L. HAND, Circuit Judge.
These three suits were tried and decided together, and the appeals were taken and argued at the same time. They concern three patents issued to Joseph Sachs, and all cover devices for housing electrical switches and fuses, and their wiring. The first was to keep inclosed against interference by unauthorized persons everything except the fuses which lead from an electrical meter to the fixtures which consume the current. These will at times blow out, and it is convenient that the customer shall ha.ve access to them, so that he may put in new ones. Everything else ought to be beyond his power to reach, both for his own safety, and because he might otherwise steal the current by by-passing the meter. The wires leading to the meter, their switches and the fuses which are often put there as well, are called the “service-side” of the installation; the wires leading from it, including the switches and fuses always put on this side, are called the “load-side.”
The first patent disclosed a metal box which contained the “service-side” wires with knife-blade switches, which could be actuated from outside the box, fuses carried by the blades, and holes in the blades, expressly provided for use in testing the meter, of which more hereafter. -• These were housed in the upper part of the box which the meter sur
mounted. A little more than midway down the box was a solid horizontal partition through which the “load-side” wires passed to the “load-side” switches, also knifeblades, carrying fuses on their blades, on which there were also terminal testing holes. 'When the front was open, the box thus presented two compartments unconnected, except where the “load-side” wires passed through the horizontal partition. Over the whole front the patentee placed a metal cover to be locked in place, which would keep out any one who did not have the key or break the seal, exeept that opposite the “load-side” fuses there were holes in the cover of substantially the same bore as the fuses, through which they passed and which made them available to the consumer in case of a blow-out. lie could not, however, reach anything but these, and this was one object of the invention.
Meters must be tested and during the test the consumer must still be supplied with current. This is done by attaching a wire, called a “jumper,” from a “service” to a “load-side” terminal, thus by-passing the meter. The test meter is then put into the circuit; for this it is necessary that there be a gap, which may be on the “load-side,” and perhaps also on the “service-side” (Swallow & McCoy, 927,384), though this the plaintiff denies. As wo have said, the patent discloses switches in both “service” and “load-side,” whose movable blades carry fuses; in these blades there are the holes to which the “jumper” can be plugged, after which both switches can be opened without interruption of the circuit, with both fuses still in place for purposes of safety. The testing meter is, normally at any rate, inserted in the gap caused by opening the “load-side” switch. The disclosure thus provides specific means for testing.
Of the prior art we need consider only the patent to Wurdack, 1,041,398, his catalogue, and the so-called “Baltimore Use.” Taking up the last first, we may dismiss it, since it was not a single housing containing the “load-side” fuses with the “load-side” switch and the “service-side” fuses. Again, it had no “service-side” switch at all, and while this may not have been necessary, the patent prescribed it and the defendant has seen fit to use it. If there was anything of merit in the invention, it consisted in assembling all the switches, fuses and wires into one compact housing. This had been tried several times before, and when Sachs succeeded, as he did, ir producing a unitary device, portable as such, with all the necessary details, which gained a very extraordinary popularity, it seems to us that the testimony of the past overcomes any inherent doubts we might have, based upon the apparent ease and inevitability of the changes he made. We should have to disregard settled doctrines to treat the installation at Baltimore as an anticipation.
Wurdaek’s patent is closer, and with his catalogue, the nearest in the record. We have held catalogues to be printed publications within the statute, when their distribution is adequately proved, Jockmus v. Levitón, 28 F.(2d) 812; and while the evidence here leaves something to be desired, we are content to accept the catalogue as a reference. It disclosed a single housing with a cover which could be taken off, exposing the whole inside. In the cover was a door which the consumer might open, and which made accessible the “load-side” fuses, eight in number. It however also showed some of the terminals, and was not, like Sachs’, confined to the fuses alone. This of itself would not have been important, as appears in our consideration of the third patent, but it was possible for a person, expert in such matters, to make attachment between a “load-side” and a “service-side” terminal and so by-pass the meter, stealing the current. This would have been a troublesome job, no doubt, and often unsuccessful, but the possibility was there, whose avoidance may well have had something to do with the success of the patent in suit. It is true that it would have been a simple thing to put a horizontal partition above the “load-side” fuses, which would have prevented access to the “service-side” terminals. However, we cannot judge an invention by the ease with which the prior art may be changed, once the idea is conceived. It seems to us that the absence of the partition assumes importance if one considers the relative success of the two disclosures. Wur-daek’s housing was not accepted; Sachs’ has gone into immense use.
Be this as it may, there was no provision for testing the meter in either Wurdack’s patent or his catalogue. That it was possible to make a test with his device is probably true. Terminals could be found to which the “jumper” could be attached, and a gap could be made by the “service-side” switch in which a testing meter could be inserted. But there was no “load-side” switch for that purpose, and while we are not clear why it could not be inserted in the “service-side,” the patent disclosed it in the other form, and the de-' fendant has followed the disclosure. Wur-
daek’s invention did not therefore have any means for testing, but left this to the ingenuity of the trade. It does not appear that this was so obvious as to make the specific provision for it a mere redundancy, or a matter of convenience. Even so, no one had thought to do it before. It seems to us therefore that the earlier efforts which did not go so far, and the surprising success of the patent, ought to overcome any doubts arising from the apparent simplicity of the changes, at least against literal adoption such as the defendant’s.
There remains the question of the disclaimers. Pending the suit the plaintiff filed a disclaimer to claims seven and ten, so as to exelude from them all but devices in which the “load-side” switch contained a fuse which should alone be accessible without opening the door. The figures showed this and the specifications described it in this way (page 1, lines 55-57; page 2, lines 53-58; page 3, lines 100-103; page 4, lines 53-62; lines 108-114). We doubt that the disclaimers were necessary at all; claims seven and ten in saying that the “service-side unit” was to be “inaccessible while permitting access to the fuse of the ‘load-side’ unit” could by implication mean that the rest of the “load-side” was inaccessible. It is quite true that more than the fuses was shown to be accessible in some of the figures, but some, especially, of the language cited (e. g., lines 53-58; page 3, lines 100-103; page 4, lines 111-114), seems to indicate that one form was intended to prevent a customer from reaching anything but the fuse. When the- specifications themselves distinguish between alternative forms, the patentee may disclaim one. Strause Gas Iron Co. v. Crane Co., 235 F. 126 (C. C. A. 2); Permutit Co. v. Harvey Laundry Co., 279 F. 713 (C. C. A. 2); Permutit Co. v. Wadham, 13 F.(2d) 454 (C. C. A. 6); Campbell Metal Window Corp. v. Pomeroy (D. C.) 300 F. 872.
The disclaimer to claim four adds another feature which appears to mean no more than that the “load-side” switch shall carry a fuse which is not disturbed by the movement of the blade. This was the only form shown in the drawings, and the switch was so described. It was disclosed as like the “service-side” switch (page 5, lines 31-36), and the “service-side” switches were to carry fuses on .their blades (page 4, lines 127-128). The specifications in the passage cited do indeed say that other switches may be used, but the disclaimer goes no further than to limit the claim to the disclosure, abandoning any implied and undeseribed scope. We think the patent valid and infringed and we affirm the decree below.
The next patent was for certain details of the mechanism within a housing for a switch and its fuses. It concerned- the connection between a switch blade and a bail bar to throw the blade in and out, the bail being in turn moved by a handle outside the box. The bail bar was connected with the blade by a lirtk of fibrous or other insulating material, but in such wise that the blade might have some side play, so as to accommodate itself for sure entrance into its clip to make the circuit. The specifications described the link as rigidly fixed to the blade with an open hook at its other end to engage the bail loosely, but they defined the invention vaguely as covering either a loóse connection between link and bar, a loose connection between link and blade, or a loose connection at both ends of the link (page 2, lines 117-129). The main purpose seems to have been to avoid the familiar cross-head between several blades or links from such blades, which moved the blades together, and which allowed no relative movement between them. This apparently interfered, or was thought to interfere, with the proper seating of a blade if it did not exactly register with its clips. Eor this reason the blades were held independently and had each its own motion laterally.
There were a number of earlier efforts in the art of like kind. Hope’s British patent, 4,397, of 1908, showed a slotted link pivoted on the blade, and by the slot loosely coupled to the handle which threw the switch. In Railing’s British patent, 26,009, of 1910, the handle moved the blade at its own pivot, being attached to an axled block to which the blade was also attached, the block being loosely mounted upon the axle. In Wirt, 801,288, two switch blades were loosely connected by a cross-bar near their swinging end which allowed them some accommodation relatively, unlike a cross-head. In Levin, 1,370,656, the blade was loosely engaged near its pivot by a block which was attached to the handle. Each blade thus got independent lateral play to accommodate itself to its clips. Horton, 1,407,416, adds nothing to the rest.
Of the claims in suit, 20 and 21 prescribed that the link should be fixed to the blade, and the disclaimer by the words, “rigidly and fixedly secured,” added nothing except to claim fourteen. However, it was valid as to that claim, since the specifications had clearly disclosed alternative constructions, and the situation is like that as to the
first patent. In so far, however, as the disclaimer limited the claims to a bail bar which went across the blades, and to a link with an open end, we may assume, though we do not decide, that it was invalid, on the theory that these, though disclosed, are nowhere claimed in the specifications as a constituent part of the invention. However, wo do not see that these limitations were in any sense necessary to escape the prior art. A disclaimer is to abandon some part of the invention of which the patentee is not “the first and original inventor.” If he has claimed originally too much, so that the claims are invalid under the prior art, the part disclaimed must be clearly separate in the body of the specifications; if he wishes to recast the whole, he must go to a reissue. But there is no objection to his limiting a valid patent as his fears may dictate; that does not make valid what would be invalid without it. It is a timorous retreat from positions which he could have successfully maintained. We can therefore disregard all of the disclaimer except so much as was necessary to limit .claim fourteen, and so much was valid.
The distinction between a link fixed to the blade and what the art had known is not indeed very great, but it was a hitherto unknown variant in a field where as we have seen- there had been much experimenting. Whether it was useful or not we need not inquire; the defendant chooses to employ it. The trade has accepted it generously, and we see no reason why the inventor should not keep others out of the narrow confines to which he has withdrawn. The decree is reversed as to the second patent.
The third patent is for a housing for a fusible switch, the invention this time being in an automatic lock to keep the fuses inaccessible while the switch blades bridge their gap, and to allow the consumer to reach them, when they do not. Over the fuse opening in the cover, into which the fuses project, a smaller cover is hinged which must be closed to allow an outside handle that moves the inner bail bar to be swung into closing position. This handle crosses the whole box on the outside, sweeping across the fuse outlet. As the closing movement of the handle is from the hinged side of the cover, the cover when open blocks the passage of the handle, but when closed, allows it to pass above it, locking it in place and closing the switch at the same time. The fuse block completely fills the opening in the main cover.
The defendant’s box is of different construction ; the cover which gives access to the fuses slides in grooves on the inside of the front, and an inside bail bar, which moves the switches, stops its opening when the blades are swung into place. When the blades open, the bail bar is disengaged, and the consumer is free to slide the cover down and open the fuses for examination or replacement. In its manufacture the plaintiff has adopted this form, rather than the disclosure of the patent.
It seems to us that, if construed in the sense needed to cover the defendant’s devices, the claims are invalid, though taken verbally they would serve. Krics’ patent, 1,224,880, was for a box in which the fuses were in one compartment and the switches in another, separated by a solid partition. The switch handle locked the hinged cover over the fuses when the blades closed the circuits, but unlocked it when they did not. This is the defendant’s operation. The plaintiff argues that the reference is not good, because the fuse block did not fill the opening through which the fuses projected, and this is true; but it does not seem to us invention to shape the main cover so as to expose only the fuses, or to make the fuse block fill the space. Cole, 1,135,358, showed a somewhat similar device in which the cover opened and closed the switch. It was quite different in mechanism, and we are aware that it is in general not enough to pick here and there from the art and so anticipate an invention. On the other hand the plaintiff is trying to push the disclosure beyond what Sachs described. Before giving it the needed scope, we must see how far the generic ideas were already known. Everything was in the art except the detail we have mentioned, and while it is always embarrassing to know when the new feature should prevail, there must come an end to what we can call invention merely because it is new. Exhibit 20, one of the supposed infringements, shows how the main cover itself can be depressed so as to cover all but the fuse outlets. It appears to us to make no difference whether the result is accomplished in that way, or by building the base block so that it fills the opening. These are alternatives optional with any designer.
Jessen’s patent, 1,209,445, answers verbally better than Kries, barring the fact that it is the edge of the fuse block which “cooperates” with’ the cover. However, it was of quite another design and seems to us in
structure less like the defendant’s switches than Kries. Hill, 1,373,656, is very like Kries except that it was possible to reach one terminal of the switch, which might be the live end. Again, we can see no invention in bringing forward the fuse block, so that the fuses fill the space, or in making the opening small enough to expose only them. These .references at any rate struck close around the target, and while that is ordinarily rather proof of the aim of him who makes the hit, it may be relevant to limit the extent of his victory. We do not hold the patent invalid, but it seems to us too narrow to cover the defendant’s boxes. The decree upon the third patent is affirmed.
We have not dealt with the defendant’s argument that the first patent is for an “aggregation.” Frankly, we are unable to attach a definite meaning to that word. The notion that the parts of an invention must co-operate is certainly very persisting in the patent law, and it must correspond to some underlying idea. So far as it means that the whole complex claimed must be a unit in use, each part of which shall be necessary to the ■common result, we can understand it. So far as it rests upon an implied reference to mechanics, that is, that each part must give or take a strain, it seems to us a false lead. The test is more practical than that, because inventions are to answer human needs, and the elements may be mechanically inert. The cooperation of the means necessary to create an invention is to be measured by the purpose to be fulfilled, not by the interaction of the parts. Each factor must indeed be a condition to that result, but the whole may be a mere assemblage; the co-operation between them all may be no more than their necessary presence in a unit which shall answer a single purpose. Therefore, we can find little advantage in a discussion of what is or what is not an “aggregation.” In patents, as in other branches of the law, the question is of the interests involved; inventions depend upon whether more was required to fill the need than the routine ingenuity of the ordinary craftsman. SucE a, standard is no more of' a will-o’-the-wisp than others which the law adopts, reasonable care, reasonable notice and the like; the effort is to fix that standard by recourse to average propensities, dispositions and capacities. Any attempt to define it in general terms has always proved illusory; it is best to abandon it.
The decree upon the first patent is affirmed ; those upon the second are reversed; those upon the third are affirmed.
On Petition for Rehearing.
PER CURIAM.
On the merits we see nothing to add to what we have already said except as to Horton’s patent, 1,407,416, which we dismissed by saying merely that it added nothing to the others. In February, 1915, Sachs had completed drawings of the invention which later appeared in patent No. 1,292,081, and these he put in proof. They bear the date, were properly authenticated, and are unimpeachable as evidence. They antedate Horton’s application, which was filed on March 24, 1915, though Sqehs’ model was completed later. Indeed, his drawing of December, 1914, would probably have been enough alone. Unless Sachs unreasonably delayed the prosecution of his application, this proof was enough. Automatic, etc., Co. v. Pneumatic Scale Corporation, 166 F. 288 (C. C. A. 1). That is a matter which must be pleaded by special notice (U. S. Code Tit. 35, § 69 [35 USCA § 69]), and the answers do not contain anything of the sort. It has not been argued before us, and, so far as we can see, it was not mooted at the trial. We decline to consider it.
However, we did not dispose of the question of costs. Three suits were before us; No. 1911, against the Hartford Electric Supply Company on patent 1,292,081, and patent 1,301,175; No. 1912, against the Service Electrical Supply Company, on patent, 1,-292.081, and patent 1,294,176; No. 1913, against the Union Light Company on patent 1.292.081, and patent 1,294,176. Pending the suits the plaintiff filed disclaimers in the District Court as to certain claims in 1,301,175 and 1,292,081, but not as to 1,294,176. The plaintiff succeeded in the_ District Court in No. 1911 as to 1,301,175, but failed as to 1.292.081, and was awarded no costs. Both sides appealed, and we have affirmed the decree as to 1,301,175, and reversed it as to 1.292.081, By the weight of authority, the plaintiff should be allowed its costs in this court. Kahn v. Starrels, 136 F. 597 (C. C. A. 3); Johnson v. Foos Mfg. Co., 141 F. 73, 89 (C. C. A. 6); Excelsior, etc., Co. v. Williamson, etc., Co., 269 F. 614, 619 (C. C. A. 6); Bankers’ Utilities Co. v. Pac. Bank, 22 F.(2d) 680 (C. C. A. 9); Excelsior, etc., Co. v. Meyer, 36 F.(2d) 447, 450 (C. C. A. 7) . The Third Circuit decided to the contrary in Novelty Glass Co. v. Brookfield, 172 F. 221, but without notice of its earlier decision in Kahn v. Starrels, supra. The situation may be different, when the disclaimer is made of a claim first held invalid upon ap
peal. Liquid Carbonic Co. v. Gilchrist Co., 253 F. 54, 58, 59 (C. C. A. 7).
This seems to us in principle correct, because the plaintiff, having put itself in the District Court in a position to demand a decree upon both patents, was forced to prosecute its appeal as to patent 1,292,081, and defend the defendant’s appeal as to patent 1,301,175, in order to secure and defend its rights. The claims in suit were free from invalidity after the disclaimer, though invalid when the suit was brought, because of the matter disclaimed. There is no reason why the mistake so corrected should any longer deprive it of re-imbursement for the necessary costs of maintaining its position; and the statute does not demand so much. "We allow to it the costs in this court; it will take no costs in the District Court.
In No. 1912 the plaintiff succeeded on the appeal as to patent 1,292,081, but failed as to patent 1,294,176. There should therefore be no costs of this appeal. The defendant should have costs in the District Court; this because the plaintiff could not in any event have costs in that court as to patent 1,292,081, and because the defendant has succeeded as to patent 1,294,176. The plaintiff’s success as to patent 1,292,081 should not offset the defendant’s as to patent 1,294,176, in view of its commencement of the suit as to patent 1,292,081, at a time when strictly it was not entitled to sue upon that patent; the claims in suit being invalid without disclaimer.
The same disposition should be made of No. 1913.
Settle mandate in accordance with the foregoing.