both applications, and were there maintained while the Patent Office, conceding in effect its former error, held that the claims could be made, but required a divisional application. This was duly made and prosecuted to allowance. No adverse claim appeared to the method or to any machine which would involve the method until, at the earliest, two years after the method patent issued.
*841 This case is plainly not within the Splitdorf rule, as we understand it, either by parallelism or by strict analogy. Defendant points out, as a general analogy, that the machine and method had been in public use by the applicants, who became the patentees, and under licenses from them, all for more than two years before the divisional application was filed. This is true. When it is remembered that the divisional application is normally entitled to the benefit of the original filing date, it is not easy to see how public use under license from the applicants can amount to that laches under which adverse rights may arise; or how the Splitdorf rule can apply unless the delay is after the public appearance of an adverse claimant, or, perhaps, one who may be an adverse claimant; but we find it unnecessary to decide that question. It is further apparent (from the claims) that the question of double patenting, as between product and method is not the same as between machine and product; and also that whether this method, if the claims are construed broadly enough to make them separable from the operation of the machine, involves any patentable invention, is a debatable question; and, if they must be confined to a method which can be practiced only by the machine or its equivalent, again doubt at least arises.
Coming to the workholder patent, we meet a similar question. If it is to be construed broadly enough to reach the defendant's structures, and is to be disassociated from its status as an essential part of the machine patent, the presence of patentable invention can be forcefully challenged. We see no occasion for deciding these questions of validity of the method patent and the validity and infringement of the workholder patent. So far as has been pointed out or as we observe, there has not been and is not likely to be any commercially important infringement of either of these patents, which does not also infringe the machine patent. They were all issued on the same day and expire at the same time. Questions of accounting, so far as we see, are the same under claim 25 of the machine patent, as we have construed it, as under any other claims of that or the other patents; and while theoretically the right to injunction might be important, we see no indication that it will be of practical value, because other methods would cost too much. In this situation we have frequently declined to enter upon difficult questions which are practically moot. If we misjudge this situation, and our further consideration of the other claims in suit under any patent is reasonably necessary to settle the controversy between these parties in its present or in any probable modified form, counsel may suggest the necessity and we will give it attention.
In the case against Gibbons, involving these three patents and the lap-corner box patent, the District Court in Illinois dismissed the bill without opinion. The Circuit Court of Appeals [25 F.(2d) 363, 364] affirmed this decree. Appellant's counsel now insists that a reading of the original opinion of that court demonstrates that the court considered the lap corner later filed and copending patent as a part of the prior art, supporting its conclusion that the machine invention was not patentable. This analysis by counsel may or may not be correct; but the original opinion and the per curiam upon rehearing together demonstrate that the conclusion of noninvention was finally based upon the undisputed prior art. As we read the opinions, that court thought, as we do, that the substitution by the patentees of the one-machine assembly and treatment of preformed cleats for the Rosback one-cleat, two-machine combination possessed the true inventive character[9]; but it allocated this merit to the lap-corner patent, which turned out to be the popular form of the product resulting from the change in the machine. We have indicated our reasons for thinking that the dominating new result, which gave inventive character is to be found in the new combination of machine elements which permitted the automatic production of any kind of a wirebound box, rather than in the particular form of product box which proved to be the best seller. Convinced as we are that this is the right view, we find ourselves constrained to reach a different conclusion from that adopted in the Seventh Circuit. Of course, this is not to deny that the lap-corner box blank, as patented in its specific form, had specific patentable merit; but, even if the box patent was erroneously sustained, that error would not justify another.
Out of the very numerous other points made in appellee's 250 pages of briefs, we find none requiring further discussion.
The decree below will be reversed, with instructions to enter the usual decree for injunction and accounting, based on claim 25 of the machine patent, and to dismiss the bill, without prejudice, as to the method and workholder patents. The other claims of the machine *842 patent may be ignored in the decree. Appellant has largely succeeded in this appeal, but in large part has not. We award it half costs in this court. Costs below to this time are discretionary in that court. Whether our disposition of them here furnishes analogy helpful as to costs below that court will decide.
KNAPPEN, Circuit Judge (dissenting).
I am unable to agree in the conclusion of the majority opinion, which reverses the decree of the District Court (which found the machine patent No. 1,128,145 invalid, if infringed by defendant), and directs decree for injunction and accounting based on that patent, and dismissal of the bill without prejudice as to the method and workholder patents.
I am constrained to agree with the opinion of the Circuit Court of Appeals of the Seventh Circuit in the Gibbons Case, 25 F.(2d) 363 to 366, inclusive, which holds the machine patent (as well as the workholder and method patents) invalid for lack of invention. The statement of reasons and the logic of the Circuit Court of Appeals of the Seventh Circuit in finding those patents invalid, both as stated in the original opinion (pages 364-365) and in the per curiam opinion (pages 365-366), seem to me convincing.
This makes it unnecessary for me to discuss the art prior to Inwood and Lavenberg, except to say that while in making the boxes by hand, preformed separate sheets and premitered cleats had been used before the patents in suit had been applied for, they were not regarded as practicable for machine manufacture. It remained for Inwood and Lavenberg to discover that the use of such preformed separate sheets and premitered cleats in machine manufacture was practicable. This use of preformed separate sheets and premitered cleats runs all through the four patents. The contrast between the method of manufacturing wirebound boxes under the patent and the then present method is plainly set up in the reissue patent No. 12,275. The drawings of the patent, in connection with the language of the specifications, clearly indicate the process to be employed in making the product of the patent. The claims of the reissue patent seem to me to indicate plainly to one skilled in the art the process to be employed. As concluded in the original opinion of the Circuit Court of Appeals of the Seventh Circuit in the Gibbons Case, 25 F.(2d) at page 364: "The entire story of the appellee's invention is told in this reissue patent, and that the solutions of the problems demonstrated in the other three patents were obvious, and involved the exercise of merely ordinary mechanical skill"; and as stated in the per curiam opinion, 25 F.(2d) at page 366: "It [the reissue] was the culmination of their [patentees'] efforts, and therein lies the entire story of their invention, as we have stated in the original opinion." Upon the theory of that opinion, the question of co-pending application as prior art is not involved. Nor am I impressed that the validity of the machine patent is governed by the rule that generic and specific inventions may, at the option of the patentee, be secured by separate patents. The controlling objection is not that there has been double patenting, but that there is lack of invention in the machine patent. As stated by the Circuit Court of Appeals of the Seventh Circuit, 25 F.(2d) at page 366:
"The moment the long cleat was discarded and supplanted with short ones,[10] the workholder and machine readily appeared as the logical solution of the mechanic to the problem confronting him. We cannot believe that in this situation invention was achieved, or that anything more than the exercise of mechanical skill was involved. Novelty and utility in the combination did not arrive until a further improvement was made, that of preforming the side materials and so arranging them that they would overlap. This last step accomplished an improvement, and the patent for the combination, with this improvement, was by us sustained."
Nor is it important that, as asserted, the workholder and stapling machine inventions were conceived earlier than the reissue box patent; because previous to the invention of the box patent the other two inventions were wholly impracticable. After that invention they became useful, but only by virtue thereof.
In view of what is above stated as to the validity of the machine patent, the question of validity of the method patent (No. 1,128,252) is perhaps not very important as the issue is developed on this present review. It would seem enough to say that, in my opinion, the method patent is invalid for the further reason that it is for the process which results in the product which is covered by the reissue patent. Mosler Safe & Lock Co. v. Mosler, Bahmann & Co., 127 U.S. 354, 8 S. Ct. 1148, 32 L. Ed. 182; Miller v. Eagle Mfg. Co., 151 U.S. 186, 187, 14 S. Ct. 310, 38 L. Ed. 121. Nor does the rule forbidding priority in case of copending applications apply to this defense. Indeed, in Mosler Safe & *843 Lock Co. v. Mosler, Bahmann & Co., supra, the applications were copending. The question as presented in the instant case is readily distinguishable from Century Co. v. Westinghouse Co. (C. C. A. 8) 191 F. 350, 352 et seq. The conclusion that the method patent is invalid for the other reason herein suggested makes it unnecessary to consider the correctness of Judge Raymond's holding that the patent is invalid by reason of delay in filing divisional application; nor the effect upon that defense of our decision in MacGregor Co. v. Vaco Grip Co., 2 F.(2d) 655. Nor is it necessary to consider the question of infringement by appellees as to either patent.
[1] It took 66 figures and 12,000 words to describe this machine.
[2] One machine, making boxes 24" by 12" by 18" (for Quaker Oats) turned out nearly four boxes per minute; another, 300 per hour.
[3] In 1904, just before and even after Inwood was brought in, Healy spent $1,200 building Flora's solution of the trouble. Flora adhered to the single long cleat idea, but discarded even the one "pusher," using instead opposite feed rollers.
[4] It would seem that with beveled cleat ends a lap corner is impossible unless the cleats are at unequal distances from the side edges; while with step mitered ends, the lap corner is inevitable.
[5] It shows both the continuous sheet sides and the four separate sides, as alternative. It does not mention the lap corner, nor clearly disclose an adjustment particularly adapted to that result.
[6] A typical claim of the product patent is: "A wirebound foldable box blank, comprising a plurality of straight edged sheets, wires secured to and connecting said sheets, and cleats secured to said sheets, to terminate short of edges of the latter a distance equal to the thickness of a sheet, whereby, when folded at a right-angle, the end of one sheet overlaps the end of an adjacent sheet." Compare with claim of machine patent, infra.
[7] In specific form, their conveying means was a pinion-driven rack, on which the workholder rested.
[8] "We wish it to be understood that this invention is not to be limited to the use of removable carriages or formers such as described, and which are in the nature of adjuncts applicable to the wire feeding and stitching machine. Any means for positioning the separate parts of the blank-material in proper relation, as broadly defined in the claims, may be substituted for the formers, or carriages described, whether removable from and replaceable in the machine, or otherwise. In operating the apparatus shown and described, the carriages or formers may be fed into the machine one after the other in a manner to obviate the necessity of stopping the machine after each blank is stitched. In this way, an endless chain of cleat and sheet holders and positioners, with the material thereon, may be caused to pass between the bed frame and stapling devices, and keep the machine working to its fullest capacity."
[9] The court said: "It was the patentees' fundamental difference in the order of assembly and methods of construction that contributed most to their invention over Rosback."
[10] Italics mine.
