DENISON, Circuit Judge.
[1] This is the usual patent infringement suit, brought by appellant, as plaintiff in the court below, and based upon patent No. 888,984, issued May 26, 1908, to Charles Fell for “apparatus for rifling pipes or tubes.” Pipes used for the transmission of liquids had always been constructed with as smooth an inner surface as possible, in order to minimize the friction. It developed that in the long-distance carrying of the heavy viscous oils of California, the adhesion of the oil to the inner surface of the pipe was a serious obstacle. In 1904 Isaacs and Speed made an ingenious discovery. They found that a film of water interposed between the pipe and the oil would remove the trouble, that if a little water was put in with the [509]*509oil and the whole liquid given a rotary motion, the resulting centrifugal force would throw the water out against the pipe and create the desired water film, and that sufficient rotary motion could be induced by rifling the pipe, so that when the contents were pumped through, giving primarily a longitudinal motion, there would be induced a secondary spiral motion. They utilized their discovery by a method patent, not now involved, and by an apparatus or device for rifling pipe. This was designed for use in the field, after the pipe was manufactured in the ordinary way, and it consisted of a rigid framework encircling the pipe and carrying four rollers or small wheels which projected inwardly radially to the pipe, and the innermost points of which were in a circle considerably smaller than the outer periphery of the pipe. It necessarily followed that as the pipe was passed through, four corresponding indentations would be made, resulting in ribs upon the inside of the pipe. These indenting wheels were so hung in the surrounding frame that their axes were not parallel to the corresponding diameter of the pipe, but were at an angle of, say, 20 degrees therefrom. It followed that the pipe, as it passed through lengthwise, would slowly rotate, and that the resulting ribs, grooves, or depressions on the inside of the pipe would be helical instead of longitudinal. This invention received distinct commercial approval, but the difficulty and expense of handling pipe in this way were considerable, and application was made to plaintiff, a large manufacturer, to produce and sell pipe in that form at its factory. The natural query was whether pipe could be put into that shape by or in connection with the original drawing and rolling process, and while it was hot and comparatively plastic; and, if so, how that could be done. These questions were submitted to Fell, plaintiff’s superintendent, and the patent in suit is his answer.
Fell built his invention upon a device very common in the rolling mill art, a pair of skewed rolls. These consist of an upper and a lower horizontal shaft, each carrying a roll with such concavity in its surface that the pipe may travel between the two, while moving longitudinally upon a line at right angles to a vertical plane halfway between the vertical planes of the two roll shafts; in other words, the upper and lower roll shafts, instead of being at right angles to the travel of the pipe, make complementary angles thereto of, e. g., 70° and 110°, the upper roll shaft having its left-hand end moved 20° forward, and the lower one having its corresponding end moved 20° backward from the right-angled position. Obviously, if the concavity in each roll was upon the arc of a circle, this nonparallelism of the upper and lower roll shafts would so distort the opening that a circular pipe could not pass through, and in order to make this opening or “pass” circular, when projected upon a plane at right angles to the pipe axis, and so suitable for the passage of circular pipe, it is necessary to make each concavity partially elliptical. Such “skewed rolls” operate at the same time to forward and to rotate the pipe, and were in common use, particularly for straightening purposes. It occurred to Fell that by putting circumferential beads or ribs upon the concave faces of these rolls he could get the desired result. Experiment demonstrated that [510]*510this was so. The difference between what Isaacs and Speed had done and what Fell did is made clear by the following sketches:
(Fig. 1, Isaacs & Speed.) (Fig. 3, Fell.)
Fell’s patent has 14 claims ; the bill alleged infringement of all. In the court below plaintiff relied upon nearly all, but in this court it practically confines itself to the first three; and we shall find it necessary to refer to the others only for tire salce of applying the differentiation test, in order to determine the meaning of the first three. They are as follows:
“1. In apparatus for rifling pipes or tubes, a pair of skewed rolls formed witA circumferential grooves or passes for the pipes or tubes having circum-ferentially extending rifling beads or ribs, substantially as described.
“2. In apparatus for rifling pipes or tubes, a pair of skewed rolls formed with circumferential grooves or passes for the pipes or tubes having circum-ferentially extending rifling beads or ribs, and means for positively rotating both of the rolls, substantially as described.
“3. In apparatus for rifling pipes or tubes, a pair of rolls having each a groove or pass for the pipes or tubes with circumferentially extending rifling beads or ribs, the two rolls having their axes crossing each other, and means for positively driving the said rolls, substantially as described.”
We see no substantial distinction between these claims. The element added by the second claim, “means for positively rotating both of the rolls,” is the statement of an element which, in some form, must be present to make the device independently operative; and the third claim merely uses different terms of description for the same rolls, so we may confine ourselves to the first claim.
The novelty of the broad result which Fell accomplished is not questioned; no one had ever before rifled pipe with a pair of skewed rolls carrying circumferential beads; upon this branch of the case, the only question is whether his solution required invention as distinguished from expert skill. There were in existence two classes of [511]*511machines, more or less analogous. One was for corrugating hollow pipe; the other was for twisting' solid bars. The first class is typified by Fife, No. 426,439, April 29, 1890. This consisted of two indenting rollers, upper and lower, .upon skewed axes so that, as the tube was fed between them, they produced a helical indentation. The axes of the rolls were almost parallel with the axis of the pipe, whereby it resulted that the helix had a very low pitch, and the product was what would naturally be called a corrugated tube; but if there is any distinction between corrugating and rifling due to the pitch of the groove, Isaacs and Speed had bridged the gap by applying the same idea to rifling, except that they used four indenting rolls, and thus provided a pass within the circle struck through these four points, instead of through merely the two opposite contact points of Fife. The other class of devices was used for making twist drills, and is best typified by Denk, No. 761,880, June 7, 1904. He uses skewed rolls, each having one circumferential, large, high rib. The round bar to be twisted was first provided with two oppositely disposed large longitudinal grooves of about the same cross-section shape as the cross section of the circumferential rib on the rolls, and these grooves were made very deep, so that the two left only a comparatively thin web of the bar between them. -As the rod is advanced through the skewed rolls, with the upper and lower ribs engaging the respective grooves, and as the rear end of the rod is gripped fast against rotation, it necessarily twists on this central web, and the result is a twist drill.
We are satisfied that Fell made a distinct and valuable advance, and one beyond the limits of mechanical skill, and that his first claim is valid. No one of the pipe corrugating devices had an indenting bead upon a roll, which roll conveyed a double motion to the pipe. In ail these corrugating devices the parts which corresponded to, and were perhaps equivalent to, the indenting bead were carried upon a stationary pipe-inclosing framework. They were adaptations of the screw-cutting lathe. The idea that this indenting element could be carried circumferentially upon the surface of a roll is not disclosed by any of the corrugating patents, had not occurred to these inventors, and to us seems not obvious. If it did occur to one, it seems likely that the thought would be rejected because all these corrugating and indenting rollers pressed inwardly on a pressure line radial to the pipe; and, while the same thing would be true of one centrally located bead in the upper and lower rollers, this would produce only two grooves— not enough for efficient rifling — and as soon as the beads were put anywhere except in the center of the rolls the indenting pressure became tangential and not radial. That this apparently abnormal action would produce a satisfactory result in pipe of this character is due only to the fact that well-shaped grooves are not necessary, and that a slight indentation is sufficient.
The twist drill machine is, upon the whole, no closer. It has skewed feed rolls, carrying one circumferential bead, but this also, being central on the roll, is radial to the pipe; and — more important than that~it was not designed to, and did not operate to, make an indented groove in pipes or in anything else. The groove was made by other [512]*512means, and, having been so provided, was then used as a means of twisting the weakened central web. It was a small device, used for making small things in a tool factory; and to take this device and enlarge it to be suitable for rolling-mill purposes, and to substitute tangential, small, slightly indenting beads for the single, radial, comparatively large gripping bead, producing eight-inch, rifled pipe, instead of a small twist drill, impresses us as clear and strongly marked invention, which, while not primary in every sense of the word, yet is so distinct and so meritorious as to entitle Fell to a fairly broad and liberal construction of his patent. The advance was great enough and the invention clear enough so that we do not need to place reliance upon the great practical utility of the device, or upon the fact that defendants, after experimenting with several other forms, were apparently forced to adopt the same general plan of manipulation; though both of these features are present in this case.
The question of infringement requires an interpretation of the terms of the claim, which may or may not cover defendants’ structure, according as the interpretation is liberal or narrow. The defendants, having received an order for a large quantity of this rifled pipe from the same customer who had been buying from plaintiff, and on the condition that they would sell at a less price, first tried a modification of the Isaacs and Speed device, and then made some other changes, settling down upon and continuing to use a structure which, instead of the conventional roll, is built up from discs of different sizes rigidly attached to a shaft. This structure is shown in comparison with plaintiff’s by the following sketches made by plaintiff’s expert, and which, though not áccurate, are not essentially unfair (only the upper roll of the pair being shown):
It is obvious that if there was indenting pressure exercised vertically upon the upper and lower parts of .a hot and ductile pipe, the circular cross-section would be distorted and the sides would be bulged out horizontally, unless resistance against such distortion was provided at this [513]*513point. The plaintiff’s rolls, extending as they do to the same central, horizontal plane, provide this resistance by the extreme parts of the semielliptical concavity; and when these rolls are projected upon an intermediate vertical plane, as has been done in the above sketch, these extreme parts appear to be and may not improperly be thought of as flanges on the rolls, and they provide a rolling, frictional contact with the pipe, resisting the tendency to side expansion along and near its central, horizontal plane; and, while we recognize that the term lacks perfect accuracy, we shall, for convenience, speak of this part of the roll as the “flange.” The defendants’ rolling structure does not have this part. The largest disc on the upper roll is the one which makes the lowermost indentation in the upper half of the pipe, and as it does not extend below this point, it cannot provide resistance for the tendency to horizontal expansion; but, as this resistance is vitally necessary, the defendants provide it by guide-plates (4 and 5) fixed to the framework and accomplishing the same ultimate function by a sliding friction instead oí a rolling friction. The defendants say of their device: First, that it is not a roll at all; second, that the roll flanges are an essential part of plaintiff’s structure, and that these the defendants have omitted. It is also contended, rather as part of the first point just named, that the patent contemplates and is confined to a roll which is in contact with the walls of the pipe throughout the greater part of the roll surface between the beads, so as to have throughout its surface the forming function. The considerations which require that construction of the claim which will support these contentions are said to be found on the face of the patent, in the Patent Office proceedings and in the prior art.
We see nowhere any reason for doubting that the defendants’ built-up structure comes within the broad and general definition of “a pair of skewed rolls with circumferential 'grooves having circumferential beads.” As the beads are the chiefly effective parts in the rifling operation, it can make no difference whether they are built up by discs, leaving the intervening parts open, or whether all parts were first cast solid and then the intervening parts cut away. The points which grip and indent and forward and rotate the pipe are just the same whether the body of the device is in solid or in skeleton form. The development of defendants’ structure demonstrates that it is “a pair of skewed rolls.” They took their ordinary sizing rolls with semicircular groove, cast ribs in the grooves, and skewed the axes. This device would not work, because the pass was not circular. The extremities of the groove- — what we have called the roll flanges — were, by the skewing, projected into the pass on the sides. These flanges being thus in the way, the defendants cut them off and substituted the guides, but by this change the rolls did not lose their general character; they continued to be skewed rolls, with a circumferential groove having circumferential beads. The defendants then cut away some of the face of the roll between the ribs, making the ribs higher, whereby it became apparent that they were, as they always had been, the edges of discs imbedded in the roll; but the structure did not, because its surface contour was thus changed, cease to be a roll carrying beads. It also [514]*514continued to have the circumferential groove. Obviously there must be grooves, or else the pipe could not pass through. A groove is an opening with a bottom and with sides, and these parts are sufficiently marked by the three discs. If these were close together, or if two more graduated discs were inserted between the central one and each side one, it would be perfectly apparent that there was a circumferential groove; and we think it no less true because the surface is not completely marked. We are not sure that defendants intend to deny that their device responds to a definition of the terms of the claim, so broadly considered.
Are the side-supporting flanges of plaintiff’s rolls to be considered as performing a function essential to the device and as a feature which defendant must use, in form or by equivalent, before there is infringement? We think they are. It is true that the first claim does not call for these flanges in so many words, and that some of the other claims impliedly do so by calling for a “semielliptical groove or pass,” since the groove will not be completely semielliptical unless it is carried down to the central plane; but this term, “a pair of rolls,” used in this connection, necessarily means rolls that will roll pipe; and such a construction of them or their associated parts that the pipe will retain its general form is necessarily implied. However,, the very fact that this limitation as to the form of the roll is not expressed in the claim, but that any implication on that subject arises from the necessity that the device shall operate, indicates that no narrow rule of equivalency can be applied, and that when something else is substituted which was already a well-known means of performing the same function, and when its construction did not involve any invention, the thing so substituted is a mechanical equivalent, and infringement is not thereby escaped. We have already pointed out that these side flanges or extreme ends of the semielliptical groove have no function, except to resist the outward bulge of the pipe at this point. In a fair sense they are rolling guides; and the stationary guides used by defendants (shown in cross-section in the above sketch) have precisely the same function and perform it in the same way, except that the friction is sliding instead of rolling. Further, the Root patent, No. 180,-643, of August 1, 1876, had skewed rolls, in each of which the groove or pass was not semielliptical, but more shallow, and was*provided with central stationary guides or “keepers,” all arranged (as said in defendants’ brief) “just exactly as in defendants’ machine.” The defendants, therefore, have substituted for the roll flanges not only an equivalent, but a known equivalent; and this confirms the conclusion that infringement is not thereby avoided.
We find nothing upon the face of the patent requiring the claim to be limited to rolls which had these flanges attached to and integral with the rolls. It is true that neither by drawing nor specification does Fell suggest the performance of this function by stationary parts of the frame, but this is not necessary. In the absence of something clearly showing that the patentee did intend to have his grant confined to a specific form, a broad and generic claim may rightfully stand on a mere specific disclosure; and the invalidity of such a claim (if [515]*515it is invalid) will result, not from the applicant’s failure to use more sweeping language in his specification, but from the state of the art limiting the actual invention. The claims are part of the description required by statute, and in them, and not in that part of the description which is now commonly called “specification,” is the proper place in which to define the breadth of the invention, as was most accurately apprehended by Fell’s solicitor when he (though quite unnecessarily) said that various changes might be made “without departing from my invention as defined by the appended claims.”
[2] Nor can we find the supposed claim limitation to these specific roll-flanges in the fact that Fell’s specification and drawings showed rolls having flanges, and that he claimed the rolls “substantially as described.” The stress which defendants’ counsel in argument have put upon this quoted phrase, and the constant recurrence of the question in cases coming before us, make it advisable to consider and determine what place this phrase should have in the rules of claim construction. We think a review of the controlling principles, of the authoritative cases, and of the established Patent Office practice, in the light of which its grants must be considered, demonstrate that, at least in all recent patents, and save in exceptional instances, the presence or absence of this phrase is of no interpretative importance.
Upon principle the matter seems clear. The statute requires that the specification should describe the invention, and that the patentee should specifically claim what is new. He cannot claim that which he has not substantially described; and it seems necessarily to follow that the idea — not “exactly as described” or “precisely as described,” but “substantially as described,” that is to say, “not varying in any vital way from the description” or “cither in the form shown or in its substantial equivalent” — is necessarily inherent in every claim. So, in Mitchell v. Tilghman, 86 U. S. (19 Wall.) 287, 391 (22 L. Ed. 125), referring to these words, it was said, “Words of such import, if not expressed in the claim, must be implied”; and again, in Hobbs v. Beach, 180 U. S. 383, 400, 21 Sup. Ct. 409, 416 (45 L. Ed. 586), speaking of the rule that these words refer us back to the specifications, that court says, “This rule, however, is equally applicable whether these words be used or not.” To say that these words must be implied, if not expressed, and to say that if expressed they import an otherwise nonexistent limitation, is to propound a dialectic paradox. We cannot lightly assume such a self-destructive contradiction in the law, and its seeming existence calls for a careful analysis of the apparently pertinent and authoritative cases.
in Seymour v. Osborne, 78 U. S. (11 Wall.) 516, 547 (20 L. Ed. 33), the question was whether the claims were void as being for an effect or a function, as, for example:
“Discharging the cut grain from a quadrant-shaped platform on which It falls as it is cut, by means of an automatic sweep-rake sweeping over the same, substantially as described.”
The court unquestionably placed reliance on this phrase, “substantially as described,” in reaching die conclusion that the claims were not functional, but it is not apparent that the court would have reached [516]*516any different decision if the words had not been there; and the language of the opinion on this point is written by Mr. Justice Clifford, who, only three years later, in Mitchell v. Tilghman, supra, 19 Wall. 287, 391 (22 L. Ed. 125) found it necessary to use the language which we have already quoted:
“Words of such import, if not expressed in the claim, must be implied, else the patent in many cases would be invalid,” etc.
It seems quite clear that Mr. Justice Clifford could not have regarded his language in Seymour v: Osborne as laying down any rule of law to the effect that the actual presence of words, always constructively present, could have controlling importance. It is further to be noted that Seymour v. Osborne was decided in 1870, construing patents issued before 1865, and both issue and construction were before the Patent Office rule on the subject became finally settled as hereafter pointed out. In the Corn Planter Patent, 90 U. S. (23 Wall.) 181, 23 L. Ed. 161, the claim involved called for a combination of several parts “substantially as and for the purpose set forth.” From the discussion and disposition of the matter on pages 217, 218, and 219, it is apparent that the court construed the claim by reading into it what was necessary to make it an operative device of the general class and for the general purposes described; and it is difficult to see that any different course could have been taken if the concluding phrase had not been present in the claim. The remark that this phrase “throws us back to the specification for a qualification of the claim” seems rather casual, because clearly we will be thrown back to the specification for that purpose without any propelling aid from this clause. In Westinghouse v. Boyden, 170 U. S. 537, 18 Sup. Ct. 707, 42 L. Ed. 1136, the opinion was by Mr. Justice Brown, and his remark on this subject, on page 558, referred only to Seymour v. Osborne and to the Corn Planter Patent, omitting reference to Mitchell v. Tilghman. The question as to the proper construction of the claim had already been decided, and the presence of this phrase is only referred to as one of the “other facts which have a strong bearing in the same connection.” Just as Mr. Justice Clifford, when the matter was the second time before him, did not seem to be controlled by his comment the first time, so Mr. Justice Brown, going into the subject more thoroughly in Hobbs v. Beach, was unwilling to say that these words had any meaning.
In Hobbs v. Beach, supra, 180 U. S., at page 399, 21 Sup. Ct. at page 416 (45 L. Ed. 586), the court was confronted more squarely than ever before with the question of what force should be given to these words. This question was one of those upon which the two courts below had differed, and if they were held to import full limitation to the form of the specifications and drawings, there was no infringement. It seems a fair conclusion from the opinion that Justice Brown found himself embarrassed by what had been said in the three last above cases, on the one side, and on the other by Mitchell v. Tilghman and the inevitable logic of the situation, which compelled him. to say that the rule of going back to the specification for a qualification of the claim and for the elements of the combination “is equally applicable whether these words be used or not.” He, therefore, [517]*517while unwilling to say that they are “absolutely meaningless,” did conclude that “it is difficult to say exactly what effect should be given to these words,” and that “the authorities really throw but little light upon their proper interpretation,” and finally concluded that “without determining what particular meaning, if any, should be given to these words,” they did not limit the patentee to the exact mechanism described, but that he was still entitled to the benefit of the doctrine of equivalents. In spite of the uncertainty shown by the language of the opinion, and on studying the specification and claim of the Beach patent and the scope that must be given in order to sustain the charge of infringement by defendant’s device, and also the opinions of the courts below, it is difficult to see why Hobbs v. Beach is not a decision that a claim containing these words is entitled to the same breadth of benefit of the rule of equivalents as if the words were not there.1
Since Hobbs v. Beach, the Supreme Court has mentioned the question only once, so far as we find. In Singer v. Cramer, 192 U. S. 265, 284, 285, 24 Sup. Ct. 291, 48 L. Ed. 437, Mr. Justice White clearly indicates his supposition that the presence of this phrase in a claim did have a limiting effect; but what is said on this subject is without reference to Hobbs v. Beach, and is said after a discussion of the Patent Office proceedings and the limitations accepted therein had led him to the conclusion that the limited construction of the claim must be given. We do not think this case should be taken as a decision that the presence of the words necessarily imports a limitation. It may be usefully noted that in most, if not all, of those Supreme Court cases which give claims a broad construction and do not limit them to the form shown in the specification these words were present. See (e. g.) Winans v. Denmead, 56 U. S. (15 How.) 330, 14 L. Ed. 717; Clough v. Barker, 106 U. S. 166, 1 Sup. Ct. 188, 27 L. Ed. 134; Morley Co. v. Lancaster. 129 U. S. 263, 9 Sup. Ct. 299, 32 L. Ed. 715; Hoyt v. Horne, 145 U. S. 302, 12 Sup. Ct. 922, 36 L. Ed. 713.
In the decisions of this court, statements are to be found looking in both directions. In Stilwell Co. v. Eufala Co., 117 Fed. 410, 414, 54 C. C. A. 584, Judge Day was illustrating the rule that a claim should be construed by the specifications, and cited some of the above-named Supreme Court cases referring to “substantially as described.” So far as ‘the case indicates, he did not, in reaching the result, place any dependence upon that phrase; on the contrary, the defendant’s claim of uoninfriugement was overruled and the claims were given a construction sufficiently broad to cover defendant’s apparatus. In Sanders v. Hancock, 128 Fed. 434, 436, 63 C. C. A. 166, it was said of this phrase that it “carried into the claims the description of the specification.” Judge Severens again was only illustrating the propriety of a [518]*518.decision that when an element was named in a claim without description, except the name, resort should be had to the specification to see what the element was. That no decision was intended as to the definite effect of “substantially as described” is evident from the fact that of the four cases cited by Judge Severens one (Stilwell Co. v. Eufala Co.) treated the point only incidentally as above noticed, two (Soehner v. Favorite Co., 84 Fed. 182, 28 C. C. A. 317, and Canda v. Michigan Co., 124 Fed. 486, 61 C. C. A. 194) contain no reference ’to the point, and the remaining one (Lamb Co. v. Lamb Co., 120 Fed. 267, 269, 56 C. C. A. 547) says it is not necessary to enter into this “beclouded question.” When, however, in General Elec. Co. v. International Co., 126 Fed. 755, 758, 61 C. C. A. 329, the limiting force of this phrase was somewhat more directly involved, and the then recent case of Hobbs v. Beach was apparently first considered by this court, it was held that these words did not import into the claim all the details of the specification, but only meant that if the parts named in the claim operated substantially as described, “that is all that is necessary.” So, in Houser v. Starr, 203 Fed. 264, 269, 121 C. C. A. 462, 467, we said (though unnecessarily to the decision):
“Whatever parts are named in the claim are of necessity intended to be named with reference to the specification and drawings. The reference cannot be made narrower by saying, ‘as described,’ nor broader by saying, ‘substantially.’ The words ‘substantially as described’ do not create this necessity for construction by the entire patent; they are only a formula of recognition of the rule.” 2
[3] If, however, there was a doubt about the force to be given in the present case to this phrase as one of limitation, that doubt must be set at rest by consideration of the rules of construction in force in the Patent Office when this patent issued. The meaning put upon possibly ambiguous language of the grant, both by the Patent Office and by the patentee, becomes a binding rule of construction of that grant. This is both by the general law of contracts and by the rules of estoppel. It seemingly must be equally true of a fixed and universal construction, applied all the time to all patents, and presumably known to all pat-entees, as of a special meaning fixed by the special action. Just as a statute adopted from another state after interpretation by its highest court must, in the adopting state, receive the same construction, so when the Patent Office has a settled rule of construction under which a certain phrase in a patent grant has a fixed meaning, acquiesced in by the patentee, the courts cannot afterwards rightfully say that the phrase means something more or something less. In 1902 the practice of the Patent Office on this subject was considered, and, so far as there had been any uncertainty, was finally settled by Commissioner Allen in his decision in Ex parte Shepler, 102 O. G. 468. The substance of this ruling is that thereafter these words in a claim shall be considered as “inconsequential,” and that a claim with them in is no narrower [519]*519iban a claim witlicnit them. The importance of the opinion merits its reproduction in full in the margin.3 This general view of the matter was not new in 1902. As early as 1869, Commissioner Fisher (In re Rubens & Co., Com. Dec. of 1869, p. 107) said:
“The practice of allowing claims for inventions which can only be distinguished from those previously invented or patented by the construction to be given to such words as ‘substantially as described.’ ‘as herein set forth,’ etc., cannot be too strongly condemned. The claim should state all the elements of the combination intended to be patented, and if the parts are the same in name and nuinber as in some prior machine, and the improvement consists in some modification of one or more of those parts, the claim should distinctly state that modification. It should not be concealed in ambiguous phrases. The words ‘substantially as described,’ and the like, have no fixed legal meaning. They may serve to expand or contract the claim. In general, they are employed for the latter purpose, so that a claim may appear to be broad upon its face which, in truth, by virtue of this mysterious phraseology, is exceedingly narrow. Sometimes they are added by inventors as a matter of course, by way of rounding off the claim, and have no particular meaning, and again, they are intended to signify that the patentee claims that which he describes and everything substantially like it. They are, at the best, ambiguous and dangerous phrases, and. however else they may be used, they must not be employed in the granting of letters patent to make a claim good that, without them, would be bad and unpatentable.
[520]*520“In conformity with the current practice of the office * * * the concluding words of the second and third claims should he ignored in considering the novelty of what is therein claimed.”
In 1872, Commissioner Thatcher said (In re Sperry, Com. Dec. 1872, p. 221):
“It has become the settled practice of the office to reject the clause as a limitation unless so placed as to obtain special significance.”
Again, Commissioner Deggett said of these words (In ré Collins Co., Com. Dec. 1872, p. 251):
“There is no objection whatever to their use where they make sense; but they have no legal effect either to enlarge or limit a claim properly drawn; and, so far as the grant of the patent is concerned, they should have no influence one way or the other. The claim should be sufficient in its terms either with or without them, because, either with or without them, its meaning and effect are to be determined in the courts by the light of the specification.”
The only exception 'to this rule — that the words are “inconsequential” — which seems to inhere in the situation is that if a patent was issued which contained two claims not distinguishable from each other save by the presence in one claim of this phrase, it might be the duty of the court, in giving effect to every part of the patent, to find in them [521]*521some distinguishing force; but if such instances exist at all of late years, it can only be by inadvertence in the Patent Office.4
The foregoing considerations seem to us fully to justify the proposition with which this discussion is opened — that at least in all recent patents, and save in possible exceptional instances, the presence or absence of, this phrase is of no interpretative importance.5
Limitation to Fell’s specific form of roll and differentiation from defendants’ three-disc structure are also thought to be found in the drawings, showing that practically all of the roll surface between the beads was in contact with or “hugged” the pipe surface, and in the specification referring to this portion of the roll surface as a “forming wall.” It appears to be a fact beyond dispute that this contact could not occur to anything like the extent shown in the drawings, unless the pipe material was too soft to be capable of rolling without a mandrel, and that in the practical operation of plaintiff’s device there is contact, if at all, only at one point midway between the beads; while the successful use of defendants’ device demonstrates that even so much contact is unnecessary. It is clear enough that the solicitor and the draftsman supposed that the rolls would operate in the specific way shown in this figure 3 of the drawing, and Fell is probably chargeable with the same supposition; but this comes to nothing more than an illustration and description of the supposed best form of device, with specific claims limited to that form, but accompanied by broad claims lacking such limitation. Claims 10 and 12, for example, cannot be effectively distinguished from the first three claims, except by the added limitation:
“Tlie walls of said passes being arranged to support the pipe or tube adjacent to {lie points of contact of the bending ribs or beads therewith and between such ribs or beads.”
[4] Where a patent contains both a broad and a narrow claim and suit is brought on the broad claim, we cannot construe into it a limitation not therein expressed, but which is expressed in the narrower claim and by which alone one is distinguished from the other. To do so would be making over the contract between the public and the patentee. Bresnahan v. Tripp Co. (C. C. A. 1) 102 Fed. 899, 900, 43 C. C. A. 48; O’Rourke Co. v. McMullen (C. C. A. 2) 160 Fed. 933, 939, 940, 88 C. C. A. 115; National Co. v. American Co. (C. C. A. 3) 53 Fed. 367, 370. 3 C. C. A. 559; Lamson v. Hillman (C. C. A. 7) 123 F. 416, 419, 59 C. C. A. 510; Mast, Foos & Co. v. Dempster Co. (C. C. A. 8) 82 Fed. 327, 333, 27 C. C. A. 191; Duncan v. Cincinnati Co. (C. C. A. 6) 171 Fed. 656, 663, 96 C. C. A. 400; Sheffield Co. v. D’Arcy (C. C. A. 6) [522]*522194 Fed. 686, 691, 116 C. C. A. 322; Scaife v. Falls City Co. (C. C. A. 6) 209 Fed. 210, 214.
[5] The history .of the application in the Patent Office furnishes no support for the conclusion that the roll flanges in the form shown are inherent in the first three claims. It is sufficient to say of these proceedings that, in so far as there is indicátion of any attempt to impose such a limited construction on these claims, the record shows that, as is so often the case, the examiner and not the applicant yielded. It is not necessarily of any importance that, when the examiner rejects a claim on reference to an earlier patent, the applicant thereupon amends the claim. Such a record is of importance and raises an estoppel against the patentee only when it additionally .appears that the effect of the amendment was to narrow the claim so that it no longer depended in material part upon any theory of patentable novelty, upon which same theory the patentee afterwards attempts to support the validity or scope of his amended claim. Amendments which, when examined, are found to be mere matters of form, or which do not have a limiting effect in the particular which is involved in the later judicial inquiry, do not estop the patentee from receiving such a construction of his patent as will protect his actual invention, and so far as permitted by the claim language finally adopted. Scaife v. Falls City Co. (C. C. A. 6) 209 Fed. 210, and cases cited in notes 8 and 9, p. 217.
The first three claims were originalty filed in the same form as when issued, except that there were added, by amendment, and by way of further description, the words “skewed” as to the rolls, “circumferential” as to the grooves or passes, “rifling” as to the beads or ribs, “for the pipes or tubes”' after “passes,” and “circumferentially” before “extending beads or ribs.” Every change except the last was obviously immaterial, because the added words were already inherent in the claims; and the limitation to “circumferentially” extending beads was inserted to distinguish from a reference having beads, or something similar, longitudinal on the rolls. With this limitation defendants are not concerned, because they use “circumferentially” extending beads or their equivalent. It is true that these three claims were once rejected on Denk, above described; but Fell not only declined to make any change whatever in them because of this rejection, but insisted that the examiner had “misunderstood the reference,” and that “applicant’s method of rifling pipe is believed to be broadly new. It is believed he is entitled to the broad claims presented.” He added in this letter:
“Furthermore * * * the reference does not show the walls of the passes arranged to support the article.”
This sentence has no possible effect to limit these three claims to passes which have walls so arranged, -not only because it is added as mere makeweight to the sufficient reasons already given, but also because it had special and appropriate application to the further claims which already contained this limitation, and which also had been rejected upon the Denk reference. When later these three claims and others were further rejected on reference to Fife’s patent for corrugating tubes, above described, Fell amended these by inserting the pro[523]*523vision that the grooves or passes should be circumferential. We have said above that this limitation already existed, and so Fell wrote:
“In fact, while claims 1, 2, and 3 have teen amended for the purpose of making them more clear, the pertinency of the reference to any of the claims is not clear.”
The remainder of his amendments made at this time, and in answer to this citation, refer to other claims, and cannot affect the first three. The claims were again rejected on reference to a patent which had rolls parallel with the pipe axis, and in which the spiral corrugations in the pipe were formed by the spiral corrugations in the mandrel. In answer to this reference Fell points out this distinction, and (again unnecessarily) inserts “skewed” with reference to his rolls. At the same time he made other comments, which might or might not have a limiting effect on other claims, but which could not affect these three.
Our review of these Patent Office proceedings satisfies us that they furnish no support for the theory of such limitation as will avoid infringement of claims 1, 2, and 3; their whole tendency is in the contrary direction.
The prior art has nothing closer than these earlier patents already discussed.
The decree below must be reversed, with costs in favor of appellant, and the usual decree for injunction and accounting entered as to claims 1, 2, and 3.