operates to press down that portion of the lower stretch with which it contacts so as to bring it into close frictional contact with the surface to be polished. This head reciprocates by sliding on a guidebar, working outward from the center each way pressing the belt to its work. The guidebar of defendant is not connected with a rod arranged to be operated vertically by means of a lever. The guidebar is fixed having no vertical motion. In order, however, to secure the relative motion of head and carriage by means of which the belt is pinched between head and working surface, the carriage is arranged so that it may, by means of a lever, be raised or lowered, thus increasing or decreasing pressure at the working surface, precisely as it is increased or decreased in the device of the patent. This reversal of motion between two parts, one of which remains fixed while the other advances or recedes, is a well-recognized form of equivalency and will not avoid infringement unless there is something in the record which requires the claims to be so narrowly construed that even this slight change will be sufficient to take a device, in other respects identical, out of the scope of the claim.

We do not find anything in the record which constrains to so narrow a construction. As originally filed, claims 1 and 3 called merely for "a head" to press the belt to its work. They were rejected and three prior patents were cited: Huseby, No. 637,121, November 14, 1899; Totman, No. 348,177, August 24, 1886; Morris, No. 29,293, July 24, 1860. In each of these there is a head located within the belt. In Huseby this head is spring-pressed downward; the pressure of the springs may be varied by turning nuts, and when once so regulated remains constant until the nuts are again turned; there is no regulation of pressure while the machine is working by means of a head movable vertically by means of a lever. In Totman there is a head within the belt, which has such a vertical movement as will adjust it to variations in the position of the surface being operated upon, such as may be caused by differences in the thickness of panel, but has no means of regulating pressure of the head at the will of the operator while the machine is running. In Morris an elastic roller is located within the polishing belt and is provided with means for raising and lowering it from or towards the lower stretch of the belt. But neither in Morris nor in the other two patents is there any reciprocating head. In consequence, as soon as Gale modified his claims by limiting them to a reciprocating head, his patent was issued.

Two other patents are relied upon by defendant. In Hess No. 412,616 the patentee provides a hand implement, entirely disconnected from any machine which is operated by hand. The operator takes this implement and places it within the loop of a belt, and then by grasping the two handles he uses it as a means for holding the abrasive surface of the belt against the work piece. There is nothing in this which calls for any modification of the claims of the patent in suit. The only remaining patent to be considered is Clark No. 570,866. The machine shown in this patent provides a pair of tongs in which a tube, such as a portion of a bicycle frame, may be gripped and held without being rotated. A polishing belt is wrapped partially around the tube thus held by the tongs, and as said belt is driven the guide pulleys, by which the belt is held against the tube, are moved by oscillating the plate upon which they are mounted, so that the position of the belt is changed and it is caused to come in contact with every portion of the stationary tube which is being polished. The guide pulleys operate against the inner side of the polishing belt, and by thus pulling against it they cause another part of the belt to press against the work. But we cannot concur with the District Judge in the conclusion that the Clark patent requires a construction of the patent in suit which would confine it to a device where there is a "guide bar on which the head travels vertically to press the belt to the material operated upon." The Clark patent has no head which presses the belt to its work, pinching the belt between the head and the work; nor is there any reciprocation of the head. The belt no doubt yields somewhat when the work-piece is pressed against it as did the springpressed head in Huseby; but there is no relative movement between the holder for the work-piece and the oscillating plate which carries the guide pulleys. Certainly if Clark's was a junior patent, it could on no reasonable theory be held to infringe the Gale device or that of defendant. Clark's device may very well necessitate a construction of claims 1 and 3 which will restrict them to a machine which operates by pinching the polishing belt between the head and the working face, but it is in precisely that way that defendant's device operates. We are of the opinion that such device infringes claims 1 and 3.

[2] In the Gale patent the reciprocating head may be set to have a predetermined length of throw, but that length cannot be varied during operation; to do so the machine must be stopped and readjustment effected. It is desirable to have the head adjustable during operation. Thus, if a round table-top is being polished, a short throw only is required when work begins at the side of the table-top; a longer throw is required as the polishing belt reaches the center, and thereafter the length of throw may be again reduced. Yarnell improved the Gale machine by so arranging the head and connected parts that the throw, or stroke, of the head could be regulated by the workman while the machine was running and doing work. The seventh claim of his patent reads as follows:

"7. The combination in a rubbing and polishing machine, with a polishing belt, of a reciprocating head arranged to engage said belt and press it to its work, suitable mechanism for reciprocating said head and means for regulating the throw of the head without stopping the machine substantially as described."

Yarnell was the first to adopt such means in machines of the type under consideration. We concur with Judge Hazel in the conclusion that this claim is valid and is infringed by the device of defendant. It seems unnecessary to add anything to his discussion of the question.

The decree is reversed, with costs of this appeal to complainant, and cause remanded, with instructions to decree for complainant in conformity with the views expressed in this opinion, with costs.

SUNDH ELECTRIC CO. V. INTERBOROUGH RAPID TRANSIT CO. (Circuit Court of Appeals, Second Circuit. July 8, 1912.)

No. 201.

1. PATENTS (§ 127*)-VALIDITY-PRIOR PATENT OR PUBLICATION-PRIOR IN

VENTION.

A later patent was not in the prior art, so as to invalidate a patent in suit as a prior patent or publication, although it was granted on an earlier application; but it may be proved by competent evidence that the later patentee was the original and first inventor by being the first to both conceive the invention and reduce it to practice, and such prior invention may be shown by his prior application, where there is no evidence to carry the invention of either patentee back of the date of his application, provided the invention of the patent in suit and of such application is the same.

[Ed. Note.-For other cases, see Patents, Cent. Dig. §§ 179, 180; Dec. Dig. § 127.*]

2. PATENTS (§ 328*) -VALIDITY-PRIOR INVENTION-ELECTRIC CONTROLLER.

The Sundh patent, No. 733,564, for an electric controller, held void on the ground that the patentee was not the original and first inventor of the device patented.

Appeal from the Circuit Court of the United States for the Southern District of New York.

Suit in equity by the Sundh Electric Company against the Interborough Rapid Transit Company. Decree for complainant, and defendant appeals. Reversed.

said levers and cams being constructed and timed so that said levers 54, 55, and 56 are successively actuated by said cams to close circuit and the lever 62 to open circuit."

Arthur B. Seibold and W. Clyde Jones, for appellant.
William B. Whitney, for appellee.

Before LACOMBE, COXE, and NOYES, Circuit Judges.

LACOMBE, Circuit Judge (after stating the facts as above). An electric controller is a device for regulating the amount of current delivered to a machine which such current operates. When, by closing a circuit the current is turned on, it is desirable that its entire force should not become operative at once at the working point. Therefore in the circuit there are placed one or more resistance coils, which reduce such force until they are cut out by the closing or opening of other circuits, usually in succession. A familiar instance is the handoperated controller on a trolley car where the driver moves a lever over successive contact-points, thereby gradually increasing or reducing current. It is often desirable to have the controller automatic. Thus in the case of a water-tank on the roof of a building which is fed by an electrically operated pump, there will be a device which will close circuit and set the pump going when the water sinks to a determined level, and which will open circuit and stop the pump when the water reaches another determined level. The current thus automatically turned on or off is regulated by a controller, also automatically operated at the proper times. The various parts which are combined to produce these results, independent circuits, solenoids, contact-terminals, shafts, gearing, cams, levers and what not are old and well known in the art.

Judge Hazel quoted from a description of the patented device a sufficiently full specification of its parts and their operation, which reads as follows:

"Briefly stated, the throw of a hand switch closes the circuit through the upper winding of a double solenoid, which, when energized, draws up its core and closes the main switch in the motor circuit and thereby starts the motor, with all the resistance in the circuit. At the same time the lower winding of the solenoid is energized, and, drawing up its core against the retarding action of a sash-pot, by means of a rack cut along the back of the core and meshing with a pinion on a cam-shaft, causes the rotation of the shaft, whose cams act successively on the arms of a series of pivoted lever switches, making 'butt' contact with a series of fixed contact-terminals, to close these switches one after another, and thereby short-circuits and cuts out step by step the sections of resistance (and with the last section the series field winding of the motor), and gradually brings the motor from rest up to a condition of full speed. Finally, the last cam on the shaft to operate throws open a normally closed lever switch in the circuit of the lower solenoid winding and cuts out that winding; but its core is held in its elevated position, at the top of its upward movement, by the upper solenoid winding, and in turn holds the camshaft so that its cam will maintain the switches in position, the resistance switches closed and the solenoid circuit switch is open. When it is desired to stop the motor the hand-switch is opened; whereupon the core of the lower solenoid winding drops by gravity, rotating the cam-shaft back to its initial position and allowing a spring to close the solenoid circuit switch and the resistance switches to fall back to open position by gravity, and then the core of the upper solenoid winding drops."

The conclusion we have reached upon the whole case makes it unnecessary for us to discuss the details of the structure. We may state, however, that besides the particular combination or combinations covered by the claims above cited, the patentee also showed the same combined with a "magnetic snap action," so-called in the briefs, which prevented sparking between levers and circuit-terminals, which snap-action is not covered by these claims, is not employed by defendant and need not be discussed. It is unnecessary also to review the prior art. Judge Hazel's opinion may be consulted for a sufficiently full résumé of its disclosures. He held that none of the prior patents were anticipatory, in which conclusion we concur. He stated the crucial question in the case as follows:

"Did it involve invention at the date of the Sundh application to gear the electro magnetic engine in the form of a solenoid, having a core and dash-pot, to the cam shaft of the lever to make butt contacts with the cooperating fixed contact (circuit) terminals in such a way as to start the motor and move the levers successively?"

In his conclusion that this question should be answered in the affirmative and in the reasoning by which he reached such conclusion we also concur.

[1] Among the patents introduced by defendant was one to Ihlder, No. 742,031, which, as was said by the Circuit Court, "so closely resembles the Sundh patent in suit that the supposition arises that both Ihlder and Sundh were engaged at the same time in solving the problem and each invention performs substantially the same function." If either of these devices were in the prior art, there would certainly be no invention in devising the other. It was held, however, that this patent was not in the prior art. Such holding was correct. The chronological sequence of events is as follows:

June 10, 1902, Ihlder application filed.
March 16, 1903, Sundh application filed.
July 14, 1903, Sundh patent issued.
October 20, 1903, Ihlder patent issued.

No testimony carried the date of either invention back of the date of filing application. The nature of the defense of prior art is succinctly expressed in section 4920, U. S. Revised Statutes, subdivision third:

"That it [the device of the patent] has been patented or described in some printed publication prior to his [the patentee's] invention or discovery thereof, or more than two years prior to his application for a patent therefor."

Ihlder's device was certainly not patented before Sundh's invention, which was completed by reduction to practice on the day the latter filed his application. Nor was Ihlder's application, although it completed his invention by reducing it to practice on the day such application was filed, a "publication" within the meaning of the statute, because it remained unpublished in the secret records of the Patent Office until after Sundh's application had been filed. The Ihlder patent, therefore, was no part of the prior art.