after making the foregoing statement, delivered the opinion of the court.
Following the testimony of the expert, Bates, the prior' art in dogging devices may be summarized as follows: The Freeman patent of 1858 has a rotating carriage, with a dog near the periphery, which is forced inward to hold the bolt, and an arm, pivoted to the dog and to the frame of the carriage, which serves to operate the dog. A cam acts upon the arm, and forces the dog inward, causing it to bite the bolt, and another cam, acting upon an extension of the arm, throws the arm and dog outward to release the bolt. The Freeman patent of 1859 has a rotary carriage, a dog near the periphery, an arm attached to the dog, and bearing a friction roller, and two cams, one of which acts upon the roller to move the dog outward and to release the bolt, and the other to produce the opposite effect. In the patent of Kinney and Parker there is a rotary carriage, a dog, a bent arm pivotally connected to the dog, and a cam acting on the arm to release the dog from the bolt, the dog being near the middle of the carriage, and the cam acting upon the outer end of the arm, instead of the reverse, as in the Perkins device; but, if transposed, as they might be without making any change in them, they could be described in the language of the Perkins claims. The Everts patent has a rotary carriage, movable dog, bent arm connected with the outer end of the dog, and acted upon at its inner end by a cam to cause it to release the dog from the bolt. The Clark patent has a rotary carriage, a dog near its periphery, a spring arm attached to the dog, a friction roller on the arm, and two cams acting upon the roller, one of which withdraws the dog, and the other causes it to bite. The Palmer patent contains almost the exact mechanism of Perkins’ fourth claim in its essential details, namely, a rotary carriage, with a dog near its periphery, a bent arm connected at its outer end to the dog, extending inward, with a cam or incline on the frame, which acts upon the inner end of the arm, and a spring connected to the projections or abutments on the arm and the carriage, the spring and the central cam both causing the dog to bite and hold the bolt, and a second cam near the periphery of the carriage, acting upon the arm, to withdraw the dog and release the bolt. This is the entire combination of the fourth claim except the anti-friction roll, but such rolls had theretofore been common in shingle machines and in other kinds of machines. The Clark patent has a rotary carriage with peripheral dog, a bent arm to actuate the dog, and a cam near the center of the machine to act upon the arm. The O’Connor patent has a rotary carriage with a dog near its periphery, and an arm or lever pivoted to the carriage, connected at one end to the dog, and carrying an anti-friction roll at the other end. There is a rod connected to the arm or *531lever, and two abutments, one adjustable on tbe rod and tbe other fixed upon tbe carriage, and a spring, surrounding tbe rod between tbe abutments, serves to move tbe dog inward and cause it to bite tbe bolt. A cam on tbe frame, against wbicb tbe anti-friction roll bas a bearing when the carriage rotates, moves tbe dog outward to release tbe bolt. Tbe only changes from tbe device of O’Connor necessary to convert it into that of Perkins are suggested by tbe prior patents mentioned, and especially by tbe Palmer; that is, tbe substitution of a bent arm and a cam near tbe center of tbe machine for tbe pivoted straight arm or lever and'peripheral cam of the O’Connor design. Tbe O’Connor patent bas all tbe parts of tbe fifth and forty-fifth claims except that tbe portion of the arm wbicb extends inwardly is not quite long enough to reach past tbe stationary dog,—a circumstance wbicb does not affect its function, purpose, or operation, and therefore is not a material difference.
The foregoing propositions, and especially tbe conclusions supposed to result from them, are strenuously controverted by tbe opposing expert, Mr. Powers, who after pointing out particulars of difference between tbe devices of O’Connor and Perkins, says:
“I am aware that the witness Bates states that the missing elements not shown by the O’Connor patent may be found in some other patents in the record, but I do not understand that an anticipation can be created by taking up elements in different patents, and combining them for the occasion; but, on the other hand, in order to anticipate the claim of a patent, all of its elements, either identically or substantially, must be found in the same relation and combination with each other in some one patent or device.”
To tbe doctrine of selection be refused to subscribe, and for that reason failed to find tbe invention of Perkins in tbe patent of O’Con-nor. In matters of fact tbe entire testimony of tbe witnesses shows them to be in substantial accord. Their differences of opinion are explained by Mr. Powers’ mistaken understanding of tbe rule by which tbe patentability of combinations of old devices should be determined. That tbe mere bringing together, in a new combination, of old devices or elements, especially if they belong to tbe same art or to arts kindred to that to wbicb tbe combination belongs, does not constitute invention is well settled. “It is not enough that a thing shall be new, in tbe sense that in tbe shape or form in wbicb it is produced it shall not have been before known and that it shall be useful, but it must, under tbe constitution and tbe statute, amount to an invention or discovery.” Thompson v. Boisselier, 114 U. S. 1, 11, 5 Sup. Ct. 1042; Hill v. Wooster, 132 U. S. 693, 10 Sup. Ct. 228; Burt v. Evory, 133 U. S. 349, 10 Sup. Ct. 394; Pickering v. McCullough, 104 U. S. 310: Florsheim v. Schilling, 137 U. S. 64, 11 Sup. Ct. 20; Adams v. Stamping Co., 141 U. S. 539, 12 Sup. Ct. 66; Deere & Co. v. J. I. Case Plow Works, 9 U. S. App. 567, 6 C. C. A. 157, 56 Fed. 841. Tbe differences between tbe dogs described in the patent of Perkins and those of tbe earlier patents are differences of form and arrangement which produce nothing-new in function, result, or mode of operation. It is certainly not a matter of invention that a bent arm is operated by a centrally located cam instead of an outside ring, there being no essential difference in tbe result; or that one dog, rather than tbe other, is made *532movable; or that an arm is not radial, extending centrally down in a machine to operate npon an interior cam; or that levers extend over and are actuated outside of instead of inside the periphery of the carriage; or that a spring is shown without means of adjustment for greater or less tension, such means being well known; or that a radial arm extends outward a considerable distance beyond the periphery of the carriage, and has its outer dog not pivoted to, but integral with it; or that a radial arm does not extend over and beyond the inner dog; or that a dog is adapted to be released beyond the periphery of the carriage instead of by an interior cam; or that a rectangular frame is employed instead of an arm; or that a cam instead of a spring is used to push a dog into place. It is testified, and perhaps would be evident without proof, that of these various forms of construction and arrangement of - parts, while one may be better than another, the differences are not such as to affect the question of invention.
Of the claims for spalting devices, the special feature of the first —No. 26—is “two movable sections, side by side, and forming part of said tracks, adapted to be displaced from normal positions under the bolt.” In the O’Connor patent there is a single movable section in one of the two tracks, and it follows that “two movable sections, side by side,” can be regarded only as a duplication, unless their being adapted to be displaced from normal position under the bolt is a novel and patentable improvement. That cannot be conceded, because it is a mere matter of adjustment whether the displacement shall be in advance of or directly under the moving bolt. As examples of movable sections in parallel or concentric tracks, reference is made in the testimony of Mr. Bates to the store railway or cash carrier, as illustrated in the Holbrook patent, No. 282,320, the machine for sorting and loading lumber, shown in the Davies patent, No. 238,200, and to the well-known movable sections in railway tracks. If, therefore, there is patentability in this feature of the Perkins patent, it must be found in the other claims.
The twenty-seventh claim adds to the twenty-sixth hinged posts to support the movable sections and lever mechanism connected to the posts, whereby the sections may be swung radially in opposite directions, substantially as stated. The one section of the O’Con-nor patent is moved in a radial direction by means of a lever mechanism connected to it. It is not supported on hinged posts, but, as Bates testifies, hinged posts were old devices, of which harvester reels and buggy tops were examples, and therefore it would require no invention to mount the O’Connor movable section on hinged posts like the harvester reel or buggy top, so that both ends of it would move equally, instead of pivoting it at one end, so that one end moved more than the other. Swinging gates, if supposed to move in perpendicular instead of horizontal planes, afford a familiar illustration.
The additional features of the twenty-ninth claim are “a movable bar, outside the rotating carriage,” connected to the movable section of the way, and “a trip on the carriage,” adapted for the purpose stated. There are corresponding parts in the O’Connor machine, except that the operation is reversed, the lug or trip being *533made adjustable, and the bar stationary, instead of the bar adjustable and the lug stationary. “The function, operation, and purpose,” the expert says, “are the same, and the use of the one or the other is merely mechanical expediency.”
The thirtieth claim adds to the twenty-ninth “a stop on the frame in position to throw the trip out of operative position”; but that there was no novelty in that feature is clear. Stops and trips are “old and well-known devices for all sorts of purposes.”
The thirty-first claim differs from the twenty-ninth and thirtieth in that the block-supporting way is described as “consisting of two tracks,” with a section of each in advance of the saw made movable. This, at most, is mere duplication.
Without further pursuit of details, it is evident that the important feature of both the O’Connor and Perkins patents is the movable section or sections in the tracks or guideways. As between the two certainly, and, upon the evidence here adduced, as against the world, O’Connor is the pioneer, and the question is whether the Perkins patent shows a patentable improvement or advance in the art. If so, it must be found in the fact that his movable section or sections in the process of operation are withdrawn from the track, not in advance of the approaching block, but under it, at the instant of passage, so that the spalt, or remnant of the block, will drop squarely through the opening, and not be tilted or turned sidewise, as in the O’Connor machine. The invention is not to be looked for in the hinged posts, levers, lugs, trips, and cams, or combinations thereof, employed to move the sections. Those, as set forth, were mere expedients, open alike to the use of every inventor. The single movable section of O’Connor was pivoted at one end, and it is plain that large pieces were liable to be caught, and made the means of retarding or obstructing the operation of the machine; but it is not true upon the evidence that the invention of O’Connor was a failure. On the contrary, it was as compared with anything before it a great improvement, and, to say the least, was so far successful as to require that the invention be accorded full significance in the prior art. It is not invention to improve a defective contrivance by the exercise of merely mechanical skill, however great the success. It is clear in fact as well as in law that, a movable section having been introduced into one of the two rails, constituting the track of a shingle-machine, it required no invention to introduce a like section in the other rail, side by side with the first; and it is equally clear as matter of fact that it involved nothing patentable to make the section in O’Connor’s rail, singly or in duplicate, movable equally at both end's; and we have already seen that it was a mere matter of expediency to so adjust the movement that the sections should be withdrawn when immediately under, rather than in advance of, the block. We are therefore forced to the conclusion that these claims are invalid.
It remains to consider the forty-third claim, the essential feature of which is a wooden bearing block, seated in an oil-retaining trough. The specification, it is to be observed, says that these pieces are preferably blocks of wood cut to present the edge of the grain toward the wheel which rests on them; that the blocks are inserted in troughs, *534eacli of which has a lubricating cup attached to and connected with it; and that the cup may be filled with oil, which will pass into the trough around the bottom of the block, and will be carried up the pores of the block by capillary attraction, so that the surface of the block will be constantly lubricated. The cup, it will be noticed, is not mentioned in the claim, implying a trough of different construction from the one specified, to the extent, at least, that it should be without openings at the bottom to admit oil from the cup, and leaving it to be presumed, though unexplained, that there must be in the trough itself a space, not filled by the block, capable of use as an oil receptacle. The claim departs from the specification in the further particular that it does not require the block to be so placed as to present the end of the grain to the wheel to be supported. It is, therefore, not an essential feature of.the claim that the oil shall be carried to the surface of the block by force of capillary attraction. But with that feature eliminated, there remains no reasonable pretense of patentability. It is conceded that the O’Connor patent shows bearing blocks used for the same purpose as those of the Perkins machine, but without stating of what substance they should be made. The use of wood for that purpose was certainly free to all, and it required no invention to provide a metal receptacle, socket, or trough, in which the block of wood should be placed and held in position, and, when in place, to oil it; and if by accident or design there was room in the receptacle for the accumulation of oil and an accumulation happened and capillary attraction was brought into effective play, it could hardly be said to have been invention, because, the operation of capillary attraction being well known, the result, to the extent possible in the particular kind of wood used, was just what was to have been and by a man of ordinary understanding would have been expected to happen. We are of opinion that none of the claims here brought into question contain patentable novelty. The decree below is, therefore, reversed.