This is a suit for the infringement of letters patent No. 577,639, issued to Albert J. Bates, February 23, 1897. The patent has been before the courts of this circuit in former litigation between the same parties. (C. C.) 160 Fed. 108, and 169 Fed. 793, 95 C. C. A. 259. It was there sustained, and defendant’s machine involved in that suit was held to infringe. After that litiga-tioii was closed, defendant devised and put into commercial use four new machines which form the subject of the present suit. The trial court found no infringement and dismissed the bill. Complainant appeals.
The structure of plaintiff’s machine, and the scope of its patent, are described and defined in the opinion in the first suit. The novelty of the invention over the prior art consists in applying a plurality of stay wires in straight line sequence across the strand wires of a fence, and intercoiling the ends of these stay wires with each other and about the strand wires. The object attained had been sought by manufacturers of wire fence for half a century. It had been approached by slow stages. See patents cited in the former litigation. [C. C.] 160 Fed. 108; [C. C.] 160 Fed. 125. The Edenborn patent, No. 558,787, came near the Bates invention. He used substantially the same mechanisms, but did not combine them in the requisite plurality to give his stays a straight line sequence across the strand wires. His stays were staggered, alternating between the different strands. The result was a feeble support. The Bates invention took the step which avoided this defect. It cannot, however, be spoken of with propriety as a pioneer patent.
In the first suit defendant charged that the claims of the Bates patent were so general as to cover the functions of his machine instead of the machine itself and its mechanical equivalents. The claims standing alone give color to that defense. They run:
“In a wire fence maeliine the combination of (1) mechanism for intermittently feeding a plurality of longitudinal strand wires. (2) Mechanism for intermittently feeding a plurality of stay wires simultaneously and transversely of the strand wires. (3) Mechanism for cutting off suitable lengths of the stay wires to span the space between the strand wires. And (4) mechanism for simultaneously coiling the adjacent ends of the stay wires around the strand wires.”
This language is broad enough to cover every possible mechanism which would construct the kind of fence produced by the Bates machine. We upheld the patent by applying the familiar rule which *119requires claims to be read in connection with specifications. The patent was thus confined to the invention, and limited to the machine described in the specifications and its mechanical equivalents. AH the elements are old not only in the general field of industry, but in the form adapted for use in fence making. Bates’ combination is his invention.
The patent is for a combination of four groups of mechanisms, each of which performs a separate part in the production of a wire fence. First, there are two mechanisms which simultaneously feed out a plurality of strand wires, and transversely to these a plurality of stay wires. When this function is performed, these mechanisms stop. Then a cutting mechanism which has been stationary, starts up, severs the several sections of the stay wires, and stops. Next the coiling mechanism which has been stationary starts up, engages the ends of the stay sections, and intercoils them with each other, and about the strand wires. Then the coders stop. This completes a section of the fence. Thereupon the structure is moved forward, and at the same time the feeding mechanisms start up, and repeat their function. The distinctive feature of the Bates machine, both in its structure and its principle of operation, will thus be seen to be a succession of complete stops and starts. This intermittency is expressly claimed in most of the claims of the patent, and must he read into the others from the specifications, in order to sustain such claims. It is in this machine that the Hates invention lies, and not in the product, in determining the question of infringement, attention must he directed to the structure of this machine and its principle of operation.
The distinctive feature of defendant’s machines 3, 4, and 6 is con-tinuousuess of action in every part. The moment the feeding of the wires begins, the entire machine starts, and continues in action so long as the process of manufacture is carried on. The several mechanisms that do the feeding, the cutting, and the coiling all synchronize together so as to continually produce the complex result of a completed fence fabric, filíese machines are fundamentally different from the Bates machine, both in their structure and in their principle of operation, and they also possess in the quality of contiimousness of action a distinct mechanical advantage over the wear and jar and wasted energy arising from the rapid succession of starts and stops of the Bates machine. For these reasons, we think they do not infringe.
[1] Counsel urges that it cannot be declared as a matter of law that a machine which operates continuously cannot be the equivalent of one that operates intermittently, and we are not disposed to question that statement. Much would depend upon the structure of the particular machine. If it was simple, the change from intermittent to continuous action might be produced by the mere substitution of a well-known mechanical equivalent, or a slight mechanical adjustment. But the likelihood of the result being thus produced decreases as the complexity of the machine increases; and, when mechanisms attain the complexity of those involved in the present suit, the difference in the mode of operation is very strong evidence, indeed, that *120there is such a difference in the machines as will avoid the charge of infringement. The two capital criteria by which to determine the question of infringement are structure and mode of operation. Where both of these are substantially the same in two machines, their identity for purposes of the patent law is established; but, when either is absent, the requisite identity to constitute infringement is as a rule wanting. It can hardly be questioned that continuous action constitutes a radical difference from intermittent action in the mode of operation, and also implies a corresponding difference in structure. In the machines here involved this implication is fully sustained by an inspection of the machines themselves. Dryfoos v. Wiese (C. C.) 19 Fed. 315, affirmed 124 U. S. 32, 8 Sup. Ct. 354, 31 L. Ed. 362, tends strongly to show that a machine which is continuous in its operation does not infringe one which is intermittent.
[2] It is true, as counsel says, that we did not sustain the Bates patent in the former suit because of its intermittency of action. That property was old. As the opinion clearly shows, our decision is based on the fact that the Bates machine combined the old elements for feeding, cutting, and coiling in the requisite plurality and simultaneity of action, to produce the stay in straight-line sequence across all the strands, thus producing a new and useful result. Defendant’s machines produce the same result, and employ a like plurality of parts operating simultaneously, but they differ-from plaintiff’s machine in the feature as to which that machine was at one with the prior art, namely, by substituting continuous for intermittent action. Even if it were true (as it is not) that defendant’s machines were identical with plaintiff’s in every feature which distinguished plaintiff’s machine from the prior art,, there would be no inconsistency between our present and former opinions, if defendant’s machines differ from plaintiff’s in the feature of intermittency as to which it was at one with the prior art. Infringement would be avoided if to substitute continuous for intermittent action required a machine new in structure and mode of operation. All would depend on the importance of this change.
Much of the brief for appellant is devoted to pointing out likeness in details between the elements of the patented machine and defendant’s machine. The force of these arguments is greatly weakened, if not wholly destroyed, by the conceded fact that the elements of these machines are old in the art of fence making. All that Bates did was to combine them i'n the requisite plurality and particularly the mechanism for feeding the plurality of stay wires simultaneously and transversely of the strand wires in straight line sequence. The modifications which he made in the old elements in order to combine them in this plurality certainly required, when compared with the prior art, no greater exercise of inventive talent than was used by defendant in making the modifications necessary to substitute continuous for intermittent action. Both used the same old elements and hence it is an easy matter to point out likenesses.
Counsel for appellant attempts to meet the apparently fundamental distinction between its machine and respondent’s which we have pointed out, by three arguments:
*121(1) lie says, first, that defendant’s machines do not continuously feed the stay wires. He bases this argument upon the fact that in the Denning machines the stay wires are not fed immediately against the strand wires, but are fed in at a distance of several inches from them, and there appropriate sections are cut off, and then these sections are pushed across to the strand wires, moving in a line at right angles to the feed of the stay wires. The mechanism which performs this operation pushes the stay sections across, and then returns to the place from which it started, ready to receive the next section. Notwithstanding the feeding mechanism, and this machine which pushes across the stay sections, are in constant motion, counsel says that the intermittency in presenting the stay wires to the strand wires is the same as the intermittency of the .Bates machine. In our judgment this argument forsakes the action of the machine, and looks to the movement of the stay wires as elements of the fence structure. The form of the fence forbids that the stay wires should be continuously applied to the strand wires. They are separated from each other by several inches, according to the size of the mesh desired, and this necessitates a corresponding succession in their application. But the feed of defendant’s machines is continuous both as to the movement of the stay wire and the stay sections, and the mechanism which produces those movements. There is a succession of stay wires, but no intermittency in the machine.
(2) Defendant says that both in the Bates machine and in the defendant's machines, now under consideration, there is a time when the strand wires and the stay wires are in transverse movement with respect to each other, and that this is succeeded by a stage when the strand wires and the stay wires are held firmly together, and are relatively to each other motionless; and he urges that this succession of relative movement and rest, as between the strand wires and the stay wires, constitutes the real intermittency of the Bates invention. Again, we answer that this argument diverts attention from the Bates machine, where his invention lies, to the movement of parts of the fence fabric in the process of manufacture. Because the strand wires in hotli machines are at one time in motion relative to each other, and at another time stationary, in no way impairs the basic fact that the Bates machine in producing this result along with the other results of the process, operates by a succession of stops and starts, whereas the defendant’s machines proceed in a continuous motion,
(3) It is urged that the feed of the stay wires in defendant's machines is in fact not continuous because the knives which sever the sections move in a plane transverse to the movement of the wires; hence, it is urged, there must be a brief instant while the knife is passing across the wire when the wire is stationary, and that the succession of stops and movements thus produced creates in the defendant's machines the intermittency of action which is found in the Bates machine. This contention is made notwithstanding the manifest fact that the mechanism which feeds the wire forward is in continuous action. The argument seems to us purely fanciful. It substitutes a metaphysical intermittency for a mechanical one.
*122We also think that in defendant’s machine No. 6, the “automatic hand” by which the stay sections are transferred from the cutting machine to the strand wires, is not the mechanical equivalent of the tubes which guide the stay wires against the strand wires in the Bates machine.
Defendant’s machine No. 5, referred to in the evidence as the “Hopper Machine,” has the same intermittency of action as the Bates machine. It is distinguished from that machine by wholly omitting the cutting mechanism. Stay sections are cut in the various lengths needed for different kinds of fence, in a stock machine, built expressly for the purpose, and which was sold and used for many years before the date of the Bates patent. These sections are deposited in quantity in separate hoppers, one hopper for each space between the strand wires, and pass by gravity down a chute into notches in the upper edges of horizontal bars which carry them sidewise against the strand wires into position for coiling. The charge of infringement is, of course, avoided, if the cutting mechanism of the Bates combination is wholly omitted, and not supplied by a mechanical equivalent. Counsel for appellant urges that the device by which the stay wires are separated from the mass in the hopper, and transferred to the strand wires, is really the mechanical equivalent of the Bates cutting mechanism, and ingeniously suggests that the “cutting out” of the stay wires from the quantity in the hopper is the mechanical equivalent of the cutting off of the sections of the stay wire in the Bates machine. Without in any way questioning the good faith of counsel for appellant, this seems to us a far-fetched argument. No one but a patent expert in extremis would ever think of applying the term “cut” to the operation by which the individual stay wires are separated from the mass in the hopper. To find such a use of the term reference is made to the figurative language of the cowboy who speaks of “cutting out” a steer from the herd; and it is urged that this is the same cutting as the sevening of the sections of the stay wires in the Bates machine. When words are thus paltered with in a double sense, there is an end of all profitable reasoning. We use the same word, but we mean different things. There is hardly a common word in our English speech which does not range through a vast variety of meanings. But in the exact science of mechanics, when we use such a term in describing the operation of machines, we select one of its meanings and adhere to it. Any other practice can lead only to confusion. In the Bates machine there are knives and elaborate mechanism for their operation in cutting off the stay sections from the stay wires. This mechanism is clearly claimed as an element of the Bates combination. It thus becomes a part of his invention, and any machine which wholly omits it escapes the charge of infringement. It seems to us too plain for argument that defendant’s hopper machine does this.
Appellant complains because after the decision in the former case respondent deliberately set about devising, with the aid of patent experts, a machine which would produce the Bates fence, without infringing the Bates patent. It is sufficient to answer that complaint *123in the language of Mr. Justice Grier in Burr v. Duryee, 1 Wall. 531, 574, 17 L. Ed. 650: “Every man has a right to evade a patent, provided he does not invade the rights of the patentee.”
The decree is affirmed.