Lead Opinion

BUFFINGTON, Circuit Judge.

In this case the plaintiffs, the Pless-Bright Manufacturing Company and others, owners of patent *724• 822,723, granted June "5, 1906, to Robert Conrad for a ball bearing, charge Fichtel & Sachs with infringing the eighth and ninth claims thereof. On final hearing the court below, in an opinion reported at 209 Fed. 867, dismissed the bill on the ground of noninfringemant. Thereupon the plaintiffs took this appeal. By reference to such opinion and to that of Judge Holland in 177 Fed. 435, where the same patent was in issue im another case in this circuit, its subject-matter will be seen. It will be noted that Conrad’s patent is for a finished ball bearing adapted for use, and not for a method of assembling. Thus, iii his patent he says:

“I do not claim in the present application the described method of assembling the parts of my improved ball bearing, this method being claimed in a divisional application filed May 18, 1906.”

And, as showing that his patent was for a finished ball bearing, adapted for use, he says:

“The principal advantage of the new bearing lies in the continuity of the sides of the groove, which insures the regular runmng of the balls, and consequently great durability of the bearing,”

Such being the avowed advantages claimed by the patentee, let us turn to the art and inquire whether Conrad’s bearing was new, whether that novelty did lie in “the continuity of the sides of the groove,” and whether it was this feature of the groove continuity that made the bearing durable. If those features of novelty and utility be established, and the advance thus made involved patentability, our next inquiry will he whether the patentee duly claimed and the defendant has appropriated this groove continuity. Turning first to the question of novelty, it is apparent, of course, that prior to Conrad’s patent the mechanical construction known as a ball bearing, wherein an axle or ring was provided- with an exterior groove and an enveloping ring with an interior groove, which two grooves overhung loose bearing balls, was old. So, also, we must assume the possible eccentric relation of these two- rings was also theoretically well known. The-then common mode of ball bearing construction, however, was to provide one of the rings with a removable section through which the balls were placed in the groove. After this such opening was closed to prevent ball escape. This device is shown in the patent No. 669,124, to Riebe, of 1901. The other method was to introduce the balls through a sidewise slot or channel in one of the rings cut as deep as the groove path. This device is shown in patent No. 796,871, of Sachs, of 1905. While these methods of construction made a reasonably satisfactory hearing under light strain, they proved unsatisfactory when the bearing was subjected to the high speeds and heavy wear due to the radial, lateral, and axial strain incident, for example, to automobile service. • In this regard the practical ball bearing art as then practiced prior to Conrad’s patent was well described by Judge Holland in 177 Fed. 436 in these words:

“Many patents have been issued for ball bearing devices, which have not been entirely satisfactory, for the reason that the tracks or ways were interrupted, and the balls consequently could not travel freely therein. It was old to have inner and outer rings with opposing grooves, but the sides of these *725grooves were interrupted in one way or another to permit the introduction of the balls. In some eases filling openings were provided, and in some instances these were filled up or plugged after the balls had been introduced in order to prevent the escape; but these prior devices were defective, in that the raceways would crumble or wear at the interrupted parts of the raceway, and then the injured balls would cause undue wear to the remaining portions of the raceway, and thus the bearing suffered a rapid depreciation, and often an entire failure in a comparatively short time, and where the filling openings were plugged to prevent the escape of the ballsi the plugs could not be given precisely the same temper as the rings forming the remaining portion of the raceway, and unequal wear would ensue, which resulted in injury to the bails and raceway, and an undue shortening of the life of the hearings. These bearings could safely be subjected only to light loads, and were entirely unsatisfactory, and not fitted for use in heavily built, rapid moving vehicles.”

The proofs show that such wear was sometimes caused by twisting, sliding, or uneven pressure of the balls upon the noncontinuous portions of the grooveway. When this wear once began, the bearing would disintegrate very suddenly, indeed, within a few hours’ time. Without entering into detail, it suffices to say the proofs in that regard are;

"It was, therefore, impossible to use ball bearings for ordinary loads and speeds as found in ordinary machinery, prior to the Conrad bearing. The Conrad bearing was, therefore, the pioneer in being the first instance in the art of a ball bearing capable of withstanding or enduring the loads and speeds of ordinary machinery for a period of time as great as the life of such machinery. The quality of the Conrad bearing which enabled it to accomplish this new and pioneer result was its quality of substantial wearlessness. It must be understood that this wearlessness or substantial wearlessness has a different relation and significance than might be casually thought when such a quality was mentioned. It is not tjiat wearlessness is desirable for its own sake, but that wearlessness is essential in a ball bearing, because the moment that any substantial wear has taken place the period of rapid disintegration arises and the life of the bearing is ended. * * * This makes an accuracy essential in the balls and races to a standard of about one ten-thousandth of an inch. Accordingly, if a bearing is not almost absolutely wearless, it will within a very short time wear more than one or even several ten-thousandths of an inch, so that, even if it were originally made with accuracy, its accuracy is terminated by this very slight wear, .fust as soon as this accuracy is lost, the bearing commences to quickly disintegrate. Accordingly a ball bearing to bo capable of successful general use must bo almost absolutely wearloss, because it cannot be allowed to wear even to the extent of a few ten-thousandths of an inch during its entire life. Hence the importance of an almost absolutely wearless bearing is not merely to provide a long life, but in order to provide practically any U]e at all. Thus all bearings prior to Conrad did not differ merely in degree to the extent of having merely a shorter life, but they did not have practically any life at all. It was only the quality of wearlessness produced in Conrad’s bearing which gives the bearing its life equal to the life of the mechanism in which it may be used, * * * The Conrad bearing was the first wearless ball bearing, meaning, of course, practically or substantially wearless, under normal conditions. Of course, I do not mean to affirm that the Conrad bearing will not wear somewhat if subjected to overloads or conditions where grit is encountered, or acid in the oil. As bearings go into the custody of all sorts of persons, who use different machinery, a slight wc-ar may occur even with the Conrad bearing; but eliminating this aspect of the matter, what I wish to affirm and make clear is that prior to the Conrad bearing all ball bearings showed rapid wear, no matter how carefully they were made or used, and that the Conrad bearing was the first in. stance of a ball bearing capable of enduring ordinary loads and speeds without substantial wear.” .

*726One cause of this disintegration was, as we have seen, the noncontinuity or break of the groove path to provide for ball admission. In that regard the proof is:

■ “An interruption in one form, or another in one or both the races, to permit the insertion of the balls, always left a weak point, much less able to withstand loads than the other portions of the race not so weakened. Disintegration and crumbling of the bearing was likely to occur within a short time at the weak point of the race. As soon as this disintegration or crumbling commenced it proceeded rapidly to destroy the bearing. It is obvious that this result would occur, because even a slight injury to tfie race produces an injury to the balls, which causes further injury to the races, and in turn to the balls, and so on.”

In this state of the art Conrad disclosed the device of the patent. It is exceedingly simple, and in the after-light of its simplicity it is remarkable that no one had suggested its earlier use. Narrowed to what it really is, Conrad simply brought into the art a metallically unbroken groove pathway for the balls. This pathway he kept unbroken by introducing the balls by eccentric displacement of the rings; the retention of the balls and the concentricity of the rings he effected by ball spreaders. The elements he used were individually all known. There was no originality in a spreader to keep rollers or balls from contracting with each other, as also there was none in placing two concentric rings in eccentric relation. Conrad’s novelty and contribution to the art consisted in disclosing the actual use of a continuous or unbroken groove as a pathway for a ball bearing. As a practical means of putting his continuous pathway in use he suggested ball loading by eccentric displacement, and as a means of preventing unloading by eccentric displacement he showed the use of spreaders. This resulted in the production of a unitary, workable device, to wit, a ball bearing, for which his patent was granted.

A study of the patent shows groove continuity was the substantial disclosure of Conrad’s specification. For example, it said:

“The invention provides a ball bearing having concentric grooved rings, the sides of the grooves being uninterrupted throughout their circumference."

From this it will be seen that, while the rings themselves have concentric grooves, it is the grooves in the rings and the sides of those grooves that are unbroken or uninterrupted throughout their entire circumference. This feature of unbroken groove continuity is further emphasized. Thus Conrad says:

“The principal advantage of the new bearing lies in the continuity of the sides of the groove.”

He also asserts that it is this continuous groove continuity that mechanically makes the ball bearing durably effective. Thus, speaking of “the continuity of the sides of the groove,” he says it “insures the regular running of the balls, and consequently great durability of the bearing.” That the unbroken groove continuity itself was the essential feature, and that such groove was differentiated from the face or ungrooved surface of the ring, is evidenced by the several claims. For example, in claim 1 we note, “A ball bearing including two concentric rings having opposing grooves on their adjacent faces, the sides of *727said grooves being uninterrupted throughout their circumference;” in claim 2, “a ball bearing including two concentric rings having opposing grooves on their adjacent faces, the sides of said grooves engaging the balls to prevent substantial lateral movement, said sides being uninterrupted throughout their circumference;” in claim 8, “a bearing comprising two concentric rings, balls between said rings, each ring having a groove, both sides of which overhang said balls and are continuous and practically integral throughout their circumference’— terms also carried into claim 9. From these extracts it will be seen that the designated, effective, and claimed feature of Conrad’s disclosure was the groove with unbroken, integral, continuous sides. In the old art it was the cut into the wearing side of the grooved pathway that had developed an exposed wearing weakness, and it was this groove incision or groove entry that Conrad eliminated and claimed. It was, of course, obvious that this device of an unbroken, side-wearing surface of his groove showed, in his method of applying his unbroken groove, that he also obviated a cut in the ring face beyond the groove edge, for he says:

“The sides of the grooves, and in fact, all the parts of each ring are continuous and practically integral throughout the entire length of the ring.”

But a careful scrutiny of the patent makes it clear that this unbroken feature of the ring was merely a mechanical statement made in properly describing the unbroken character of the groove, and that nowhere in the specification or claims was any other reference made to the unbroken character of the ring face, or was any functional effect attributed to or claimed for an unbroken ring as such. Indeed, Conrad’s sole reference to an unbroken ring is where the specification says:

“Each ring has a groove, the sides of which overhang the balls to a slight extent. The sides of the grooves, and, in fact, all the parts of each ring, a re continuous and practically integral throughout the entire length of the ring.”

[1] It is contended, however, that the proceedings in the Patent Office should in some way narrow the effect to be given the claims here involved. There is no doubt that where, pending allowance of a patent, an applicant is forced to narrow his claims, he will not be permitted to put an after construction on them which would make them include what he had been forced to avoid by narrowing. But that principle has no application here, for a critical examination of this file wrapper shows that Conrad came out of the Patent Office with broader claims than those originally made. In other words, the several amendments and the withdrawal and substitution of claims resulted in an elimination of extrinsic matter, in pointing out the essential feature of his disclosure, and in the grant of claims embodying that essential feature and eliminating from his claims such extrinsic, nonessential features. The specification as originally made, together with all the claims, were not satisfactory to the Office — a fact possibly due to the nonfamiliarity of the foreign applicant with American patent requirements — and in the subsequent proceedings each and every of the original claims were rejected. The entire original specification was then *728withdrawn and another substituted. These facts are all important, for the contention of the defendant in effect is that the claims granted should be given the effect the claims originally made would have had. That the device itself for which Conrad received his final claims was the same one for which he sought his original claims there is no question. The eight figures of his drawing and their lettering remained unchanged. But, while Conrad had in view the broad, general character of the ball bearing he there disclosed and illustrated, we think his specification as originally drawn did not specifically and with exactness define wherein the precise inventive feature of his device lay, nor did he properly confine his claims to such inventive feature. As we have seen, the gist of the ball bearing disclosed by Conrad was the unbroken, continuous, unrecessed,• and integral grooved ball-pathway. When, however', he came to make all three of his original claims they were each and all not for such grooves, but for rings “unrecessed and unbroken.” While» of course, such an “unrecessed and unbroken” ring undoubtedly physically had on its face Conrad’s “unrecessed and unbroken” groove, yet the unbroken face of the ring was a mere mechanical strengthening incident; for, no matter how much the ring face, as a face, was unbroken and unrecessed, the invention did not lie in the integrity of the ring face, but in the continuity of the sides of the groove. Such being the case, it is clear that the claim as originally made for rings “unrecessed and unbroken” needlessly embodied in the claim the narrowing limitation of necessitating the ring surface to be unbroken, when the real invention lay in requiring the groove side to be unbroken and unrecessed. And, anticipating what will later bé apparent, we may here say that, if Conrad’s claims had thus remained for “unrecessed and unbroken” rings, the defendant would have avoided infringement of this narrow claim by merely slotting its rings, though in doing so it had cut into and broken the continuity of the sides of the groove. Presumably attention must have been called to this fact by the Office, since'in the new specification and the claims which met the Office’s approval this oversight was remedied, and in the new specifications as earlier quoted herein it was shown that “the principal advantage of the new bearing lies in the continuity of the sides of the groove,” and new claims were made, not for rings “unbroken and unrecessed,” as before, but for “opposing grooves on their adjacent faces, the sides of said grooves being uninterrupted through their circumference” — “each ring having a groove both sides of which overhang said balls and are continuous and practically integral throughout their circumference.”

As originally made, the pathway secured by the claims would have been restricted to a pathway in a ring which was unbroken and unre-' cessed. As allowed, the requirement, but not the claim, was narrowed, the ring feature was eliminated, and the “unbroken and unrecessed” requirement was restricted to a groove. In such case the lessening of the requirement was a broadening of the claim. That the claims granted were broader in scope than the ones originally made is evidenced'-by the fact that, to show infringement of the claims as originally made it would be necessary to show the infringer used rings— *729“said rings being unrecessed and unbroken”; whereas, with the claims as granted, it is only necessary to show the use of “a groove both sides of which overhang said balls and are continuous and practically integral throughout their circumference.” Such being the case, it follows that the claims granted should receive the construction their language naturally imports (Dodge Needle Co. v. Jones [C. C.] 153 Fed. 189, and 159 Fed. 715, 86 C. C. A. 191), and that no statement or action of the patentee in obtaining his patent estops him from claiming to the full extent what his claims on their face purport. We think the court below therefore failed to give proper construction to this patent when it held that “it must be confined to the rings, solid and unbroken throughout, upon which Conrad laid repeated and emphatic stress,” for the fact is that in his specification as eventually accepted Conrad laid stress, not on his rings being solid and unbroken, but on the unbroken character of the grooves.

[2] The device of Conrad has proved meritorious and it has gone into exceptionally wide use. Its worth and the validity of his patent have met with wide commercial recognition in the licenses taken under it. We find nothing in the way of prior patenting or use that suffices to shear Conrad’s device of patentable novelty; for, without entering into a discussion of the prior art, we may say that, while ball bearings were known and used, there was prior to Conrad no use of ball bearings in the high speeds and heavy loads which his device has made possible. No prior device left any impress on the ball bearing art to which his device applied. Indeed, as the proofs show, Conrad’s device made practical the use of such bearings in the case of high speeds under heavy loads. We therefore hold the claims of his patent here involved are valid.

We next turn to the question of infringement. Stripped of all extrinsic matter, that question practically and mechanically turns on the effect and sufficiency as a ball confiner of a steel dam or barrier so slight in height that its gradual rise from the central line of the balls’ grooved pathway ceases when 44/ioooo of an inch is reached. In other words, the defendant forms an unrecessed, continuous, integral grooved pathway for its balls until a level of 44/ioooo of an inch is reached, and from that point until the outer edge of its ring is reached, not the whole ring face, be it observed, but a channel less in diameter than the ball, is cut and recessed in a narrow cross-section of the ring face.

The defendants’ device is made under the patent to Blin, No. 818,-734, of April 24, 1906, for a ball bearing. Describing such device from the specification of that patent, we note that, according to Blin’s invention :

“ * * * The outer portion of the inner ring is provided with a notch or cut-away part, which starts from one of the sides of the rings and, being of less depth than the groove containing the balls, does not therefore extend to the bottom of that groove. In the same way on the inside of the outer ring is made another notch or recess, so that when the two rings are placed concentrically with their two notches facing each other the distance separating the two walls of the notches is smaller than the distance between the two track surfaces on which the balls roll.

“When a sufficient number of balls has been introduced into the bearing in order to keep the two rings .concentric, a ball is introduced between the two *730notches when they face each other, Fig. 1; but it cannot enter the ball track unless forced in. However, owing to the elasticity "of the material, the ball can be driven in, where it remains. All the remaining balls will be introduced in the same manner. Thus the two rings by themselves constitute a complete ball bearing and ball retaining device without the use of auxiliary-parts.

“In order to remove a ball, the two notches, 5, d, are brought opposite each other, and the ball opposite the said notches is forced out from the other side between the two notches. The other balls can be removed in the same 'way until a sufficient number are removed to allow the bearing to fall apart. The result of this arrangement is that it is not necessary to close the notches in order to prevent the balls from escaping. It will, moreover, he understood that in this construction the traveling path of the balls does not present any lack of continuity due to joints between various pa/rts.”

It is clear, therefore, that after all the balls are inserted all the parts co-operate precisely the same as if] the channel did not exist for all the .ball contacting pathway of the groove is unbroken and continuous throughout its entire circumference. In that regard the defendant, in 'describing its ball bearing in its prospectuses, says:

“F. & S. bearings are made with diagonal side entering slots slightly smaller than the sizes of the balls themselves. The balls are therefore sprung in under greater pressure than they ever receive in service, the races yielding sufficiently to let them "enter. Once in, the balls never again touch the side slots, which virtually then ño longer exist.”

Such device would appear to differ in three respects from Conrad’s, viz.: Its ring face is channeled, notched, or recessed; a larger number of balls can be inserted through such channel than by Conrad’s unchanneled ring method; no spreaders are required to maintain defendants’ ring concentricity. As assembled ball bearings, however, both devices operate in the same way, and the unbroken grooves in which the pathway of the balls of both devices lie would seem equally effective in avoiding the evils of the noncontinuous groove of the prior art. Under such facts is infringement shown ? The court below thought not, and, as we understand it, held that the claim element, “each ring having a groove both sides of which overhang said balls,” was not found in defendants’ structure, because a groove, the sides of which only rose 44/ioooo of an inch, did not overhang the ball. In that regard, it said':

“How such a minute fraction as 44/ioooo of an inch would ‘overhang’ is not perceptible,” and “I do not understand how The sides’ of the raceway can be continuous and practically integral if they are cut in half, or so nearly in half that 44/ioooo, of an inch further would finish the job, and this is the conceded situation of the defendants’ bearing.”

Turning, then, to the question whether the defendants’ groove does overhang its balls, we should note that, however slight the overhang is, it is sufficient to cause this result. Indeed, as we have seen, the defendants not only concede that fact, but positively assert it:

“Once in, the balls never again touch the side slots, which virtually then no' longer exist.”

At bar, counsel for complainant stated that if this defendant cut its channel entrance deeper by 44/ioooo of an inch all claim of infringement would be •abandoned. Slight, therefore, as the difference is mechanically, it is evident its presence is a matter of functional substance. *731That such minute exactness as a metallic cut of 1/ioooo of an inch can under modern machinery methods be made is well known. And in that connection it must not he overlooked that the slight rise in the indented plane of the thread of a screw, as afterwards shown, was the dam or harrier that prevented the Reiss telephone from transmitting human speech. It is the function of the barrier, not its size, that counts as a measure of invention. That the groove does “overhang” the ball is, we think, self-evident. In the nature of things the overhang or lesser diameter of the exterior of an inclosing groove than the diameter of the inclosed ball is the only agency that keeps the ball from escaping. Since, therefore, the balls of defendants’ bearing in fact do not escape, since there is no other agency than the groove sides to prevent escape, and since the diameter of the balls is greater than the diameter at the edges of the groove, it is physically evident that the groove sides do overhang and prevent the escape of the balls. Moreover, it must not be overlooked that, while the end of the channel through the ring is but 44/ioooo of an inch higher than the bed of the groove at that particular point, the rest of the entire surface of the ring is solid at every other point and makes a much deeper sided groove at other points than the channel mouth, and, as noted above, “in order to remove a ball, the two notches are brought opposite each other, and the ball opposite the said notches is forced out from the other side between the two notches,” manifestly this barrier, slight as it is, requires force to pass it.

It is therefore clear that the balls of defendants’ structure when in operation do travel in a groove which overhangs such ball. It is equally .clear that such pathway of the balls is a groove whose sides are “continuous and practically integral throughout their circumference.” It would therefore seem that, so far as that feature of the claim infringement is shown, unless the defendants’ structure does not answer the other calls of the claim, viz.:

“The number of balls being such that they can be inserted in the space between the rings when the latter are displaced from their normal positions,” .and “means for distributing the balls throughout the length of the groove, whereby the two rings are held together against axial displacement by the engagement of the balls with the overhanging walls of the grooves and the parte are held together so as to form a unitary device.”

Now it is clear that defendants’ two concentric rings are adapted to eccentric displacement, and by such eccentric displacement a ball bearing may be assembled which is accurately described by the foregoing element, viz.:

“The number of balls being sueh that they can be inserted in the space between the rings when the latter are displaced from their normal position.”

Indeed, the proof shows that in actual practice the defendant so filled its rings, viz. :

“Both of the rings were eccentrically placed. Then as many balls as possible were introduced between them; then the rings were put concentrically, and the additional balls were put in through the filling notches.”

The defendants’ mechanism then being such as to eccentrically introduce in the same way as Conrad the number of balls Conrad requires, ■does the defendant, as it contends, relieve itself of infringement by fur*732ther providing a channel through which an additional ball or additional balls may be introduced? While Conrad did not show a slot for such additional ball charging, it is clear he did not restrict himself to the number of balls that could be introduced into his bearing, for he expressly says:

“The number of balls which can be introduced can be increased by effecting a slight elastic deformation or tilting, and at the same time pressing an additional ball between the others.”

Indeed, the introduction of the additional balls by the defendant through the medium of its slotted channel is but the mechanical equivalent of the additional ball introduction by deformation which Conrad pointed out; for whether the size of an opening between two opposite surfaces be increased by chiseling or cutting the opening wider, or by forcing the sides apart, is, generally speaking, a matter of mere mechanical detail. By the slots the defendant may have increased the scope' of ball introduction: but such improvement or advance, if improvement it be, is but an added or alternative means of getting the ball bearing into working shape. When that is done by any of the several methods of introduction, viz., by eccentric or slot introduction, or by deformation introduction, the results in the unitary structure of Conrad and the defendant are precisely alike, in that each ring of both then has “a groove both sides of which overhang said balls, and are continuous and practically integral throughout their circumference.” And it is this unbroken groove, and the unbroken pathway the ball thereafter travels upon, and not the pathway by which the ball reached that channel, that is the gist and substance of this case. Conrad’s was the real low grade roadway over which tonnage was carried; Blin’s was but a switch to get additional tonnage on the main track. Moreover,, it is to be borne in mind that in increasing, over Conrad the number of balls introduced the defendant has utilized the same elements that Conrad had used to introduce a smaller number; in other words, the defendant has used rings adapted to eccentric introduction. Instead, however, of stopping with the Conrad ball number, the defendant goes a. step further and introduces additional ones. In other words, the defendants’ ball bearing has the number of balls that “can be inserted in-the space between the rings when, the latter are displaced from their normal position,” thus measuring up to the claim, and it has also the further balls introduced through the channel slot. But advance beyond or even improvement upon a patented device does not cease to infringe,, so long as the combination of the patented device is itself utilized in. the advance.

So, also, in the other element quoted, viz.:

“Means for distributing tbe balls throughout the length of the groove,, whereby the two rings are held together against axial displacement by the engagement of the balls with the overhanging walls of the grooves, and the parts are held together so as to form a unitary device.”

It will be noted that it is the maintenance of a proper ball distribution which prevents recurrence of eccentricity. In Conrad’s method' the number of balls eccentrically introduced was such that, unless some means were provided to maintain the proper spacing or distribution of *733such balls around entire ring circumference, the eccentrically charged balls were liable to be in the same way eccentrically discharged. This difficulty Conrad met with his spacers, saying:

“Tlio edges of the rings, however, are spaced so far apart from each other that they may he displaced eccentrically relatively to each other in the manner shown in Fig. 3, leaving a crescent-shaped space of sufficient width to permit the introduction of a limited number of balls. The crescent-shaped space is marked d. The rings may be then restored to their concentric position and spreaders or distributing devices introduced into the spaces between the balls, so as to distribute them entirely around the raceway and to prevent their return to a position such as Fig. 3, which, would permit the escape of the balls. * * * It will be seen that the spacers f hold the balls in the position of Fig. 1 — -that is to say, in the distributed position — preventing the balls from running together, and thus allowing one of the rings to fall down against the other and release the balls through the crescent-shaped space.”

But this specific means of distribution was not carried into Conrad’s claim as quoted above, but the generic term, “means for distributing the balls,” etc., was used. It is contended, and indeed the court below found, that the defendants’ bearings lack a hall-spacing, device, which is not necessary for preventing the hearing from falling apart, and has no such function. That the defendant has no spacer such as Conrad showed is a fact, but that‘it has “means for distributing the balls”— that is, means for distributing the limited number of balls liable to eccentric discharge — -is also a fact. That both Conrad and the defendants’ devices do not eccentrically discharge, and that they are both, when charged with the eccentrically chargeable quota, liable to discharge, is clear. Why, then, when they are used as a ball bearing, do neither of them eccentrically discharge ? In Conrad’s device the reason is obvioxxs. Conrad’s spreader compels ball separation, and so prevents the balls from gathering in one part of the ring, and thus forming the objectionable “crescent-shaped space of sufficient width to permit the introduction”' — -and therefore the escape — “of a limited number of balls.” And how7 does the defendant meet this difficulty ? What means does it use to prevent the formation of this “crescent-shaped space of sufficient width to permit the discharge of balls” ? Analysis shows that its means is the combined use of a slot and of the extra channel-introduced balls. In other words, the slot, or, more accurately, the additional or slot-introduced balls, are the defendant’s means for distributing the other “balls throughout the length of the groove, whereby the two rings are held together against axial displacement.” Conrad’s device by spreaders prevents the eccentrically charged balls from gathering in a section of the ring where they would form a crescent of discharge. The defendant allows the same number of halls to gather in the same dangerous place, but by its additional channel-introduced balls, which cannot get into such space, prevents the crescent from opening and discharging. It is therefore clear that the slot-inserted, additional balls of defendant’s device are the mechanical equivalents of Conrad’s spreaders. For if we take Conrad’s device we can dispense with his spreader by simply cutting a lateral channel of such depth as will permit slot introduction. And after all said and done this is simply what the defendant has done. In the self-confessing words of the defendant’s prospectus quoted above, viz., “Once in, the halls never again *734touch the side slots, which virtually then no longer exist,” and in the statement of the Blin patent, “It will be understood that in this construction the traveling path of the balls does not present any lack of continuity due to joints between various points.” It is clear that Blin and the defendant use the body and soul of Conrad’s ungrooved pathway and to- it add’ “a notch or cut-away part, which starts from one of the sides of the ring and, being of less depth than the groove containing the balls, does not therefore extend to the bottom of that groove.” That the extra, slot-introduced ball or balls are so introduced simply as a preventive of eccentricity is, we think, conceded and rightfully conceded by Blin’s patent. Thus he says:

“When a sufficient number of balls has been introduced into tbe bearing in order to keep the two rings concentric, a ball is introduced between tbe two notches when they face each other, Fig. 1; but it cannot enter the ball track unless forced in. However, owing to elasticity of the material the ball may be driven in, where it remains. All the remaining balls will be introduced in the same manner. Thus the two rings by themselves constitute a complete ball bearing and ball retaining device without the use of auxiliary parts.”

The “sufficient” number of initial balls makes ring concentricity, and the slot-driven balls maintain such concentricity. And in that connection it should be borne in mind, .as bearing, on the question whether the minute rise of 44/ioooo formed a barrier which kept the ball in its path, it will be evident that if the height of Blin’s notch or cutaway was sufficient, as he says it was, that a slot-introduced ball “cannot enter the ball track unless forced in,” it is equally clear that the ball cannot leave the pathway unless it be forced out, and the only way in which that barrier can be overcome and the ball leave the trackway he points out:

“In order to remove a ball, the two notches are brought opposite each other, and the balls opposite the said notches are forced out from the other side between the said two notches.”

It goes without saying that if the ball has to be thus forced out when a double-sized opening is presented by the two cuts in the two aligned notches, that the half-sized presented by one notch and a solid opposite ring makes a doubly higher grooved pathway. Such being the case, we think it clear that the sides of defendant’s groove do- overhang the balls, and that its overhang, be'it only 44/ioooo, which in the practical application of the Blin patent the defendants use, is an effective barrier-to hold the revolving balls to their orbit.

The decree dismissing the bill must therefore be reversed, and the case remanded, with directions to the court below to enter a decree holding the claims in issue valid and infringed and directing an injunction and accounting.

Rehearing

Petition for Rehearing.

PER CURIAM.

The motion for rehearing has had our joint consideration. The defendant firm was formed but a short time before this suit was brought, and is only liable to account for the period subsequent to its formation. Bearing in mind the several prior cases against others in which the rights of the patent here involved were vigorously prosecuted, and the not undue time that elapsed between the *735successful termination of such litigation and the commencement of this suit, we do not find any such delay as should preclude the plaintiff from the customary accounting. The defendant firm volunteered to enter the field of infringement, and accounting is a necessary consequence of its act. As to the other matters involved, we see no reason to depart from the conclusion reached and stated in the opinion heretofore filed.

The motions are therefore refused.