The patent is for an improvement in a printing head used for consecutive numbering. Its purposes as declared are to prevent the adjacent digit wheels from sticking together when one is moved, and from being “overrotated or otherwise moved at improper times.” There had been earlier efforts in such machines to hold the wheels which were not to be moved, but the invention was for a double lock, primarily, a separate detent for each wheel, locked by the adjacent wheel, and secondarily, a bar common to all, which held the detents until the moment came when a number was to change. Then the detents upon the appropriate wheels were released, and the secondary lock was thrown off all. As soon as the change was completed, the detents and the secondary lock went into place again, holding the wheels against all disturbance. The opinion below gives a full and accurate description of the prior art and of the patent in suit. We agree with it except in so far as it extends the scope of the claims to cover the supposed infringement.
This was a device of the same general kind, with separate detents for each wheel, locking, however, only by means of springs, a method disclosed early. The defendant supplemented these by a bail bar which locked all the detents at once. It was fastened to arms at either side of the frame, which were free to swing with a limited movement, relieving the detents of the bail when a number was to change. When, however, the pawl frame which actuated the wheels was not in operative position, it hooked one of the arms of the bail against movement away from the detents, and so in turn locked the detents themselves. Thus it insured a locking of all the wheels as soon as the pawls had effected a change of number. It did not and could not, however, prevent "the rotation of wheels of higher order when a wheel of lower order was moved, because at that moment the bail was off all the detents and must be, if the pawl was to move any wheel at all. At that instant the defendant relied only upon the spring several to each detent. The same is true of the secondary device of the patent in suit and is indeed inherent in any lock which operates on all the detents simultaneously. Thus the primary purpose of the patent, to prevent turning when the wheels stuck, was not realized by the infringing device.
The secondary lock, the only common new part, was organized quite differently in the patented machine. Kern disclosed a rotating bar running along the base of all the detents and normally holding them positively in place by the heel of each, which eoacted with the periphery of the bar. (The primary lock was by a toe in the detent which eoacted with a disk on each wheel, not necessary to describe.) The rotation of this bar, through a mechanism actuated by the pawl frame quite different from that of the defendant, brought beneath the heels of the detents a cut-out segment at the moment when the numbers were to change. The detents were thus freed of the secondary lock; such of them as were then in phase for a change of number being also freed of the primary lock. The secondary lock was replaced by a turn of the bar, as soon as the pawl frame moved back after changing the number. In function the two machines are alike; in mechanical organization they are entirely unlike.
In determining the scope of claims the *462relevant comparison is between the prior art and the supposed infringement, not the disclosure. The claims speak in general terms and thus make the disclosure only an instance or example of its more or less abstract aspects. In some sense all claims are necessarily functional, and indeed the chief problem for courts is how far generalization is permissible. The prior art is a limit in any case beyond which they cannot go; their language is another, dependent first upon its proper interpretation, and next upon the latitude allowed under the doctrine of equivalents. Thus a reference may be relevant on the score of infringement which is immaterial to validity, as is well understood. Moreover, if the patentee must extend his claims to cover functions disclosed in his specifications in order to include the supposed infringement, prior art is a sufficient answer which discloses that same function, though the references be different in detail from the infringement, for they are as good to anticipate, as the patented disclosure is to extend the monopoly. The patentee, being obliged to resort to functional similarities between his disclosure and the infringement, must be content to accept the same measure as to the prior art.
In the case at bar we think that though Hart’s patent, No. 1,233,166, was a little different in organization from the defendant’s and was in a different kind of machine, it showed all that was common to it and the patent in suit, except in one feature, o which Bates, iNo. 484,389, supplied. Hart was for a computing typewriter, where the digit wheels must rotate more than one number; they might turn through their whole periphery, less one space. Hence there was a greater danger of overrun than in such devices as those in suit. Nevertheless, if that danger was met in the same way as the defendant met it, pro tanto it was an anticipation. Hart provided a series of detents, the ends of which ran over a ratchet at one side of each number wheel. The locking end of the detent was kept in contact with the ratchet by a spring at its other end, so that the locking end rode over the ratchet continually and tended to prevent an overrun so far as the strength of the spring would do so. In order, however, to avoid the resultant clicking, if the spring was made strong enough to act as a proper brake, Hart provided a positive cheek or lock at the far end of the detent, which came into play at the moment when the wheel came to its proper position. This was in the form of a bail bar held by two arms at the side of the machine which locked the several detents simultaneously by camming against their under side. The ■phase was momentary, only while the lever was held in place which set up all the wheels, and the function was only to prevent an overrun, though the detents could be made to lock as long as the lever was kept in extreme position, The device was substantially the same in structure as the defendant’s, except that the pawl frame and its action was very differently conceived. True, Hart’s machine was a combined typewriter and computing machine, and we understand that the totals were never printed. So far as the patent provided for holding the types against displacement by inking rollers it was not apposite. The types might indeed be held firm for any purpose by holding over the lever in its extreme position, but it would hardly be fair to take such an unsuggestod use as an anticipation of this function.
Kern’s supposititious third purpose — • holding the types against displacement by the inking rollers — is not mentioned in the specifications, or indeed even alluded to, unless it be in the words, “otherwise moved at improper times.” However, a patentee need not be aware of the advantages of his invention, and so far as this was an advance, the patent must be credited with it. Both the patent and the supposed infringement pos-. sessed it, because the secondary lock held the wheels during the inking phase. But it seems to us that Bates fills the hiatus so left open. This was for a printing head for consecutive numbering, though its organization was quite different from either Kern’s or the defendant’s device. The detents were within the wheels, as was also the pawl frame; but, while the mechanism is complicated and the specifications difficult to understand, it appears that as the types were moved, the pawls' by contact with the detents drove them into recesses which 'held the wheels against further movement. The detents thus held the wheels during the inking and also prevented any overrun, one of the disclosed purposes of the invention (page 1, lines 25-27). There was indeed neither bar nor bail interposed between the pawl-frame and the detents, as was the case with the devices now in court; the pawls bore directly upon the detents, and the method was so far dissimilar. But unless we are to say that the claims cover all intermediate devices though they exclude the same *463function -when the contact is direct, we cannot disregard the reference. It does not seem to us that we should extend the claims to cover operation by any intermediate means and stop there. Indeed, as we have shown, Hart would prevent this so far as concerned the overrun.
As in all such cases, it may be argued that it is improper to anticipate, or even to limit, claims by catching at this feature in one reference and at that in another. We agree; and we should apply that familiar doctrine in a case where the similarity of the disclosure and the infringement were in anything but function. Here they are similar only in that, except that the lock must be actuated by the pawl frame, an inevitable condition anyway, since the lock and frame must be in phase. The summation of functions, merely as such, when each has been disclosed separately, can hardly be a basis for invention. If the doctrine of aggregation has any scope at all, it is certainly applicable to such a situation. All that we can see that the defendant has borrowed from Kern is the general notion of a positive lock to avoid overrun, and to hold the types during inking. The second was present in Hart; both were in Bates; and the claims must be limited more specifically to the disclosure in order to stand. We think that they are not infringed, though we have no doubt that the patent is valid.
Decree reversed; bill dismissed.