along a downward incline or chute, L, to the rollers, B. A bridge, H, extends between the carrier, C, and the rollers, D, E. Below the roller, D, is a pan. N, for catching any juice which may run from the cane. The operation is as follows: Cane is carried on the cane carrier, C, and delivered on the bridge, H, thereby bringing it within reach of the teeth of the breaking and cutting rollers, D, E. These preferably run faster than the carrier, C, and will draw the cane between them, their teeth cutting and breaking it and partly squeezing it. Whatever juice is expressed will fall into the pan, N. The broken cane will fall on the incline or chute, L, and slide along the latter to the rollers of the grinding mill or rollers, B, which press out as much of the remaining juice as possible. By regulating the distance between the breaking and cutting rollers, D, E, the cane may be only broken, and the teeth made to only partly cut into it, or the cane may be cut up into short pieces, as may be desired. I prefer the teeth of the rollers, D, E, to run diagonally or in zigzag lines, because, as the cane is usually thrown on the carrier in various posi tions, it frequently reaches the rollers in a position parallel thereto, and, entering between the teeth thereof in the same parallel line would be delivered from the rollers without being cut; but, if the cane is placed on the carrier so as to reach the rollers in a position perpendicular thereto, straight teeth running parallel therewith can be used. The two cutting rollers may be placed in close proximity to the pressing rollers, and be even placed in the same pedestals therewith. They may be placed at any distance from the pressing rollers, or even in the position perpendicular to that of the pressing rollers. It may be desirable to have more than two cutting rollers used, so as to break and cut up cane in very small pieces." The machine used by the defendants, and which it is claimed by the appellants infringes theirs, was manufactured by the Birmingham Machine & Foundry Company, of Birmingham, Ala., but has not been patented. In their answer the appellees denied infringement, and submitted the following additional issues: "(1) Nonpatentability of the matters described in the specification and in the claim in the patent in suit, if tested by its own literature, or by the state of the art to which it relates. (2) Invalidity by reason of anticipation, prior publication, and public knowledge of the device, combination, or mechanism sought to be patented, and its prior public use before alleged invention of Krajewski." The cause was heard upon the pleadings and proofs, and the court directed a decree dismissing the bill. From the decree thus rendered the appellants have

appealed to this court.

Edwin H. Brown and J. D. Rouse (Wm. Grant and Charles S. Jones, on the brief), for appellants.

J. R. Beckwith, for appellees.

Before PARDEE and SHELBY, Circuit Judges, and MAXEY, District Judge.

SHELBY, Circuit Judge, after stating the case as above, delivered the opinion of the court.

There has been in use for a great many years what is known as the "three-roll mill" for compressing the juice from sugar cane. It consists of three rolls, mounted and framed, rotated by gear wheels. The rolls were made sometimes with and sometimes without grooves. The purpose of the grooves was to firmly hold the cane as it was drawn between the rolls. Cane of full length was passed between the upper and front lower rolls, and the crushed mass was then made to pass between the top and rear lower rolls. The juice pressed out was caught in a pan, and the bagasse removed by a belt. The mill was sometimes so constructed as to place the rolls upright in the frame, and then, when the rolls contained grooves or channels, their function was not only to firmly hold the cane, but also to receive and carry the expressed juice to the pan or collecting trough below.

In the use of these mills, unaided by any machine or process to prepare the cane, there was great loss, on account of the failure of the mill to express all of the juice. This is shown by the evidence, and seems to have been a matter of common knowledge. To secure better results, it was evident that the cane must have preliminary treatThe record shows that several attempts preceded Krajewski's to make a machine to prepare the cane for the mill. The British patent, No. 2,586, of 1882, is a device for splitting and partially crushing the cane. Faure's patent, No. 250,720, Ferron's patent, No. 279,235, Chapin's patent, No. 321,007, Newell's patent, No. 186,100, and the Hungerford patent, No. 346,817, are all devices intended to prepare the cane for the mill. Each of them may briefly be described as a cane shredder. They are devices for tearing or shredding the cane. The result was usually produced by toothed rolls geared to rotate at different speeds. The result of the work of the shredders is to break or tear or split the stalk longitudinally, and so reduce the work of the three-roll mill. None of these machines was constructed to cut or break the cane in short lengths transversely. The shredders did not cut the cane in short lengths, so as to secure a uniform feed of it to the mill. The Krajewski machine is not a shredder, but a machine for breaking or cutting cane. It is not made to tear, shred, or split the cane, but to cut or break it transversely in short pieces, so that the three-roll mill may be fed with it in a uniform manner. It does not tear the cane longitudinally, or lacerate the fibre, so as to produce a mass of shredded stalk, irregular in volume. The result produces pieces transversely cut or broken. The fact that the Krajewski machine produced results so different from the shredders indicates a difference in its parts and mechanism. The Krajewski machine contains two rolls having teeth which are zigzag and intermeshing, and when intermeshed equidistant from each other, and of such sharpness that they can transversely cut or break cane into short pieces. The rolls are maintained in position and proper relation by gear wheels, so that the teeth of the two rolls do not come in contact side to side, but pockets or spaces are left between them. Each roll with its teeth is different from the shredders. The two, being used together, are wholly unlike the shredders in the result produced. The Krajewski machine is much used in connection with the three-roll mill. When so used the function of the mill is to press the cane which is already cut and broken by the machine. The evidence shows that the use of the machine greatly increases, and probably doubles, the output or capacity of the three-roll mill. The machines have been on the market since 1886. They have been sold and are in use in Texas, Louisiana, Cuba, Mexico, Hawaiian Islands, and British West Indies. It has been introduced into nearly all sugar-growing countries, notwithstanding the fact that the cost of establishing a plant, including the three-roll mill, is from $8,000 to $13,000, according to the size of the rolls and the conditions under which the plant is erected. The Krajewski patent contains the following claim:

"The combination of rollers provided externally with more or less sharp teeth having a zigzag trend in the direction of the length of the rollers, said

teeth intermeshing but not coming in contact with each other, and when intermeshed being equidistant from each other, said rollers being geared together in order to maintain their relative positions toward each other, substantially as specified."

A brief comment on some of these words and phrases is required. The patent does not require the rolls to be formed in sections. The specification provides that the rollers are "preferably composed of a number of collars keyed to or otherwise fastened on the shafts." Features described as preferable do not constitute a limitation of the claim. "When the inventor says, I recommend the following method,' he does not thereby constitute such method a portion of his patent. His patent may be infringed, although the party does not follow his recommendation, but accomplishes the same end by another method." Sewall v. Jones, 91 U. S. 171, 186, 23 L. Ed. 275. It will be observed that the teeth of the two rolls are to intermesh, but not come in contact with each other. This language does not mean that the teeth of one roll may not come in contact with the other roll. The phrase "more or less" has been commented on as being vague and uncertain. The rollers are provided with teeth "more or less sharp." If the word "sharp" had been used alone, it would have been subject to the same criticism. The phrase "more or less sharp" is indefinite, and necessarily so, because of the limits and imperfection of language. Taking the words in connection with the context and the specification (the only fair way to consider them), they mean sufficiently sharp for the purpose sought (the cutting of the cane). The phrase is in current use to describe or qualify. Mr. Justice Field, in Glue Co. v. Upton, 97 U. S. 3, 6, 24 L. Ed. 985, uses the phrase: "But to render the article new, in the sense of the patent law, it must be more or less efficacious, or possess new properties by a combination with other ingredients." No word, perhaps, exists, that the learned justice could have selected, that would show with certainty how efficacious the article must be. And no word can be selected to show exactly how sharp the teeth must be. It seems to us sufficient to say "more or less sharp." and that this means sufficiently sharp for the purpose shown by the specifications. The question of infringement, which will be consid ered later, has, in some of the arguments presented, been made to turn on the meaning of the word "equidistant" as used in the patent. The teeth when "intermeshed being equidistant from each other," what is meant by these words? We must, of course, look to the connection in which a word is used, to get its intended meaning. The drawings and specification show that the teeth of one roll fall between the teeth of the other, and in the depressions between them. The purpose of this adjustment is, in part, to leave spaces on opposite sides of the teeth, adapted to receive the pieces of cane cut or broken off by the teeth. Was it material that these spaces should be of the same size on each side of the teeth? If slightly different in size, they could still receive the cane. This fact indicates that the meaning was that the teeth of one roll were to fall approximately rather than exactly midway between the teeth of the other. The word "equidistant" must, we think, include any position that is sufficiently midway to attain the purpose of the in

vention. The meaning of letters patent, like other grants or written instruments, must be ascertained by the language employed, as applied to the subject-matter. Robinson v. Sutter (C. C.) 8 Fed. 828; Adams v. Iron Co. (C. C.) 26 Fed. 324; Sayre v. Scott, 5 C. C. A. 366, 55 Fed. 971.

It is claimed by the defendants that the Krajewski machine contains no patentable novelty; that the combination is merely the aggregation of well-known devices. A machine must be constructed of known things. Unless the aggregation is clearly obvious, such as would be suggested to any one skilled in the art, it cannot be an answer to the claim of novelty in a machine consisting of many parts to segregate its parts, and say that each of them is well known and has been long in use. Invention would not often be equal to the test of originality and novelty if the parts of the machine invented are subjected separately to that test. The parts of a machine when segregated may be ancient and well known, and yet the completed device in the aggregate altogether novel. Originality sometimes consists in new combinations. If it be conceded that the machine is only a combination of old and known devices, the patent is not invalid, unless the combination was an obvious one for obtaining the advantages proposed,-one that would occur to any one skilled in the art. The idea of producing a combination or machine that would cut the cane in short lengths did not occur to any one of the many mechanics who designed and made the various shredders. They were intended to prepare the cane for the mill, but the machines made were shredders, and not cutters and breakers of the cane. To the mechanics making the shredders the combination of devices adopted by Krajewski was not obvious. When made, it appears to be very plain. It accomplishes the end sought in the art. It increases greatly the product of the three-roll mill. In Loom Co. v. Higgins, 105 Ú. S. 580, 591, 26 L. Ed. 1177, an improvement in looms for weaving pile fabrics was held patentable which consisted of such a new combination of known devices as to give to a loom the capacity of weaving 50 yards of carpet a day, when before it could only weave 40. In that case the supreme court said, speaking through Mr. Justice Bradley:

"It is further argued, however, that, supposing the devices to be sufficiently described, they do not show any invention, and that the combination set forth in the fifth claim is a mere aggregation of old devices, already well known, and therefore it is not patentable. This argument would be sound if the combination claimed by Webster was an obvious one for attaining the advantages proposed, one which would occur to any mechanic skilled in the art. But it is plain from the evidence, and from the very fact that it was not sooner adopted and used, that it did not, for years, occur in this light to even the most skillful persons. It may have been under their very eyes, they may almost be said to have stumbled over it, but they certainly failed to see it, to estimate its value, and to bring it into notice. At this point we are constrained to say that we cannot yield our assent to the argument that the combination of the different parts or elements for attaining the object in view was so obvious as to merit no title to invention. Now that it has succeeded, it may seem very plain to any one that he could have done it as well. This is often the case with inventions of the greatest merit. It may be laid down as a general rule, though perhaps not an invariable one, that if a new combination and arrangement of known elements produce a 44 C.C.A.-37

*

new and beneficial result, never attained before, it is evidence of invention. It was certainly a new and useful result to make a loom produce fifty yards a day, when it never before had produced more than forty; and we think that the combination of elements by which this was effected, even if those elements were separately known before, was invention sufficient to form the basis of a patent."

The aim of all these inventors and mechanics who invented and made the various machines for preparing the cane was to make one that would so prepare the cane that the three-roll mill could press out of the cane approximately all of its juice. For many years machines were produced that would only tear the cane longitudinally or strip it or shred it. This lessened to some extent the labor of the mill, and was of some benefit. But it was subject to many objections, as shown by the evidence. The idea of producing a machine that would cut or break, and not tear or shred, seems to have occurred first to Krajewski, or at least he was the first to produce such a device. In view of the fact that the shredded cane in full length, or the split and torn cane, would clog the mill, producing unequal pressure from want of uniformity, it is strange that Krajewski's idea did not occur to the prior inventors. The evident fact that a uniform feed of cane in short lengths would permit the rolls of the mill to run nearer each other, allowing greater and more uniform pressure, would, it seems, have suggested the necessity of a different device or combination. The shredders failed to properly prepare the cane for the mill. The Krajewski machine succeeds. In Re Barbed Wire Patent, 143 U. S. 275, 282, 283, 12 Sup. Ct. 443, 450, 36 L. Ed. 154, Mr. Justice Brown said:

"Under such circumstances courts have not been reluctant to sustain a patent to the man who has taken the final step which has turned a failure into a success. In the law of patents it is the last step that wins. It may be strange that, considering the important results obtained by Kelly in his patent, it did not occur to him to susbtitute a coiled wire in place of the diamond shape prong, but evidently it did not; and to the man to whom it did ought not to be denied the quality of inventor. There are many instances in the reported decisions of this court where a monopoly has been sustained in favor of the last of a series of inventors, all of whom were groping to attain a certain result, which only the last one of the number seemed able to grasp."

Other authorities may be cited tending to sustain our conclusion that the Krajewski patent is valid. Schroeder v. Brammer (C. C.) 98 Fed. 880, 888; Dudley E. Jones Co. v. Munger Improved Cotton Mach. Mfg. Co., 1 C. C. A. 158, 49 Fed. 61; Consolidated Safety-Valve Co. v. Crosby Steam Gauge & Valve Co., 113 U. S. 157, 5 Sup. Ct. 513, 28 L. Ed. 939; Smith v. Vulcanite Co., 93 U. S. 486, 495, 23 L. Ed. 952; Magowan v. Packing Co., 141 U. S. 332-343, 12 Sup. Ct. 71, 35 L. Ed. 781; Potts & Co. v. Creager, 155 U. S. 597, 15 Sup. Ct. 194, 39 L. Ed. 275.

The question of infringement remains to be considered. The machines alleged to infringe the patent were made by the Birmingham Foundry & Machine Company. Moffett, the pattern maker and draftsman of the company, admits on cross-examination that he made the drawings of the machines from advertisements of the Krajewski machines, and from an inspection of two of the Krajewski