In the operation of internal combustion automobile engines, the engine pistons suck into the cylinders successive charges of -an. explosive, mixture -.'of- air and-liquid fuel-delivered through a *847carburetor to tbe intake manifold under tbe control of a throttle valve. When the engine is running at maximum speed the throttle valve is practically wide open, the mixture passing to the engine cylinders is at approximately normal atmospheric pressure, and no appreciable rarifieation or vacuum exists in the manifold. As the throttle valve is closed, the suction passage through the manifold is restricted, and since the intake stroke of the engine pistons and the capacity of the cylinders remain the same, the mixture between the throttle valve and the cylinders becomes rarified and a partial vacuum is created in the manifold. The engine usually is not stopped by the closing of the throttle valve but receives a minimum supply of explosive mixture ordinarily known as the idling mixture, sufficient to keep the engine turning over, so that it may be accelerated as its power for propulsive purposes is again desired. When the engine is so turning idly with the throttle valve closed, or almost closed, the vacuum in the suction passage is greatest, and this partial vacuum may be termed a byproduct of the slowing down of engine operation, since it is produced only when the engine is not operated at maximum speed and power.

The Dickson patent sought to utilize this by-product of available force to apply the brakes of the vehicle. On a base structure of the ordinary automobile chassis and engine, the specifications show a pipe, 25, running from a point in the manifold above the throttle valve to a suitable brake cylinder, 16, one end of which is connected with the atmosphere, and the 'other or elosed end to this pipe. When the engine throttle is elosed to the idling position, the air in the closed portion of the brake cylinder, 16, is exhausted into the manifold through the pipe, 25, and the atmospheric pressure on the other, or open, end of the brake cylinder causes the piston to move toward the elosed end. The piston is connected to the usual brake rod and applies the brakes. An extraordinary amount of static force on the brake bands is attained: the suction in the manifold is about 10 pounds per square inch; in a brake cylinder 6 inches in diameter this is increased to about 300 pounds; and the movement of the piston is sufficient to permit a leverage of 48 to 1. The application or release of the brakes under the Dickson patent is achieved by means of a three-way valve, 26, inserted in the pipe 25 between the manifold and the brake cylinder. It is, evident that Dickson understood that the introduction of any considerable amount of unfuelized air into the mixture in the manifold, through pipe 25, would tend to stall the engine. Accordingly, the three-way valve, 26, was so arranged as to permit either the suction pressure or the atmosphere to he introduced into the elosed end of the brake cylinder, but to , prevent the atmosphere from entering the manifold- at any time. In this fashion only the relatively slight amount of air present in. the elosed end of the brake cylinder could be exhausted into the manifold and this air could be readily fuelized by the liquid fuel adhering to the sides of the manifold.

The three-way valve is controlled by a foot-button, 29, which is kept in closed position by a coil spring, 30. The specifications and accompanying drawing also show the operation of the three-way valve by the hand throttle lever on the steering wheel by means of a connecting rod, 31, attached to this hand throttle linkage. Manual operation of the brake by means of the usual foot brake lever is also possible by means of a lost motion slot, 19, at the lower end of this foot lever. Infringement of claims 1, 2, 3, 5, and 6 of the Dickson patent was found by the District Court which also held that the specifications disclosed over the prior art an invention which was operable.

The Root patent1 relates to an improvement of the three-way valve indicated by Dickson. In the Dickson device the brakes could not be held, after having been slightly applied, unless the three-way valve was returned a short distance toward the open position, thus cutting off both the suction and the air from the brake cylinder, and thereby holding the piston in the position to which it had been moved. This required that the foot button or hand throttle lever be slightly released by the operator every time he desired to hold the brake. Root’s patent eliminated this additional and somewhat delicate movement by providing a valve which was automatically elosed by a movement of the piston and which held the vacuum until the suction was again introduced or air admitted to the brake cylinder. This result was achieved by using a follow-up element, 21, which was actuated by the piston itself, through a lever 22, so that when the suction was introduced and the piston moved, it automatically moved the follow-up slide to cover the suction aperture and hold the brake. Any additional movement of the foot button or hand lever (which are not shown by Root but which are adaptable to his device) would repeat this operation until the maximum braking effect had been obtained. The defense of noninfringe*849ment of the Root patent was sustained by the lower court.

*848

*849Defendant’s brake, which is manufactured by the Westing-house Air Brake Company under licenses for the use of United States patent, No. 1,539,166, issued May 26, 1925, and patent'No. 1,603,750, issued October 19, 1926, both to Albert Dewandre of Belgium, consists of a brake cylinder similarly connected to a point above the engine throttle valve, containing a piston which actuates the brake rod. Valve means for the control of the suction and air to be admitted to the closed end of the cylinder are provided, and manual operation of the brake, in the event of a failure of suction, is provided by means of a collapsible chain connection between the brake rod and the piston.

Defendant contends that the Dickson patent is invalid because of anticipation in the prior art and because it discloses an inoperable, impractical paper patent. He also urges that the Root patent is not infringed because defendant’s brake appliance does not contain a follow-up element, but achieves the same function as Root by the repeated opening and dosing of the suction valve.

1. In our judgment, the Dickson patent was not anticipated in the prior art. It is essentially a combination patent, and while each of the elements induded in it were to be found in the prior art, the combination in the manner specified for the indicated purr pose'was new, and thus constituted a patentable invention. Leeds & Catlin Co. v. Victor Talking Mach. Co., 213 U. S. 325, 29 S. Ct. 503, 53 L. Ed. 816; In re Coykendall, 58 App. D. C. 280, 29 F.(2d) 868; Craft-Stone, Inc., v. Zenitherm Co., Inc., 22 F.(2d) 401 (C. C. A. 3d). Analyzed, the invention consisted of the following elements: means for the utilization of the latent suction power present in the manifold when the engine throttle was partially closed, a brake cylinder operated by such suction and atmospheric pressure, a separate pipe connection between such cylinder and the manifold above the engine throttle, and valve means for admitting suction or air to the closed end of the brake cylinder without admitting air to the manifold.

The bulk of the early patents and publications introduced by defendant indicated no more than that it was well known that the suction present in the manifold was a latent source of power in an automobile and that it might be utilized to operate various auxiliary devices. British patent No. 5291 to Royce for improvements in lubricating system for internal combustion engines, accepted December 23,1909, shows a device, connected to the manifold of a gas engine, for controlling the admission of oil for the lubrication of the engine. British patent No. 18,145, to Smith, accepted May 5, 1909, shows the utilization of the manifold suction for blowing a horn, although it is extremely doubtful whether this device was ever operable. The Tice article in “Motor” magazine for June, 1911 (pp. 582-587), illustrates the well-known vacuum tank fuel feeding device which employed the manifold suction to raise fuel from a main tank placed in the rear of an automobile to a smaller tank above the carburetor. United States patent No. 984,-032, to Seager, and United States patent No. 983,994, to Harrington, show devices embodying substantially the same principle. The article in “La Vie Automobile” for January, 1904 (pp. 588-592), indicates another example of the use of manifold suction, here to control the feeding of fuel to the engine and thus to control its speed. All of the foregoing devices, in our opinion, indicate no more than that manifold suction was a known source of power; they do not suggest its use for braking purposes.

British patent No. 5,281, to Dare, for an improved brake for motor cars, accepted January 12, 1905, however, does show the use of manifold suction for braking. This patent discloses a device for operating the brakes of a vehicle driven by an internal combustion engine, which may be operated either by the pressure of the exhaust of such engine or by the manifold suction, or both combined. Mg. 2 shows a brake cylinder a containing a piston connected to the brake rod, a pipe, j, running from this cylinder to the engine or exhaust, and a valve f1 in the cylinder. The difference between the Dare device, where suction is used, and Dickson, is found in the fact that in the Dare disclosure the brake cylinder is interposed between the engine cylinders and the carburetor, whereas in Dickson the brake cylinder is connected in a side line, or a branch taken off from the main fuel feed line going from the carburetor to the engine. In Dare, defendant’s expert admits that everything passing through the throttle and carburetor passes through the brake cylinder. Thus, if the brake valve were closed, the engine would necessarily stall. Dickson’s contribution was a separate branch pipe containing a valve through which the brake cylinder might be exhausted, without affeeting the mixture admitted to the engine. The Dare patent discloses no such arrangement, and despite Defendant’s Exhibit M, which was a model of *850Dare showing an embodiment of this feature, we are of the opinion that Dare did not anticipate Dickson.

United States patent No. 840,193, issued January 1, 1907, to George W. Barlow, also shows a brake cylinder, open at one end, containing a piston operating on the brake rods, and connected by means of a three-way valve to a source of suction. The source of suction shown by Barlow consists of an injector inserted in the exhaust line, past which the exhaust gases rush and create a vacuum. This injector is connected through a check valve to an intermediate vacuum tank which is employed to store the vacuum. The Barlow device clearly does not anticipate; it does not indicate the use of the manifold suction, which is an engine by-product, but requires the use of an injector in the exhaust, which would tend to create back pressure and rob the engine of power. It requires the use of a large, intermediate storage tank; and, since the amount of exhaust gas is negligible when the engine throttle is closed, several successive applications of the brake while the car is running slowly would exhaust this tank and make further application impossible, without speeding the engine and wasting fupl; in short, in Dickson the maximum effective vacuum is attained when the throttle is closed, in Barlow the vacuum reaches its maximum with the throttle open and minimum with the throttle closed.

Defendant urges, however, that, since the Dare patent and auxiliary devices utilizing the manifold suction were known at the time of the Barlow patent, it required but simple mechanical skill to combine Barlow and Dare to achieve Dickson’s result. This argument cannot be accepted. As has been pointed out, Dare did not disclose the use of a separate pipe from the cylinder to the manifold, and the auxiliary devices utilized but a negligible portion of the available suction power. The Dare and Barlow patents operate on two basically different principles, Dare on the suction present in the direct intake line, and Barlow on the exhaust requiring a vacuum storage chamber. Dickson dropped the storage chamber, which was clearly impracticable for an automobile, showed a separate branch line from the manifold, and inserted his three-way valve in this line. It would reduce patent protection almost to a nullity if an infringer could, in the light of a subsequent disclosure, comb the prior art and piece together portions of earlier patents, while dropping other parts, and thereby invalidate a new combination of old elements. See Webster Loom Co. v. Higgins, 105 U. S. 580, 591, 26 L. Ed. 1177; Kryptok Co. v. Stead Lens Co. (D. C.) 207 F. 85, 95, and eases cited.

2. Defendant’s contention that the Dickson disclosure is inoperable has somewhat more weight. The claim rests upon the purported showing that, under'the arrangement disclosed, the vacuum brake would be applied every time the hand throttle lever were closed to idling position, and unless the throttle were in that position, braking by means of the foot button 29 would be impossible because of the absence of suction in the manifold. Both experts admit that if the brakes were applied every time the hand throttle is closed to idling position, slow coasting, engine braking, or travel in traffic would be well-nigh impossible. Under such circumstances, the device would be practically useless for automobiles, and substantially a failure so far as any commercial value in the art is concerned. Therefore any construction which would yield this result would be invalid as inoperable. The chief dispute is whether Dickson’s specifications and drawing necessarily show such construction. Defendant’s claim that it does would appear to be valid unless (a) the so-called Bentley scheme (Exhibit 35) for avoiding the difficulty, or some form of it, was either known at the time of the patent or is an exercise of merely mechanical skill, and is reasonably indicated in the specifications and claims, or (b) link 31, which connects the hand throttle linkage to the three-way valve, may be disregarded, and, within the specifications and claims, eliminated to permit the operation of the brake solely by the foot button 29.

(a) Plaintiff’s expert, Bentley, avoided the obvious difficulty by attaching a lost motion device to the base of the throttle lever IS, together with a stop screw, so that link 12, or its equivalent, could be moved to apply the brake without affecting the throttle lever 15. This highly ingenious device also required that the throttle linkage be pivoted differently, and that link 11 and link 31 'be attached to opposite ends of the pivoted arm, so that when the hand lever 9 was turned through a part of its are it would operate valve 15 without causing a movement of link 31; and a continuation of the movement of lever 9 through the remainder of its are past the idling position would actuate link 31 through the slip connection, without causing a movement of link 12 and the valve 15. It is agreed that this device rendered the Dickson patent wholly operable; the sole question is whether it is disclosed by the patent.

While a patentee is not bound to set forth *851in detail what was usual practice in the art (Webster Loom Co. v. Higgins, 105 U. S. 586, 26 L. Ed. 1177), there is only slight evidence that a device like Bentley’s was known to the art. Moreover, while we might agree that the linkage used is sufficiently simple to constitute merely mechanical adaptation of what Dickson shows, the wording of the specifications in the patent itself negatives any such scheme. After describing the general method of operation, the specifications continue (page 1, line 69): “The end of the rod 31 is provided with a slotted way 33 to permit the operation of the valve 26 through the operating rod 29 without affecting the movement of the throttle way 15.” This clearly indicates that but for this slotted way 33 in link 31, any movement of the foot button 29 would be reflected in the throttle linkage and in the throttle valve 15. Obviously, this would be possible only if rod 31 were definitely fixed to the throttle linkage, thus negativing the slip connection which Bentley incorporated. This theory is easily proved by the fact that Bentley’s device is completely operable without the slotted way 33, which Dickson believed to be necessary; that is, in Bentley’s device any movement of rod 31 (assuming that the hand lever 9 were also free to move) would be taken up by the slip connection at the base of throttle lever 15. In view of this statement in the specifications, Bentley’s device cannot be read into the Dickson patent. „

(b) We are, however, in agreement with the district court in holding that link 31 may be eliminated, and that operation by means of the foot button 29 is reasonably indicated as an alternative construction in the Dickson patent. While the specifications are somewhat poorly drawn, they do sufficiently indicate this mode of operation. Following a description of his brake cylinder, suction pipe, and three-way valve, Dickson states (page 1, line 60) :

“To the arm 28 is hinged a manually operable rod 29, whereby the operator may be able to open or close the valve and thus control the passage of suction through the pipe. The spring 30 returns, the valve to a closed position, admitting air to the cylinder 16 which in turn releases the brake. Through the rod 31 the arm 27 is connected with the throttle valve operating means at 32. The end of the rod 31 is provided with a slotted way 33 to permit the operation of the valve 26 through the operating rod 29 without affecting the movement of the throttle valve 15. Either of these valve-operating means may he used to the exclusion of the other.”

The difficulty here is that we cannot .with absolute certainty say to what the phrase “valve-operating means,” in the last sentence, is referable. It might mean that the engine throttle and brake valve may be operated separately, thus meeting defendant’s objection at the outset, or else that the hand and foot control of the brake valve might be used independently of each other. We are inclined to believe that the second meaning, namely, separate operation of the brake valve either by the foot button or hand control, is indicated.

Presumptively, a patent is operable, and the burden is on the defendant infringer to show that the device disclosed is impracticable. Western Electric Co. v. La Rue, 139 U. S. 601, 11 S. Ct. 670, 35 L. Ed. 294; Walker, Patents, § 85. Cf. Electrical Engineers’ Equipment Co. v. Champion Switch Co., 23 F.(2d) 600, 603 (C. C. A. 2d). In the case at bar, Dickson clearly disclosed the principle which became the basis of the modern vacuum automobile power brake, namely, the separate line from the manifold to the brake cylinder in which a form of three-way valve is placed. Historically, both his assignee and defendant utilized this invention by confining brake operation to the foot lever, and since we find sufficient indication in the remainder of the specifications and claims that Dickson meant to describe this separate mode of operation, we must ascribe operability-to the device and validity to the patent. On page 2, line 9, the patentee states: “The primary function of the valve 26 (when the foot operation is used) is to provide for applying the brake only when wanted, and not every time the engine feed is cut off.” This clearly indicates that separable operation by the foot button was contemplated. Moreover, the distinction is further shown by claims 4 and 7, not in suit; they are specific as to the alternative construction. Claim 4 specifies “mechanical connections between the engine throttle valve and the three-way valve”; while claim 7 mentions “manually operable means whereby the throttle and valve may be simultaneously operated.” The claims in suit contain merely the statement that a “manually operable three way valve” or “manually operable means” are to be provided. In our judgment, this is sufficient basis for the District Court’s finding that within the specifications and claims, rod 31 might be disregarded as, in fact, it was in the commercial development of the Dickson pat*852ent. Cf. Dashiell v. Grosvenor, 162 U. S. 425, 432, 16 S. Ct. 805, 40 L. Ed. 1025.

Defendant’s contention that rod 31 must be retained because Dickson’s drawing does not indicate that it may be removed, cannot be accepted. As was said by this court in Manhattan Book Casing Mach. Co. v. E. C. Fuller Co., 204 F. 286, 287, “a patented device is not to be held inoperable because slavish adherence to the drawings develops obstacles.” A patentee is not required to indicate in the drawings every suggested use; ■ and it obviously required not even great mechanical skill to disconnect rod 31 in accordance with the specifications. When this is done, admittedly the Dickson brake, applied through the foot button alone, is clearly operable.

3. Claims 1, 3, and 5, however, do not specify the use of a valve in the pipe connection between the manifold and the brake cylinder. Under this construction, it is patent that every time the engine throttle is closed to idling position, a vacuum will be created and the brakes applied. Por the reasons above stated, such construction would be impracticable and commercially useless; consequently we must hold these claims invalid as disclosing an inoperable device.

4. Infringement was not seriously controverted; the evidence amply sustains the finding in this respeet as to claims 2 and 6. Defendant’s brake employs a brake cylinder, open at one end, and connected at the other end through a separate pipe to the intake manifold, the pipe containing valve means for controlling the admission of suction or air to the cylinder, with such valve means operated by a foot lever.

5. On its cross-appeal, plaintiff assigns error in respeet to that part of the decree which finds the Boot patent, though valid, not infringed by defendant’s device.2 The Westinghouse valve means consists of two separate sliding valves, suction valve A and air valve B, which are separated by a spring. Valve A is actuated by a rod running through but not connected to valve B; valve B is actuated by a button which slides on the end of this rod, this button being attached to the end of lever K which in turn is actuated by link J and one-half of the split lever G — D. The function of this device is identical with that of Boot, When the foot pedal rod E is moved, the split lever O moves forward through its entire free motion and thereby closes valve B, the air valve, and opens valve A, the suction valve, to exhaust the air in the brake cylinder. This causes the piston to move in operating the brake through the lever C and at the same time returns the split lever G half way baek to its initial position.' This movement closes suction valve A without affecting the air valve B and thus holds the piston in the position to which it has been moved. Any additional pressure on the foot pedal repeats this operation.

Plaintiff contends that defendant’s valves A and B are in effect a single valve means which is the équivalent of the Boot device and that the split lever G — D is the equivalent of the follow-up element 22 in the Boot patent. Defendant contends that his valve does not infringe because the brake is applied and/or held by the repeated opening and closing of the suction valve A, that a followup slide, which is characteristic of the Boot patent, is unnecessary, and that the operation whereby a valve is seated or unseated by the movement of the piston was well known to the prior art.

The single point of similarity between the two devices consists in the control of the valve means by the brake piston to hold the brake. • In defendant’s device, the piston operates repeatedly to open and dose the single suetion valve A; in the Root patent, the piston operates on a separate, automatic follow-up slide. The conception of a separate followup element was not new. United States patent No. 530,994 to Howe and Clark, issued December 18, 1894, and patent No. 1,143,034 to Brown, issued June 15,1915, show straight line follow-up elements whieh are automatically operated by the movement of the piston. Indeed, the Howe and Clark device shows a cylinder open .at one end; had suction rather than pressure been indicated, this device would have been close to what Boot achieved by a rotary follow-up element. The automatic cut-off shown in patent No., 203,-224, issued April 30, 1878, to Wadsworth, shows a elearly defined similar follow-üp -element. The Casey and Cavin patent, No. 1,-082,733, issued December 30, 1913, shows a similar type of valve provided, in the words of the patent at line 34, with “two independently movable disks, one operated by a hand lever and the other operated by the piston.” These devices may be said to be the progenitors of Root whose contribution consisted in utilizing a rotary follow-up element to control a piston operating in a vacuum suction cylinder.

Defendant’s device, on the other hand, was developed from a different line of valve means in which a separate suction valve is caused successively to open and close and a separate éxhaust • valve is provided. This *854type of valve means is illustrated by the two patents to Massey, No. 358,866, issued March 8, 1887, and No. 351,786, issued November 2, 1886. These Massey patents operate on suetion pressure, and while the drawing shows the valves seated or unseated by means of elastic diaphragms, the .patent explicitly states that diaphragms or pistons may be used interchangeably. The Jenkins patent, No. 818,967, issued April 24, 1906, shows a similar device, operated by suetion pressure, for automobile steering gear.

*853

*854As plaintiff’s witness Bragg very clearly testified, the chief difficulty in the practical adaptation of automatic valve mechanisms to automobile braking lay in the fact that brake rods and levers were lengthened or shortened by the torque exerted by the brakes, the twisting of axles, and the up and down movement of the chassis on the springs. This might result in the air valve being reopened and the brakes released after having been applied, thus endangering the occupants of the ear. In the Root device, this result is impossible because the slide 12, controlled by the foot pedal alone, is the only means of opening or closing the air port. Piston movement cannot release the brakes, and while violent jarring of the brake rodding may move the piston and the follow-up element 22, the brakes will hold. The only way the Root valve can readmit atmosphere and release the brakes is when the operator moves the foot pedal to do so. As the witness Bragg demonstrated, defendant also avoids this difficulty due to torque; but we cannot agree that he accomplishes it in the same manner as the Root teaching, that is, by means of a separate follow-up element 22. Instead, defendant has adopted the old separate valve means, in which there are two independent valves, one for suction and one for air. To prevent the air valve B from reopening if back pressure (torque) is brought to bear by the rod going to the brake, defendant takes up this back pressure in the split-lever C-D. Thus, if the brake rod F in defendant’s device pushes back, it pushes against the fixed half, I), and cannot open the valve B which will open only when the rod E and lever C are released by the foot pedal. This split lever device secures the same functional result as does the Root device, but we are entirely clear that it does so without use of the distinctive Root separate follow-up element to hold the vacuum.

Consequently, the Root claims do not even read on defendant’s valve device unless the statement in claim 2, “means automatically operable by the piston for cutting off communication from the suetion line to the cylinder,” can be extended to include the intermittent opening and closing of the suction valve A. We are of the opinion that the District Court properly held against thus extending the claim. As the function of a device cannot be patented (Westinghouse v. Boyden Power-Brake Co., 170 U. S. 537, 18 S. Ct. 707, 42 L. Ed. 1136), and as here the result is achieved by entirely different and equally ingenious means, we agree that the defense of noninfringement of the Root patent has been sustained.

The decree of the District Court will be modified, by finding elaims 1, 3, and 5 of the Dickson patent invalid, and, as modified, is affirmed, without costs.

(dissenting in

part). I agree that it was not beyond the compass of a journeyman mechanic to eliminate the rod, 31, and make Dickson’s patent work by the treadle alone. This would preserve the kernel of the invention, which was certainly meritorious and new. Root’s patent was in detail altogether different from the defendant’s brake, yet both had this in common, that the operation of the piston locked the valve mechanically and held the pressure. Each accomplished this by an arm which closed the suetion valve, though by an entirely different meehanism. Nobody had ever applied such a device to a motor ear before. Root’s patent was, indeed, not limited to these alone, but that was its principal use, as he 'declared. It seems to me that we ought to allow his invention, to be generalized so far, and while claim two is not very aptly expressed, verbal correspondence is not important when the underlying notion has been taken. Courts keep repeating that a function is not patentable and there is indeed the highest authority for saying so. Yet I suppose that nine claims out of ten are drawn for “means” for doing this or that. What that is but a claim for a “function” I have never been able to see. If there be any such rule, it seems to me more honored in the breach than in the observance; surely it cannot be that it can be evaded by so simple a device as the use of this nearly universal word. At any rate the language of claim 2 serves well enough, I think, to cover a piston which has a mechanical connection with the suetion valve to shut it off by its motion.

I agree as to Dickson’s patent, but I think we should give a decree under claim 2 of Root’s.