DENISON, Circuit Judge.
The defendants appealed from an interlocutory decree for an injunction and accounting in the usual form, entered in an infringement suit brought by Ferdinand W. Starr upon three patents issued to him: The first being No. 533,095, dated January 29, 1895, for a drawing instrument; the second, No. 683,809, dated October 1, 1901, for a cutting instrument; and the third, No. 766,158, dated July 26, 1904, for a machine for cutting curves. All three instruments were either specially intended for, or have found their chief utility in, marking or cutting ellipses, and complainant’s commercial machines have their main application in cutting mats or glass for framing photographs. The suit involved claim 3 of the first patent, all of the 21 claims of the second patent, except claims 13 and 15, and claims 1, 2, 8, 9, 10, 11, 12, 13, 14, and 15 of the third patent. Some of these last claims were withdrawn so that as to this patent, the case was left to rest on claims 9, 10, 12, and 13. The patents must he considered separately.
[266]*266The First Patent.
[1] The particular merit of the commercial devices of each party-lies in their adaptability to describe either a circle or any desired ellipse within the size capacity of the instrument. It is characteristic of an ellipse, in process of making, that its center is constantly moving along the line of its major axis, and that this path of motion is at right angles to the .minor axis; and it follows that, if the ellipse is to be described by a point at the moving end of a revolving arm, a point at the inner end of this arm and an intermediate point must be constantly moving in paths at right angles to. each other. The elementary instrument for mechanically producing an ellipse was called a “trammel.” It is illustrated by the following sketch taken from the Century Dictionary:
AA and BB are slotted bars at right angles to each other, and represent, one, the major, and one, the minor, axis of an ellipse. C is a revolving bar carrying, at its outer extremity, a pencil, as at 1. 0 and 3 are pins fixed in the bar C .and passing through into, and traveling in, respectively, the grooves in AA and BB. Obviously, if the bar C revolves upon pin 0 while that pin is stationary, pencil 1 will describe a circle, while, if during the revolution, pin 0 travels longitudinally in AA, the described curve will be irregular, becoming a perfect ellipse, if the revolving and the longitudinal motions are maintained in proper, constant relation. This maintenance is compelled by the pin 3, traveling in the groove of BB. The pin-0 stands for longitudinal motion of the center and pin 3 stands for revolving motion of the arm. Any movement by either pin compels a corresponding movement by the other. Obviously, also, by making these pins 0 and 3 adjustable on the revolving bar and changing their relative positions with reference to the scribing point 1, any desired degree of elliptical curve can be obtained. Speaking in general terms, these- two pins, 0 and 3, are two shafts, eccentric to each other, co-operating in their revolution to produce the described curve. The resulting diameter will be varied as the distance between point 0 and the pencil 1 is increased or diminished; the resulting degree- of curve is controlled by changing the distance between pins 0 and 3; in other words, varying the mutual eccentricity of the two shafts. The longer axis can be changed at will, from vertical to horizontal, by changing, the position of the pencil so that the order of the points 1, 0, 3 will be 3, 0, 1. If the two pins are superimposed, i. e., if the two shafts are made concentric, the revolution of the bar will produce a circle.
The utility of a mechanical device for drawing ellipses was well understood long before Starr’s first application. Williams & Joslin, in 1859, by patent No. 22,910, showed a device which the Patent Office called an ellipsograph, and by which they said “curves and.figures approximating in form to ovals may be drawn with great facility and in [267]*267a perfect manner.” Their device had two co-operating, revolving shafts eccentric to each other. One of these was reciprocated in a slideway corresponding to one trammel bar, but they dispensed with a right-angled slideway and attempted to give to the other shaft its requisite right-angular motion by carrying it on the free end of a bar pivoted at the other end to the stationary frame. The resulting curve would not be a true ellipse, although a close approximation. •
On July 6, 1875, by patent No. 165,385, Toulmin illustrated and described an ellipsograh. He provided for the right-angled motion by carrying one of the shafts upon longitudinal guides and carrying the other upon the free end of a lever pivoted to the frame; but he observed, as perhaps Williams & Joslin did not, that a simple, swinging arm would carry this center, not on a right line, but .on the arc of a circle, and therefore, as Toulmin says, “this does not give a true ellipse.” Accordingly, he provided two oppositely extending arms pivoted at their outer ends to the frame and at their inner ends to the opposite ends of a crossbar, the central point of which stands for the other shaft, and so, although the opposite ends of this crossbar move in arcs, the central point moves in a right line, perpendicular to the longitudinal guides, and the desired result is reached.
In 1876, Root took out a patent, No. 181,725, on what he- called a “trammel” and the Patent Office called an “ellipsograph.” He departed from the pivoted, vibrating arm scheme employed by Williams & Joslin and by Toulmin, and returned, typically, to the trammel of the above sketch. Pie has two fixed, right-angled slideways carrying longitudinal sliding bars. On the inner end of each bar is pivoted one of the revolving shafts. The two slideways are in different vertical planes, and' between these planes a horizontal bar connects the two shafts, being fixed to one and sliding through the other. At the top of the upper shaft, a crank is attached whereby there is a co-operating revolution of the two shafts, and, when the two shafts are not concentric, any point in the connecting bar will describe an ellipse. As the lower guideway would prevent the revolution of a pencil depending from this bar, a similar bar is attached to the lower shaft below the guide way, and this lower bar carries the adjustable pencil holder.
Again, in 1883, Plottinger, by patent No. 288,810, described a machine for marking ellipses. He used a stationary knife and caused the table to turn so that the contact point described an ellipse. Pie produced this motion in the table by a typical trammel bar and sliding pin arrangement on the bottom of the table. Rawson also patented an ellipsograph, in 1890, patent No. 422,252. His machine is essentially that of Root; but, for insuring longitudinal motion in one direction, he uses a collar sliding on a rod, and in the other direction, a pin sliding in a guideway. He also provides peculiar mechanism in connection with a supplemental pencil-carrying arm.
Into this state of the art, Starr came by his application filed in 1893. He calls his invention a “drawing instrument.” He discards the two right-angled, sliding guideways of Root, Rawson, and the others, and returns to the pivoted arm idea of Williams & Joslin and of Toulmin. He changes and apparently improves this idea and calls. [268]*268this vibrating arm the “guide bar H.” At its outer end, instead of being merely pivoted to the frame, the pivot passes through a longitudinal slot in the bar, so that the bar may not only swing but may also move lengthways upon the pivot. The inner end of the bar carries one of the eccentric shafts. At an intermediate point where the bar swings over the surface of the frame, the bar is provided with a depending stud, and the upper surface of the frame is provided with a curved slot in which this stud must travel. This curved slot is not an arc of a circle, but is parabolic, and the net result of the guiding by this slot and the permitted longitudinal slip at the pivoted end is that the shaft carried at the inner end moves back and forth exactly at right angles to the longitudinal slideway which carries the other shaft. This construction is shown in figure 1 of his drawing here reproduced :
It follows from this history and description that Starr’s meritorious invention is found in the peculiar construction and movement of this guide bar H. Starr himself, in his testimony in this case, points out that:
“The friction generated by carrying tbe shaft along a straight bar [as Root and Rawson bad done] was so great that it was found to be impractical, * * * and this deficiency in a drawing instrument of that character led to the useful and effective modification set forth' in this claim 3 * * * almost entirely eliminating the friction incident to moving said shaft along a straight bar at right-angles to the lower shaft."
The defendants.are.said to make a device in which the right-angled travel of the two shafts is compelled by two right-angled rods upon each of which a sliding collar reciprocates, and each of these collars carries one of the eccentric shafts. There is no pivoted bar of any kind. There is no vibrating arm. The same ultimate result — the right-angled travel — is accomplished; but the device is distinctly of the Root-Rawson, rather than of the Toulmin-Starr, class.
The question of infringement arises under Starr’s third claim, which reads thus:
“3. The combination of the sliding bar D, the shaft L, with its integral guide arms, the sliding shaft I, the arm II, the spiral spring £ to elevate the [269]*269several parts, the operating and depressing arm K, and the pencil holder, to produce continuous or broken lines, substantially as described.”
Defendants employ this complete combination, except the “arm H”; but, in order to make out infringement, the “arm H” of the claim must be interpreted as covering any means for producing right-angled, transverse motion in the second shaft. This would be carrying the doctrine of equivalents very far. We are not prepared to say that under the principle of Hoyt v. Horne, 145 U. S. 302, 12 Sup. Ct. 922, 36 L. Ed. 713, McSherry v. Dowagiac (C. C. A. 6) 101 Fed. 716, 41 C. C. A. 627, and Metallic Extraction Co. v. Brown (C. C. A. 8) 104 Fed. 345, 353, 43 C. C. A. 568, it might not be carried even to that extent, if Starr’s substantial invention had been in some other feature of the machine, and if the “arm H” had been one of two or more previously well-known methods of accomplishing one part of the operation ; but whether or not such extreme liberality of construction could be permitted in the supposed case, it cannot be allowed where the patentee has deliberately confined himself, by his claim, to a structure containing the peculiar element which was the main feature of his invention, and where the alleged infringing structure contains, not that peculiar element, but the old substitute therefor which the inventor had discarded. To permit the claim, under such circumstances, to be so liberally interpreted, would be to make it of even more flexibility than indicated by Justice Bradley’s familiar figure of the “nose of wax” (White v. Dunbar, 119 U. S. 47, 51, 7 Sup. Ct. 72, 30 L. Ed. 303) and would be violating the settled rule (Cimiotti v. American Co., 198 U. S. 399, 25 Sup. Ct. 697, 49 L. Ed. 1100; Coupe v. Royer, 155 U. S. 565, 576, 15 Sup. Ct. 199, 39 L. Ed. 263; Brown v. Stillwell [C. C. A. 6] 57 Fed. 731, 739, 6 C. C. A. 528).
[2j So far as this record shows, Starr was the first to make a construction whereby the pencil could be depressed by the hand of the operator turning the crank, and perhaps he was entitled to a claim protecting broadly the capacity for vertical motion in the shafts, whereby the tool would vertically follow the operator’s hand, .and in combination with any curve-directing means; but he made no such claim; that issue was not presented to the Patent Office, and is not raised by this record. Patentees “cannot expect the courts to wade through the history of the art and spell out what they might have claimed, but did not claim.” Keystone v. Phoenix Iron Co., 95 U. S. 274, 24 L. Ed. 344; Macomber’s Eixed Eaw of Patents, § 226.
We do hot doubt the validity of Starr’s first patent, but we cannot find infringement.
[3] In reaching this result, we give no force to the use of the reference letter H. or to the presence in the claim of the phrase “substantially as described.” The specification, claim, and drawing are a unit. When the claim refers to “the arm,” it means the arm shown in the drawing and described by the specification, or the equivalent, and “the arm H” can mean no more and no less. Whatever parts are named in the claim are of necessity intended to be named with reference to the specification and drawings. The reference cannot be made narrower by saying “as described” nor broader by saying “substantially.” [270]*270The words1 “substantially as described” do not create this necessity for construction by the entire patent; they are only a formula of recognition of the rule, and the cases, like Pope v. Gormully, 144 U. S. 248, 253, 12 Sup. Ct. 641, 36 L. Ed. 423, and Fox v. Perkins (C. C. A. 6) 52 Fed. 205, 214, 3 C. C. A. 32, which make reference to the phrase, cannot mean anything more. See sections 214, 221, 234, Ma-comber. The meaning of “equivalent” is a question of fact to be determined by the language of a claim, in connection with the specification, the drawing, the state of the art, and the history of the application. Attempts to make arbitrary rules, as of law, for determining questions of fact lead to artificiality and seldom help.
The Second Patent.
[4] In his second patent, applied, for in May, 1900, Starr illustrates and describes over again the instrument of his first patent, but with some comparatively slight additions and changes which adapted the device to cutting out, as well as drawing, curved forms, saying of his invention that it was “especially adapted for cutting mats for pictures and similar outline work.” In the place of a pencil-holding socket carried on the lower revolving arm, he substituted a socket holding the swiveling stem of a cutting tool, and he provided such a tool which had its knife carried at the lower end of the stem on a lateral arm or bend so that the knife would' trail behind the axis of the stem, and which had its knife laterally inclined so as to cut a bevel, and its cutting edge .inclined rearwardly so as to give a draw cut. These features, in various combinations, either as constituting by themselves a bevel-cutting tool or as constituting, in combination with curve-directing mechanism like the first patent, a bevel cutter for curved shapes, are the subject of the first ten.claims.
,We are unable to find any invention in these features, alone or in cpmbination. Commercial success apparently followed the thought that a cutting tool could be substituted for the pencil of Starr’s first patent, and in this thought only is the substantial merit of these changes. This thought, in an environment consisting only of these features of the ellipsograph art which we have so far recited, might or might not have been sufficient to support invention; but the thought was not new with Starr; it had occurred to others and had been applied by others repeatedly. McAdams, by patent No. 179,039, in 1876/ showed a device for marking or cutting out ovals, having a tool-head “containing a pencil or knife,” arid he repeatedly refers to “marking or cutting.” Breach, in 1879, in patent No. 219,615, described how patterns could be either “cut put from, or marked on, a sheet of leather,” and he constantly refers to his carrier or tool-head as a “marker or cutter.” Hottinger, in his above-mentioned patent of 1883, named his machine as one for “describing or cutting ellipses, said that “in cutting or marking an ellipse, as the case may be, the-point of the tool or marker is set,” etc., and made further reference to “the tool or marker” and “the cutting or marking tool.” Cote, in patent No. 469,775, described a device for “outlining and proportioning boot and shoe patterns, and drawing the same on paper or cutting the [271]*271same from paper or other suitable material,” and he said “in one of the sockets is placed a pencil or knife-holding spindle.” He continually refers to the “pencil or knife.” King patented an ellipsograph, in 1894, No. 517,522. This was of the typical trammel type, like the Century Dictionary cut, and the drawing showed a pencil. King said:
“In some eases, If, is desirable to form ellipses from stiff paper, pasteboard, and the like, and to this end, I contemplate providing the instrument with a suitable cutter which may be carried by the block, i, in the same maimer as the pencil.”
The record shows other instances, but these are enough to demonstrate that, in 1900, it was familiar knowledge that, in such association, marker and knife were interchangeable, and that there was no inventive novelty in substituting a cutter for the pencil of Starr’s ellipsograph.
His next provision was that the tool carrier should have a socket, and that the cutting tool should have a stem so adapted that the stem would swivel in the socket; in other words, he saw that his cutter must swivel, and he provided means therefor. This idea of swiveling would seem rather obvious, as a trailing cutter -could not otherwise follow the curve; but the cutter with its swivel stem and the tool-holding socket to receive the stem had been shown by McAdams, Breach, and Cote, above described, as well as by several others in the record not necessary to mention.
If, then, in our search for novelty, we fall hack on the idea that the knife or cutter of such a tool should be carried on a lateral arm or bend of the stem so that, as the tool swivels, the cutting edge will trail behind the axis of the stem, we find this indicated in Hartford, No. 207,866, of 1878, who invented a tool for cutting out patterns, and specified that the. “cutting edge of the knife lay on one side and not across the axis of the swivel,” and fully disclosed in Heideuhain, No. 365,129, of 1887 (which was a marker and not a cutter); also, in Cote, above mentioned, who specially described the construction and said that the purpose was to have the knife “automatically maintain a position with its cutting edge to the front in whatever direction it may be moved.” The provision that the cutting edge of the knife shall be inclined ,rearwardly is the universal provision for insuring a draw cut, whether the operation is by hand or by machine. It is shown in many of the references.
Nothing remains as an independent element of supposed novelty, excepting that the knife is set laterally so as to give a bevel cut. This again approaches, if it does not touch, the obvious expedient; but it is shown by the record to have been common. Hottinger provided for his cutting tool such adjustability that it “may be set vertically or at any desired inclination.” Stuparich, by patent No. 572,320, in 1896, and for the express purpose of forming oval, beveled openings in photograph mats, provided for setting his knife laterally so as to cut any desired bevel. So, Durkel, in patent No. 660,211, on application filed in February, 1900, had shown “a device for cutting out the centers of picture mats, so as to leave the opening in the mat of circular or elliptical form, the inside edge of the mat being cut beveled,” and had [272]*272provided' therefor a knife inclined laterally as well as trailing and inclined to the rear. So, also, Gooding, No. 593,537, of 1897, had a device for cutting boot or shoe soles “and other articles of curvilinear outline from sheet material,” and provided a swivel cutting tool with knife inclined laterally and rearwardly. It follows that not even in this feature of the laterally inclined knife in such connection was there anything new.
[5] It must be admitted that no one structure shows all these features in combination, and, in our conclusion that these ten claims are invalid, we do not intend to impinge upon the well-understood rule that there may be patentable invention in selecting well-known elements in other machines, and uniting them in a new combination to- get a new result. Dowagiac v. Superior Drill Co. (C. C. A. 6) 115 Fed. 886, 53 C. C. A. 36. We are considering not so much the question of anticipation as that of invention or of double use. The most that can be said in Starr’s favor is that he took a previously existing combination and substituted, for its vertical nutting knife, the well-known bevel-cutting knife, or that he took another known combination and substituted, for its straight tool, the well-known curved or trailing cutting tool, or another existing combination and substituted the common swivel connection for a rigid connection. From no one of these points of view is it possible to say that Starr did anything more than would have, been done by any skilled mechanic to whom the primary idea had occurred. The fact that no one had before combined all these features, which fact, while not controlling, is often persuasive to show invention, cannot prevail against a clear case of the mere adoption of common expedients in adapting an existing machine to a new use, and in a case where the thought of adaptation is not new. Bullock v. Electric Co. (C. C. A. 6) 149 Fed. 409, 420, 79 C. C. A. 229.
Claims 11, 19, and 20 involve the same point. They call for the machine of the first patent, described in broad and general terms, with the addition of means for holding or clamping the cardboard or other material from which the machine cuts out the selected forms. It is apparent that, when the only thing desired was to draw a mark on paper, the paper would be easily held in position; but, when thick and heavy paper or other material was to be cut, there would be, strong tendency to push the material out of position, and this tendency must be suitably resisted. • Such resistance or clamping, to be most effective, should be applied as closely as possible to the cutting point. Starr’s machine, as commercially built, weighed perhaps 50 pounds, and-his plan for holding, the material consisted in hinging his machine at the rear to the supporting table, so that the machine, lifted by its front part, could be turned up, and in attaching to the front, supporting legs of the machine horizontally adjustable extensions or frames which, when the machine was lowered, would make contact with the material on the worktable, and carry from that contact point a part or all of the weight of the machine. These extensions could be adjusted so as to rest upon the material as closely as desired to the point of cut, and in this way the weight of the machine could be shifted to, and carried at, different positions on the material, and in the selected posi[273]*273tion where it would be most efficient as a workholder. This combination is expressed in the eleventh claim as follows:
“11. In a cutting machine, the combination of a base, a frame pivotally connected at its rear end thereto, means carried by said frame for producing an oval or a circular outline, a horizontally adjustable workholder adjust-ably connected with the front legs of said frame and adapted to be raised simultaneously with the frame and lowered upon the work to hold it while cutting the same."
Claims 19 and 20 involve practically the same elements with immaterial additions. Whether there is any patentable distinction between any of the three is not now important.
The validity of claim 11 is attacked because it is said to be for an obvious expedient lacking invention, and because it is said to cover a mere aggregation rather than a combination. The subject-matter involved is simple, and it is not improbable that similar workholding’ devices were in use; but there is no evidence on this subject. The situation is the same as though these points were raised on demurrer. There is obvious utility in avoiding independent clamps and shifting the weight of the machine to the spot where it's weight will be most efficient; and, in view of this utility and the adoption of the idea by defendant, we cannot say that the construction involves no invention.
[6] The defense of aggregation is plausible, but not good. It is true that the cutting mechanism operates just as it would if the work was held by any other kind of clamp, and that the clamping means holds the work whether or not the cutting mechanism is in operation, and with only the same ultimate result — immobility—as if it was held by any other means; but, while the machine is in operation, the same primary element, the force of gravity in the machine, enables the knife to be advanced against the resistance of the material, and causes the material at that very point to resist' against the push of the knife. The action of one part of the entire structure modifies and affects the action of the other part, and there is, during the active period, more than that mere aggregation which defeats a patent.
Defendants infringe these claims. They have perhaps made improvements, but they have appropriated the idea and clearly have the pivotal rear frame connection and the workholder in horizontally adjustable connection with the front legs of the frame.
[7] Claims 12 and 14 differentiate from the first patent only by providing that the pencil or tool holding head shall not only be adjustable longitudinally on its carrying arm, as the pencil holder was, but shall also be adjustable'laterally. No merit is pointed out in this additional adjustability, and, unless under exceptional conditions, there can be no invention iii making a tool adjustable on its carrier in four directions instead of in two. Smyth v. Sheridan (C. C. A. 2) 149 Fed. 208, 211, 79 C. C. A. 166.
[1] Claims 16 and 1-7 depend for patentability upon the-addition to the first patent oí a centering rod located centrally in the rotating head, and for the purpose, of indicating the center of the figure to be drawn. Whether this centering rod was capable, in May. 1900, of imparting patentable novelty to a claim otherwise barren thereof is doubtful on [274]*274this record; but both of these claims contain, as an element, and re- . ferring to the “arm H” of the first patent, a call for a “pivoted, oscillating member.” This cannot be read upon defendant’s structure, which had no pivoted, oscillating member, but has only the collar reciprocating on the rod. Our discussion of infringement under the first patent, referring to Hoyt v. Horne, is pertinent here. If the limitation, apparently imported into the claim by the choice of this language, could ever be neutralized by applying the doctrine of equivalents, it must be in aid of a clear invention dominating and characterizing the claim, rather than in case of small improvements upon an existing patent, which is the selected and designated basis on which the improvements must rest.
[4] Claims 18 and 21 seem to belong in the same group as the first ten. We see nothing in them to differentiate from the first patent, except, in one case, the provision that the tool be swiveled, and, in the other, the double adjustability of the tool and the trailing form of the cutter.
The Third Patent.
[8] By Starr’s third patent, applied for March 31, 1902, he covered certain improvements in his device as applied to cutting material like glass, or metal, wheré a, knife would not be operative. He relies on claims 9, 10, 12, and 13. Claim 9, typical of the group, is as follows:
“9. The combination of tbe vertical tool-socket and means for impelling tbe same along predetermined curved lines, the vertical tool-stem swiveled and free to rotate in said socket, and the rotating cutter carried on a bent or offset extension of said swiveled tool-stem.”
The part of the device to which these claims are directed is illustrated in figure 6, which is here- reproduced:
In his second patent, No. 683,809, the drawing contained figure 7. also reproduced above, and in his specifications he says:
“Figure 7 shows the tool in which the shank and arm are formed integral, and a hardened steel wheel inserted in the latter, for which, if desired, a diamond may be substituted for glass cutting.”
[275]*275He further says1.
“The machine described is particularly adapted for cutting oval and similar mats, and also for cutting similarly shaped glass for picture framing and other purposes.”
The claims of the second patent are confined to an instrument having the cutting blade laterally inclined so as to make a beveled cut, and the device of the third patent lacks this inclination; hence, the device would not respond to the claims of the second patent, and the case is not, in the strictest sense, one of double patenting. However, not only would there be no invention in making the cutting disc vertical instead of inclined, hut figure 7 of the second patent apparently <, shows this vertical position; indeed, the inclined position would seem inappropriate for glass cutting.
So far as concerns claims 9, 10, 12, and 13 of the third patent, the second patent contained full disclosure, and having been issued in October, 1901, without reservation of any kind which could preserve the subject for patenting on an application filed several months later, everything shown, and which might have been, but was not, claimed, was abandoned to the public: and these claims are invalid. Underwood v. Gerber, 149 U. S. 224, 13 Sup. Ct. 854, 37 L. Ed. 710; Ma-comber, § 12.1
[9] The bill must be dismissed, as to the first and third patents. According to the practice indicated in Herman v. Youngstown, 191 Fed. 579, 588, 112 C. C. A. 185, Starr may have 30 days, after the filing of the mandate below, in which to show that he has duly abandoned the claims of the second patent held in this opinion to be invalid, and thereupon he may have the usual decree for injunction and accounting as to claims 11, 19, and 20. Appellants will recover the costs of this court, and one-half of their costs in the court below. The partial cosis in the court below which Starr might have if he had prevailed upon one wholly valid patent, cannot be, in this casd, allowed. O’Reilly v. Morse, 15 How. 62, 121, 14 L. Ed. 601. The matter of Houser’s individual liability to injunction, damages, or costs, and the matter of any increase of damages, will be remitted to the District Court to be determined upon any accounting that may be had, or in such other manner as that court shall direct.