COLT, Circuit Judge.
These cases relate to a shoe-sewing machine, which forms a chain stitch by means of a curved hooked needle. The machine is commonly known as the “welt and turn” sewing machine, because it is used for sewing the upper and welt to the insole of a welt shoe, and the upper to the insole of a turned shoe. The sole is prepared for sew-ing by making a channel around its outer surface at a little distance from the margin, and the lin of leather between the channel and the margin of the sole is known as the “be-tweeñ substance.” In a turned shoe the needle passes through two layers of work, — the upper and the between substance of the sole. In a welt shoe the needle passes through three layers of work, — the welt., the upper, and the between substance of the sole. The needle pierces the work with a forward and upward movement, and a double line of thread is left outside the upper or welt, and a single line in the channel. In the simplest form of the machine, the parts which work with the needle in forming the stitches are a looper and a tension device. The hooked needle is first thrust through the leather, the thread is then laid into the throat of the hook by the looper, and the [394]*394needle is drawn back through the hole, pulling a loop of thread. The leather is then moved along the length of one stitch, and thg needle is again thrust through the leather, again threaded, and again drawn back, pulling a second loop of thread through the leather and through the first loop, and this process is repeated for each stitch. The thread is drawn from a spool, and extends from the spool to the leather, through the machine. The looper is usually a small finger or arm, with a hole or eye in its end¡ through which the thread always passes. The tension device is commonly a wheel with a grooved edge. The thread is wrapped around the groove of the wheel so that, as the thread is drawn from the spool, it will turn the wheel. By means of a spring and a screw, the operator can vary the extent to which the tension wheel will resist the pull of the needle. A tension device is necessary because, in pulling a loop of thread through the leather and through the last preceding loop, the needle must meet enough resistance in drawing the thread from the spool to make it draw the last loop tightly around the loop on which the needle is pulling.
For ro years prior'to 1890 the standard machine in this art was known as the “Dancel Machine,” disclosed in letters patent No. 190,709, dated May 15, 1877. Machines adapted for this work must be run at a comparatively high tension. When the Dancel machine Was run with insufficient tension the stitches would not be properly set, and when it was run with sufficient tension the thread would sometimes break, or cut the spongy between substance. These defects in the Dancel machine were caused by imposing too much work upon the needle. In the operation of the machine, the needle was forced to pull the thread for the next stitch against the strain of the tension; thereby causing the thread to render through the needle hook, which weakened the thread by chafing. Upon the needle also was imposed the duty of pulling the thread from the preceding loop through the needle hole and around the edges of the between substance, and setting the stitch by a downward pull, which tended to cut the between substance, and to pull apart the sole and upper which were being sewed together.
To remedy these defects in the Dancel machine, it became necessary to relieve the needle of a portion of it's work by means of some auxiliary mechanism. This auxiliary mechanism is known as a “puli-off” and “take-up.” A pull-off of this class, as generally understood, is a device for pulling off enough thread from the tension to form the next stitch, thereby preventing the rendering of the ,thread in the hook of the needle; and a take-up is a device for taking up the thread and setting the stitch, thereby preventing the cutting of the between substance by drawing the thread around its edges. A pull-off or take-up is commonly a wheel or truck, located between the looper and the tension wheel, which moves and presses against the thread. This truck, when it moves and engages the thread, operates as a pull-off when the thread is held at a greater tension strain on the side of the material than is exerted by the tension wlieel, and operates as a.take-up when the thread'is held at a less tension strain on the side of the material than is exerted by the tension wheel. [395]*395Thus the same truck commonly acts both as a pull-off and take-up. It acts as a pull-off when the strain on the thread is sufficient to overcome the tension, because it then necessarily pulls off thread from the spool; it acts as a take-up when the strain on the thread is less than the tension, because it then necessarily takes up the thread. The terms “pull-off” and “take-up” are sometimes used interchangeably, for the reason that a so-called “pull-off” may have a take-up function, or a so-called “take-up” a pull-off function.
While theoretically and practically a pull-off and take-up of this class should, as far as possible, entirely relieve the needle of its old duty, by pulling off sufficient thread for the next stitch, and taking up sufficient thread to set the stitch, there may be pull-offs and take-ups whose operation is more limited, and which only in part free the needle of its former work. Such auxiliary mechanism, while it pulls off some thread, does not pull off a sufficient quantity for the next stitch, and, while it takes up some thread, it is not sufficient to set the stitch. In a case where a patented pull-off and take-up perform substantially the whole work required, and the alleged infringing pull-off and take-up only a portion of the work, the question of infringement must turn largely on the scope of the patent as determined by the prior state of the art and the language of the specification and claims. If the patented pull-off and take-up are merely improvements upon old devices for doing the same work, or if the patentee has imposed a limitation on his invention by express words in the specification and claims, it is manifest that the court would not be warranted in giving, that breadth of construction to the patent which should be given in a case where the inventor was the first to devise a pull-off and take-up mechanism, or had not expressly restricted his invention by the terms of the specification and claims.
The auxiliary mechanism which may.be added to the old Dancel machine may be a device which operates both as a pull-off and a take-up, or it may be a take-up which, with some additional mechanism, acts also as a pull-off, or it may be a device which operates simply as a pull-off.
The present suits relate to patents for this auxiliary mechanism. The Spaulding suit is brought on the French and Meyer patent, No. 412,704, dated October 8, 1889, and the Cook suit on the Fowler and Warren patent, No. 564,986, dated August 4, 1896.
Taking up, first, the Cook Case, we find that the Fowler and Warren patent is for a pull-off truck which pulls off a full supply of thread for the next stitch, and so reduces the gendering of the thread through the hook of the needle to a minimum. If this patent disclosed for the first time a pull-off mechanism of this kind, which prevented rendering, and so overcame one of the two chief objections to the old Dancel machine, it would be entitled to receive a liberal construction; and, if the language of the claim was sufficiently broad, it ■ might fairly be held to cover' a pull-off device which pulled off any appreciable quantity of thread, and thereby prevented, in any degree, the rendering of the thread through the hook of the needle. But this patent does not occupy any such position in the art.' Not only at the time were there pull-off devices for doing this work, but [396]*396there existed also a device which acted both ás a pull-off and take-up. This latter device cured both defects in the old Dancel machine, and proved of such utility and value that machines with this auxiliary .mechanism fast supplanted the Dancel machine.
;It is contended that the Fowler and Warren patent is a backward step in the art. This statement may be too strong. We think, however, that it is far from representing any marked advance in the art, because there were other practical devices then in use which combined both the functions of a pull-off and take-up, while this device is limited to a pull-off action. It also appears that the device has never gone into use except to a very limited extent, and its practical utility may at least be questioned.
. The patent is simply for an improvement on old devices for doing the same work. It is for an improvement on the pull-off of the .Fowler and Warren patent, No. 564,985, of the same date as the patent in suit, and the pull-off of the prior French and Meyer patent in suit.
In the first Fowler and Warren patent there was a thread brake .which constituted a part of the pull-off device, and in the French and Meyer patent certain additional mechanism formed a necessary part of the pull-off device. The purpose of the Fowler and Warren patent in suit is to simplify the construction of the pull-off device by the • omission of the thread brake in the first Fowler and Warren patent, and the additional mechanism in the French and Meyer patent.
This is accomplished by so timing the actuating mechanism of the machine that the action of the pull-off takes place when the thread is held under strain in the hook of the needle at the completion of its backward stroke; in other words, the machine is so organized that the needle does the work of the thread brake or other additional mechanism.
The specification, after declaring that the invention is for an improved device for doing the same work as the first Fowler and Warren patent, says: “Its main novelty being that the, pull-off mechanism does its work after the needle has completed its loop-drawing stroke, and while the needle holds a loop of thread in its hook under the strain requisite to set the stitch, instead of relying upon a thread brake to hold the- thread against the action of the pull-off truck, as described in our application above referred to.” The specification also points out the two advantages of the patented device: First, “the substantive advantage qf holding the thread in the hook of the needle,” whereby “the construction is simplified by the omission of a thread brake or its equivalent”; and, second, “the advantage derived from holding the thread in the hook of the needle when the pull-off is operated, rather than holding such thread by the needle shank,” as in the French and Meyer patent, which obviates the necessity of using the additional mechanism of that patent. In the French and Meyer patent the pull-off truck operates first as a.take-up, .drawing back the thread until the loop is tightened about the shank of the needle, “thereby necessitating the use of an additional mechanism to form a bight of thread between the stock and the throat-of the needle,” while in the Fowler and Warren patent “a [397]*397supply of thread for the loop about to be drawn is maintained in the preceding loop.”
The Fowler and Warren patent in suit, like the first Fowler and Warren patent and the French and Meyer patent in suit, describes a pull-off device for pulling off the full length of thread for the next stitch, and so reduces the rendering of the thread in the hook of the needle to the least possible quantity. The machine is organized, says the specification, “to pull off thread from the thread supply, * * * and thus supply slack thread for the next stitch.” Again: “When the pull-off draws thread from the tension * * * the full length of thread is drawn from the tension, and the rendering of the thread while under strain through the hook of the needle is reduced to the minimum.”
The single claim of the patent is as follows:
“In a chain-stitch hook-needle sewing-machine, the combination of tension, looper, book needle, a pull-off mechanism between the needle and the tension, and actuating mechanism timed to cause the pull-off mechanism to make its pulling stroke after the hook needle lias completed its loop-drawing stroke, and while the loop is held under strain by the hook of the needle, substantially as described.”
Reading this claim in connection with the specification, the “pull-off mechanism” and “its pulling stroke,” “substantially as described,” must be construed to mean a pull-off which operates to draw a substantial!}? useful supply of thread for the next stitch. To £on-strue the claim otherwise would be to hold that any pull-off device which operates to pull off any appreciable quantity of thread at the particular time, and in the way described, would come within the patent. If, by so timing the actuating mechanism of the machine that the pull-off truck operates in the manner described, the pat-entees had solved any problem, or'the invention had proved to be an important advance in the art, it might be, notwithstanding the language of the specification, that this claim should be construed to cover any pull-off which acts to pull off any amount of thread at the time and in the manner described. But the state of the art, and the comparative unimportance of this improvement, forbid such a liberal interpretation of the claim; and, in our opinion, the court should not give a broader construction to the claim than to hold that it covers a pull-off truck which pulls off a substantially useful amount of thread for the next stitch, while the loop is held under strain in the hook- of the needle at the completion of its backward stroke.
Turning now to the Spaulding Case, we find that the French and Meyer patent, which is involved in that suit, covers, as we have already observed, both a pull-off and take-up. It describes auxiliary mechanism for taking up and setting the last stitch, and for pulling off sufficient thread for the next stitch. It is for an ingenious and meritorious improvement upon the Dancel machine, in that it embraces the means for overcoming both the faults of that prior structure; and machines embodying this improvement have gone into extensive use, and, to a large extent, have displaced the Dancel machine.
[398]*398■The French and Meyer device comprises a movable truck and additional mechanism. The truck is operated in such order of time with relation to the needle that it first acts as a take-up, and “alone performs the duty of setting the stitch, thus relieving the needle of the duty of drawing up the previous loop to set the stitch.” The take-up operation takes place “while the needle is in the stock, and while the loop of thread last drawn through the stock by the needle is yet on the shank of the needle.” Each stitch is set before the loop of the next stitch is drawn through the stock. Continuing its backward stroke, the truck acts as a pull-off, pulling off “enough thread for the next stitch, and leaving it to be taken up and held or cared for by an auxiliary take-up, the latter giving up the thread so drawn off to the needle as the latter is being drawn back through the stock.” Thus, the needle draws substantially slack thread while passing backward through the stock. After the needle has emerged from the stock, it gives “a sufficient pull to draw the thread of the loop being formed snugly back to the previous stitch, leaving it taut on the upper or lining.” With the auxiliary take-up there is combined a thread-holder, the purpose of which is to act upon the thread between the stock and the thread guide, and to draw from the supply of thread held by the auxiliary take-up enough thread for the next stitch.
The charge of infringement is limited to the take-up feature of the French and Meyer patent, which is covered by the first three claims. The distinguishing characteristic of this take-up is that the stitch is set by a backward pull of the take-up instead of by a downward pull of the needle, as in the Dancel machine, and so avoids cutting the between substance.
Although the needle at its extreme backward stroke may, in some degree, help finally to tighten the stitch, there can be no doubt, upon reading the specification and first three claims of the patent, that the essential and fundamental function of the French and Meyer take-up is “to set the stitch” by drawing the loop of thread “taut about the shank of the needle then in the stock.” It is to set the stitch while the needle is in the stock, and this setting consists in pulling the loop taut about the shank of the needle. The specification begins by describing the objections to prior machines in which “the needle was obliged to set the stitch.” It then proceeds to declare that the patented take-up “alone performs the duty of setting the stitch.” Again, it says, “each stitch being thus set before the loop of the next stitch is drawn through the stock”; and again it refers to the stitch as “having been fully taken up and set by the action of the take-up.” Further, in speaking of the invention, just before coming to the claims, the specification saj^s: “Prior to our invention herein described, a wax thread sewing machine using a hooked needle to make a chain stitch has never been provided with a take-up to set the stitch.” In the same manner, the action of the take-up is specified in the first three claims, as follows: In the first claim, “the said take-up drawing the said loop about the shank of the needle, as described, to set the last stitch”; in the second claim, “the said take-up completing the setting of the stitch before the [399]*399said needle is withdrawn from the stock and out from the old loop upon its shank”; and, in the third claim, “the said main take-up to draw the thread about the shank of the needle and set the stitch while the needle is in the stock, and the loop of thread last drawn through the stock by it is yet on the shank of the needle.”
It is possible that the French and Meyer patent might have been so worded as to include any take-up which operates in part to relieve the needle of the old duty of setting the stitch, by pulling up a portion of the thread in the manner described, but the patentees have seen fit, in explicit terms, to limit their patent to a take-up which has the function of drawing up the full amount of thread and setting the stitch, and it cannot, therefore, be said that a take-' up which does not substantially do this work comes within the patent. To adopt such a construction would be to hold that a take-up which operates to pull up any appreciable quantity of thread, and which falls far short of setting the stitch, is an infringement of the patent. The express and oft-repeated language of the. specification and claims forbids such a construction.
Having considered the nature and scope of the two patents in issue, we now come to the alleged infringing device.
The defendants’ machine is the Dancel machine, with an added auxiliary mechanism, which is substantially described in the Bertrand patent, No. 589 403, dated September 7, 1897. While this auxiliary device acts (or may act) both as a pull-off and take-up, it admittedly neither pulls off the requisite supply of thread for the next stitch, nor takes up sufficient thread to set the last stitch, nor does it substantially do either of these things.
The defendants’ machine is provided with a slide bar, upon which is slidingly mounted a truck connected to the bar by means of a spring. When the slide bar moves to the left, it causes the truck to engage the thread between the needle and tension, and to press against it. After the truck has been brought into engagement with the thread, the slide bar, still continuing its movement towards the left, will have the effect of drawing thread from the tension after the resistance of the spring to expansion is sufficient to overcome the tension. The truck will then continue to draw thread from the tension until the slide bar has completed its movement to the left. After the slide bar has completed this movement, the needle, wdiich during this movement has remained substantially stationary in its retracted position, advances towards the work, slackening the loop which is held in its hook. As soon as the loop is slackened, the spring, which has been expanded by the movement of the slide bar to the left, contracts, and thus acts to advance the truck further to the left, and to cause it to take up the slack of the loop. This action of the truck in pulling off thread from -the tension constitutes its pull-off function, and its action in taking.up thread from the loop constitutes its take-up function. ' So far as this device acts as a pull-off, such action must take place while the thread is held under the strain of the needle at the completion of its backward stroke, as in the Fowler and Warren patent in suit; and so far as it acts as a take-up it draws upon the loop about the [400]*400shank of the needle as in the French and Meyer patent in suit, but whether the needle is in the stock during the take-up action is a question upon which the experts differ.
It was assumed in the opinion of the court below, based upon the testimony of the experts, and it is assumed by complainant’s counsel in their brief upon these apneals, that the defendants’ device may operate as a pull-off to the extent of pulling off from one-quarter to three-eighths of an inch of thread, and may operate as a ta.ke-up to the extent of taking up from one-quarter to three-eighths of an inch of thread; and the question is whether a pull-off and take-up device, having this limited action, infringes either or both of the two patents in issue.
As the length of thread necessary for a stitch is about an inch and a half, it follows, from what we have already said, that a device which only operates to pull off from one-quarter to three-eighths of an inch of thread does not infringe the Fowler and Warren patent. Further, the purpose of the Fowler and Warren patent is to prevent the rendering of the thread in the hook of the needle, or, as the specification expresses it, “the rendering * * * is reduced to the minimum.” It is doubtful, at least, whether the defendants’ pull-off reduces to any extent this rendering, for the reason that the quantity of thread drawn from the tension is so small that it is only sufficient to supply slack thread to the needle during the time it is drawing the loop through the material, at which time there is no rendering of the thread, owing to the bite of the material upon the throat of the needle. If this be true, it follows that the needle, after leaving the stock in the defendants’ machine, performs, the same duty as in the old Dancel machine, and is obliged to draw the thread against the strain of the tension," with the accompanying rendering.
Upon the assumption that the defendants’ device does not prevent rendering, the complainant contends that the needle during its backward pull through the material subjects the thread to a double strain, namely, the strain caused by the bite of the material and the strain due to the tension, and that the supply of slack thread to the needle at this time at least relieves the needle from pulling the thread against the tension strain. It may be the defendants’ pull-off in this way helps to relieve the needle of a portion of its work, but, if it does, it is a feature which is not referred to in the Fowler and Warren patent, and which cannot now be imported into that patent for the purpose of enlarging or changing its general meaning and scope as set out in the specification and claim.
With respect to the French and Meyer patent, we have seen that it describes a take-up which prevents the cutting of the between substance by pulling taut the loop of thread around the shank o.f the needle and setting the stitch, and that the patentees, by express language, have limited their invention to a device which has this function. As the amount of thread so taken up is about one inch, it is plain that- the defendants’ device, whose take-up action is limited to from one-quarter to three-eighths of an inch, and which does not operate either- to. draw the loop taut about the shank of the needle or to set the stitch, is no infringement of the patent.'
[401]*401In a sewing machine of this type, by reason of the number, smallness, and compactness of its parts, and the rapidity with which it runs, it is difficult to determine with accuracy the exact action of its several parts upon the thread during its operation. Consequently, a conclusion respecting infringement between two pull-off and take-up devices, where there is only a slight difference in the estimated amount of thread pulled off or taken up, might not be free from grave doubt; but that is not the situation we have to deal with. Here there are marked differences between the devices. The range of action of the complainant's pull-off and take-tip mechanism is far greater, according to the evidence, than the range of action of the defendants’ device. If some doubt exists as to the real character and purpose of the defendants’ device, it can at least be said that the complainant has failed to make out a case of infringement. In other words, the proofs fall far short of establishing that the defendants’ device comes within any legitimate construction of the complainant’s patents. Nor does it necessarily follow, as the complainant contends, that, if the defendants’ device does not infringe the patents in issue, it must be a useless appendage to the Dancel machine.
In determining the question of infringement in these cases, it would undoubtedly have been more satisfactory if the defendants had frankly and fully stated their object in attaching their device to the Dancel machine, and its effect, though we know of no rule which obliged them to make such disclosure. This course of the defendants naturally subjects them to unfavorable comment, and throws a certain suspicion upon the device, which it is impossible entirely to cast aside. But this circumstance cannot influence the judgment of the court. Its duty still remains to decide these cases simply upon the evidence presented.
Giving the defendants’ device the full scope of action which the complainant’s experts think it fairly possesses, it still remains true that infringement can be sustained in the Cook Case only upon the theory that the Fowler and Warren patent covers any pull-off device where the actuating mechanism of the machine is so timed that some pull-off action on the thread is exerted, however limited, when the needle has completed its backward stroke, and that it is immaterial whether the supply is sufficient for the next stitch, or whether there is any reduction of rendering in the hook of the needle; and that infringement can be sustained in the Spaulding Case only upon the theory that the French and Meyer patent covers any take-up which draws up some thread from the loop around the shank of the needle when in the stock, and that it is immaterial whether the amount is sufficient to draw the loop taut and to set the stitch. For the reasons already stated, we are unable to give such a broad construction to either patent, and it follows that the decrees of the circuit court dismissing the bills must be affirmed.
The decrees of the circuit court are affirmed, with costs.