Motor Improvements v. General Motors Corporation, 49 F.2d 543 (6th Cir. 1931)

William Houston Kenyon, of New York City, and Charles E. Hughes, of Washington, D. C. (Theodore S. Kenyon, Nelson Littell, and Harold H. Corbin, all of New York City, on the brief), for appellants.

Drury W. Cooper, of New York City (Allan C. Bakewell, of New York City, on the brief), for appellees.

Before MOORMAN, MACK, and HICKS, Circuit Judges.

MACK, Circuit Judge.

These cases are before us on appeals from two decrees, each dismissing a bill in a suit charging infringement of United States letters patent. Two claims of No. 1,594,334, four of No. 1,594,335, both issued July 27, 1926, and nine of No. 1,624,689, issued April 12, 1927, all to appellant Sweetland, are involved in one suit; three claims each of No. 1,646,377 and No. 1,646,378, both issued October 18, 1927, to appellant Sweetland and George H. Greenhalgh, in the other.

The two suits were consolidated for hearing in the District Court and are here on the same evidence by stipulation. Sweetland is owner of the five patents; Motor Improvements, Inc., is an exclusive licensee engaged in manufacturing the oil filters and filtering systems specified in the patents. The A. C. Spark Plug Company, which manufactures the alleged infringing filters, is a wholly owned subsidiary of General Motors Corporation, on whose automobiles these devices are installed. An extended trial resulted in a very large record which includes evidence of numerous tests and demonstrations conducted by the parties both prior to and during the trial. At the conclusion of the arguments, the District Judge, in an oral opinion, held the five patents invalid because of anticipation and prior use, but stated that, if valid, all the claims in suit would have been infringed by defendants' devices.

Because of the number of patent claims and the complexity of the issues involved, the cases were very fully argued orally and in the briefs. But much of the difficulty encountered in the consideration of the intricate and voluminous record and briefs has been engendered by the parties themselves. Plaintiffs, in an effort to sustain, and defendants, in an endeavor to invalidate, the patents in suit, have freely used the specifications and claims in all of the patents to support their various contentions as to each of them. Language found in the specifications of one patent has been used to explain, support, or discredit claims in another; theories *544 expounded in later patents have been employed to supplement or destroy those asserted in earlier ones. Except where an earlier patent is cited as a reference against a later patent, issued to the same inventor, this practice of throwing everything into one general discussion is not helpful. Indeed, it serves only to increase the ever difficult task of determining nice questions of validity. In an effort to crystallize the issues involved, we shall consider each of the patents in turn, and preface the entire discussion with a brief description of the theory and art of automobile lubrication to which these particular patents relate, and a short summary of the commercial history of plaintiffs' patented device.

1. The function of a lubricant in any mechanism is to minimize friction between moving surfaces, usually bearings of one form or another. Hydrocarbon petroleum derivatives make ideal lubricants and are today almost universally used in motive equipment. Such oils readily form a film, in the form of tiny globules of oil, between the moving parts which can slide thereon with a minimum of friction. However, when this oil becomes contaminated with dirt, sand, particles of metal, or other abrasive materials, its lubricating value is not only impaired, but the oil and abrasives form a grinding compound which will wear down the lubricated surfaces and thereby loosen the bearing. Moreover, the thickness of the film depends primarily upon the viscosity of the oil, that is, upon its thickness; and since this viscosity varies inversely with temperature, the hotter a bearing becomes, the less effective is the lubrication obtained. Friction and heat increase with speed and pressure. Consequently, where bearings revolve at high speeds and under great pressures, considerable heat is generated and a continuous and copious supply of fresh, clean oil is necessary.

In the modern internal combustion engine, such as is employed in the automobile, great power is secured by the attainment of high speeds; and the necessary compactness of construction requires the use of relatively small bearings. These revolve under great pressure and, because of the combustion in the cylinders above them, are subjected to considerable additional heat. For proper lubrication a flow of oil as great as 100 to 250 gallons per hour may be necessary to secure a velocity and pressure sufficient to force the oil into each of the engine bearings. Obviously, in an automobile it is not feasible mechanically nor practicable financially, to supply this quantity of new oil and to discard that used. Only a limited quantity of oil, usually six quarts, can be carried; some system of repeatedly recirculating it had therefore to be devised.

Ordinarily the oil is stored in the bottom of the crank case and from there pumped through small pipes or otherwise supplied to all of the important bearings. The usual practice is to pump the oil to the center of each bearing, where it spreads to the edges under pressure and then emerges to drip or flow down again into the crank case. The oil pump supplies the requisite pressure; and, in addition, the movement of the cranks and connecting rods in the crank case serves to splash oil to all moving parts. The rapidly re-circulating engine oil, however, is likely to pick up dirt and abrasives from many sources. Core sand and metal particles in the new engine are washed down by the oil into the crank case and carried along through the system to the bearings. Road dust is drawn into the crank case through the breather pipe which connects the crank case with the atmosphere. Dust entering through the carburetor may in the cylinders become carbonized and form large gritty, abrasive particles which also enter the oil stream. Carbon particles form from the coking of oil cracked on the hot bottom surfaces of pistons, and these particles likewise drop into the crank case reservoir. Unburned fuel and unevaporated water, the latter produced as a by-product of engine combustion, furnish additional sources of oil contamination. This water tends to emulsify the oil, thereby reducing its lubricating value; the tendency to emulsify, moreover, is greatly increased if the oil is dirty.

These difficulties were very early recognized and were met in part by the now familiar notice to automobile owners to change the crank case oil every five hundred miles, both in summer and in winter. In addition, coarse wire mesh screens were often interposed at some point in the oil line to remove the larger particles of dirt and sand. With the continued development and widespread use of the automobile there was need or, in any event, the buying public was induced to believe in the need of some more effective method of filtering oil in situ. The practical problem was to devise a filter, small enough to be carried under the hood or elsewhere on an automobile, but nevertheless effective to remove the fine abrasive particles in the oil without impeding the free flow of oil under pressure to the bearings.

2. The filters and filtering systems developed and manufactured under the five patents *545 in suit apparently solved these difficulties, which, however, may not have been as real as the extensive and effective advertising of both plaintiffs and defendants indicated. Plaintiff corporation introduced its "Purolator" early in 1924, and despite early sales resistance on the part of both manufacturers and public, the use of the device grew steadily. Large sums were expended to meet the growing demands of the business. In February, 1924, the General Motors Research Corporation, another of defendant's subsidiaries, inquired of plaintiffs concerning the practicability of the new filter; samples were sent to it. The device was tested in defendants' laboratories through 1924. Negotiations continued and in April, 1925, the Purolator was adopted by General Motors Corporation. Beginning in June, 1925, it was made standard equipment on the Buick, Oakland, and Cadillac cars of that company. The adoption of the filter was widely advertised by General Motors. In its advertisements it stated that plaintiff's device was one of the many new developments submitted, that the Purolator had been thoroughly tested in extensive experiments with every known method of filtering, and that it would thereafter form part of the necessary equipment on these automobiles.^[1] Purchases continued up to June, 1926, for Buick, and to November, 1926, for Oakland. Approximately 360,000 Purolators were purchased by the General Motors Corporation from Motor Improvements, to which it paid more than $900,000. Beginning in the fall of 1926, however, defendant A C Spark Plug Company began the manufacture of the alleged infringing devices; purchases from plaintiff ceased, and the A C filters were thereafter used on General Motors' cars. The immediate effect on Motors Improvement was the halving of its business by the withdrawal of General Motors' purchases and the active competition of the A C filter in the open market, at lower prices, resulting in plaintiff's inability to compete at a profit, save by a change in its product at considerable expense. The A C Company advertised in February, 1928, that in addition to two million cars theretofore equipped with its filters, ten thousand more such cars were leaving the factory daily. Both in advertising and in the instruction books,^[2] defendants emphasized the utility of the oil filter and admonished proper care in its use, lest its improper functioning deleteriously affect the engine.

A study of the prior art indicates that a number of attempts had been made before Sweetland, to devise an oil filter suitable for automobile engines. Defendants rely upon these to sustain their defense of anticipation; plaintiffs assert their lack of utility. Clear it is that the Purolator was the first automobile filter to achieve commercial success. An earlier unpatented filter, the only one actually used in a standard automobile, Jardine's filter in the Royal Tourist model of 1910, was installed in but eighty cars. We shall consider it fully hereinafter as an alleged anticipation; at this time, we refer to it only to point out that, though in actual commercial use long before Sweetland's patents, it never achieved commercial success. Sweetland, therefore, when he entered the field, found no filter competitor in the automobile industry. While this is an industry in which the purchasing public may be largely ignorant of mechanical devices, and keenly competing manufacturers doubtless are ready to adopt comparatively inexpensive additions that furnish new and attractive talking points for their extensive advertising, nevertheless the ready and wide adoption of Sweetland's filter, after it had been thoroughly tested by defendants, weighs heavily in case of doubt as between utility and nonutility, invention, and mere mechanical advance. Diamond Rubber Co. v. Consolidated Tire Co., 220 U.S. 428, 441, 31 S. Ct. 444, 55 L. Ed. 527; Temco Co. v. Apco Co., 275 U.S. 319, 328, 48 S. Ct. 170, 72 L. Ed. 298; Smith v. Goodyear Dental Vulcanite Co., 93 U.S. 486, 495, 23 L. Ed. 952; Gordon Form Lathe Co. v. Walcott Mach. Co., 32 F.(2d) 55, 58 (C. C. A. 6th); Kurtz v. Belle Hat Co., 280 F. 277, 281 (C. C. A. 2d). Especially is this true if infringement be clear and the infringer itself has proclaimed and continues to proclaim the merits of the device. Sandy MacGregor Co. v. Vaco Grip Co., 2 F.(2d) 655, 656 (C. C. A. 6th); Inland Mfg. Co. v. Am. Wood Rim Co., 14 F.(2d) 657, 659 (C. C. A. 6th). As we stated in this latter case, the fact that defendant there became a licensee (here a large purchaser and user after careful study by research men) "strongly tends to prove that it believed *546 the production of such a steering wheel involved invention."

Of course, if noninvention be clear, commercial success is immaterial. Detroit Co. v. Dodge Bros. (C. C. A.) 33 F.(2d) 743, 747; Carroll-Thomson Co. v. Central Brass & Fixture Co., 22 F.(2d) 253, 254 (C. C. A. 6th). While in these days of high-pressure salesmanship, success is often more attributable to clever exploitation than to merit, the general adoption of a device after elaborate tests by experts strengthens the presumption of validity.

3. The Filter Patents.

Number 1,594,334, issued July 27, 1926, on a divisional application filed January 10, 1922, sets forth a "Method of preventing Wear of the Moving Parts of an Automotive Internal-Combustion Engine." Number 1,594,335, issued the same day on a divisional application filed April 30, 1926, is for a "Means for Filtering the Oil in the Lubricating system of an Internal-Combustion Engine." The original application for both was filed July 3, 1920. The two patents are very similar except as hereinafter stated. The drawings are identical and both patents recite, the second in somewhat greater detail than the first, the sources of contamination, the failure of wire screening to remove deleterious abrasives, and the damages supposed to result from a failure promptly to remove such substances.

The Method patent outlines the suggested technique of removing these deleterious abrasives, which is briefly as follows: The oil is withdrawn from the crank case reservoir (2) by an ordinary pump (3) and delivered to the filter (7), filtered, and then "conducted to the crank case by a pipe (8)." Reference to Figure 1, reveals that there is in this patent

no indication that this pipe 8 leads back directly to the crank case, and the specifications throw no light on this other than to mention that the crank case "contains the oil for lubricating the * * * moving parts of the engine."^[3] The remaining specifications of this first patent are devoted to a detailed description of the filter structure which is to be strapped under the hood of the automobile as indicated in the drawing.

This filter is shown in Figures 2 and 3,

and consists of a case (9) provided with a removable cover (10) through the top of which runs a simple inlet pipe (6). In the middle of this cover there is also inserted an outlet pipe (18) connected, by suitably drilled openings throughout its length, with a number of annular filter elements equally spaced thereon. These filter elements do not meet the sides of the case but permit oil to flow freely through it. Each element consists of a coarse wire frame, covered, in one form, with fine wire mesh and coated with an insoluble flocculent filter material. Another form of filter element, important as the only one commercially practiced, is shown in Figure 3. In this, the coarse wire frame is covered with a closely woven cloth fabric, such as duck, canvas, or chaincloth, which, it is stated, may be used satisfactorily either with or without the coating of insoluble filtering material. In the commercial form of both plaintiffs' and defendants' filters, the cloth alone is employed. The specifications state that this fabric may be reinforced with strands of wire to increase its strength. Other accessory devices shown are not here involved and need not be detailed.

The patent directs that the oil be withdrawn from the crank case and pumped into the filter through the inlet (6). After this *547 oil fills the filter case, it will filter through the exposed surfaces of the circular filter elements, thence through the drilled openings and out through the pipe 18. Although such oil enters under the pressure of the pump, it is supposed to permeate slowly through the filter element, at a rate stated by plaintiffs' expert to be no faster than the movement of the minute hand of a watch. This "required flow of filtered oil through the filter is obtained by reason of the extensive filtering surface which is provided by the combined filtering surfaces of the several filter elements." The patentee also states "that the removal of microscopic solids from the oil is not the desideratum of my invention but rather the removal of solid particles * * * at a rate substantially equal to or in excess of the rate of accumulation of said deleterious solids."

In summary, this first patent describes a method of using the filter structure which is inserted in the oil system of an internal combustion engine. Apart from what may be gathered from the relative proportions of the engine and filter shown in the drawings, there is no definite specification of the size either of the filter case or the individual filter elements. Nor do the claims aid materially in this respect. The second claim in suit describes a filter containing "a filter surface therein of a density to remove deleterious solids from the oil passing therethrough and of an effective area substantially in excess of prior practice, and correlating the size and density of the filter area in said casing to the volume of oil in the system, the oil pressure, the rate of oil flow through said system, and the quantity of deleterious solids to be removed, to so control the rate of deposition of said solid particles upon the filter surface that the oil content of said lubricating system may be circulated through said filter under normal engine operation, for a period, equivalent to several hundreds of miles of running, before the deposition of solid particles on the filter surface is sufficient to reduce the flow of oil therethrough, below the frequency necessary to prevent the accumulation of deleterious solids in oil in said lubricating system."

The first claim, printed in the margin,^[4] is merely a less elaborate statement of the same ideas.

In the second, or "Means" patent, the specifications are not very different from those of the first; they contain, however, an elaboration of Sweetland's theory as to the operation of his filter and a somewhat fuller statement of the prior art. This is summed up as follows: "Filters for filtering used crank case oil and removing the fine carbon sediment, dust and the like have been known heretofore, in which, however, the size of the filter necessary to filter the engine oil with sufficient rapidity to keep it clean, was much too large to be carried as an accessory on an automobile. Filters have also been known which were designed to be carried on the automobile engine, but in which the effective filtering area if of sufficiently fine mesh to accomplish filteration, was much too small to filter the crank case oil at a sufficiently rapid rate to effect any substantial clarification of the oil, and where in these latter filters, screens or coarse mesh material are used through which the oil may flow without retardation, no useful result is accomplished. It will be readily apparent that a filter of the hitherto known type if able to filter the contents of a normal crank case once every hour of engine operation must be as large or larger than the automobile engine itself as, if the filter were made of sufficiently small size to be carried on the engine, it could filter only a few ounces of oil per day and would be ineffective for my purpose. The two types of filters heretofore known are equally ineffective, the former because of its prohibitive size and the latter because the filtration of a few ounces of oil in the crank case will never clarify the entire contents or remove the impurities as rapidly as they are added by the operation of the engine." Concerning the use of cloth fabric in the alternate form of filter element, the inventor calls attention to the fact that in using these elements, "at the commencement of the filtering operation a portion of the fine suspended particles will pass through the fabric 32, but within a short time the impurities collected thereon will serve to prevent the passage therethrough of any of the fine particles with the result that only the clear oil freed from all suspended *548 matter will pass through the fabric." This coating or slime-bed of impurities is independent of any additional filtering material added to the oil; elsewhere, the inventor states that while this may be done, it is unnecessary for satisfactory operation of the cloth filter elements.

No additional or more definite directions as to the requisite size of the various parts of the filter are given in the second patent^[5] but it stated that the "invention which provides a large filtering area, large enough to filter the entire contents of the crank case once in every hour of engine operation, or more often and remove the impurities at a more rapid rate than they are formed, and yet enclosed in a casing of sufficiently small size to be carried on the engine provides a novel solution to both these problems and the adapting of the size of the casing to the size of the engine and the filter area to both the permissible size of the casing and the contents of the crank case, and the pump pressure so as to provide a definite rate of filtration as compared to the engine operation is a new idea in regard to filters." The claims of this second patent are directed to the filter structure itself; a typical claim No. 4 is set out in the margin.^[6]

We have detailed the patent specifications at length and generally in the words of the patents themselves, because the primary attack made upon the first two patents by defendants is that the claims are vague and fail to contain a sufficient disclosure of the filter structure developed by plaintiffs and adopted by defendants. The precise issue here is whether the specifications, drawings, and claims read together, offer enough information to enable one skilled in the art to construct the devices in suit. In our opinion, they do. Minerals Separation v. Hyde, 242 U.S. 261, 271, 37 S. Ct. 82, 61 L. Ed. 286; Merrell-Soule Co. v. Northland Dairy Co., 28 F.(2d) 924, 927 (C. C. A. 6th). The nature of the problem made it desirable, if not necessary, that the directions be given merely in general form. The size of the filter necessarily would vary with the size of the engine, the number of cylinders, size of pistons, displacement, etc. The rate of filtration would vary with the speed of the engine and the pressure developed in the pump; the size of the filter would have to be such that even at ordinary running speeds a sufficient quantity of oil could be filtered to comply with Sweetland's directions that the entire contents of the crank case be filtered once in every hour of engine operation. Plaintiffs' expert admitted that no specific figures as to filter area were given, but stated that a skilled mechanic would understand from the drawing that a filter having an area of from three to six square feet, inclosed in a case small enough to be carried under the hood of an automobile, was intended.

The correlation of filter to engine referred to in the specifications and claims is not a matter of absolute dimensions, but of relative proportions. The relation is clearly enough illustrated in the drawings. Essentially, once the conception that a plurality of cloth filter elements was to be so employed that the same oil would slowly permeate through only one surface and all of it at a combined rate sufficient to filter the crank case once every hour of minimum engine speed, was understood, a competent mechanic might vary size to crank case capacity. The specifications are addressed to persons skilled in the art.^[7]

Under all the circumstances of the case, in view not only of the striking commercial success of plaintiffs' filters, to which we have referred, but also in the light of the history of the prior art hereinafter discussed, and its state at the time these devices were introduced, we are satisfied that an ordinary, skilled automobile mechanic, from the disclosures of these first two patents, could have achieved a filter similar to those first marketed by plaintiffs and purchased by defendants. Compare Temco Co. v. Apco Co., 275 U.S. 319, 330, 48 S. Ct. 170, 72 L. Ed. 298; Eibel Process Co. v. Paper Co., 261 U.S. 45, 65, 43 S. Ct. 322, 67 L. Ed. 523; Wisconsin Chemical Co. v. Chute, 261 F. 89, 92 (C. C. A. 7th).

*549 Coming directly to the issue of anticipation, the prior art patents, as disclosed by this record, show three general types of filter. The first consists of an ordinary wire screen designed to operate in a high velocity stream of oil and to remove only coarser particles. Such screens offer no back pressure and ordinarily were inserted in a direct line to the bearings or moving parts to be lubricated. The rapid flow prevented the formation of a slime bed and consequently the finer abrasives were not caught. Such filters are still in common use on many automobiles. The second type of filter is that referred to as a depth type. In these, the oil or other liquid flowed at considerable speed through a deep, porous bed, consisting usually of sand, cotton, mineral wool, or other substances, which trapped the solid particles in its interstices. Such filters are usually used in commercial filtration where the solid filtrate was discarded and the filtered liquid alone desired. The third type consisted of a filter containing a dense, thin surface of considerable area, usually made of filter paper. In this type the liquid seeped through the filter, or usually a series of such filters, depositing particles thereon and building up a filter cake which was usually intended to be and was recovered.

The record reveals that between 1907 and 1920 some thirteen inventors secured patents in an endeavor to perfect a practicable filter for automobiles and internal combustion engines. A few of these attempts may be briefly mentioned. British patent No. 13,187, issued in 1907 to Winton, shows merely a horizontal series of perforated discs covered with cloth encased in a filter case which is half filled with water. In this device the oil percolated through the water and through the cloth covered discs, but all of the oil had to go through each of the discs. There is little in this patent to suggest the Sweetland filter in which an extensive filter area is obtained by having all of the oil pass through the filter elements but not through each of the many elements; some passes through one and some through another, thereby reducing the pump velocity to an almost imperceptible flow. United States patent No. 923,747, issued in 1909 to the same inventor, similarly shows merely a box containing a vertical series of fine wire mesh screens through each of which all of the oil must pass. United States patent No. 1,020,774, issued in 1912 to Nilson, is practically the same except that the space between the screens is filled with a coarse filtering material such as hair. Similar wire filters are shown in Cloudsley, British patent No. 11,727 of 1914; Weiland, United States patent No. 1,203,290; and Winton and Anderson, United States patent No. 925,258 of 1909. The other attempts cited give no details concerning the filter structure. No one of these shows a filter in which the oil could have been pumped under pressure and yet have permeated slowly through the filter elements.

Direct attack on these first two filter patents is made by defendants on the ground that there was a complete anticipation of everything disclosed by Sweetland in some eight United States, British, and German patents. All of these are offered, in the first place, to show that the use of the various insoluble filtering substances specified by Sweetland, ranging from kieselguhr to wood pulp, was old in the art. This is neither denied by plaintiffs nor is it important, since the validity of the filter patents does not, in our judgment, turn upon the particular filtering material, but upon the construction and operation of the filter. Moreover, the form of filter element commercially developed both by plaintiffs and by defendants was the one in which canvas was used. That a wide variety of materials for the alternate form of filter element was suggested by the patentee does not render the patent invalid where the essence of his invention is not dependent upon the particular substance employed. Similarly, the further charge, as to which there is no direct evidence, that the use of these other substances was not practicable, may be disregarded.

It is next contended that the actual construction and operation of the Sweetland filter are also anticipated in these patents. This contention likewise cannot be accepted. The first and closest reference is United States patent No. 335,040, issued to Piefke in 1886. It shows a filter composed of metal disks alternately drilled; clamped between each pair of discs is what appears to be filter paper. This patent contains but one feature of the Sweetland filter, namely, that the liquid filtered passes through only one surface. But there is no requirement whatever as to size, no indication as to the rate of flow, and no evidence that the filter shown, which is called in the patent a "Water Filter," could be adapted to internal combustion engines. The filter shown is an early prototype of the Royal Tourist filter hereinafter discussed as chief prior use. United States patent No. 611,507, issued to Quiggin in 1898, shows a complicated "Filter for Feed Water for *550 Steam Generators," in which the water, under high boiler pressure, appears to pass consecutively through the several surfaces. Although the specifications of this patent do refer to "a large area of filtering material in a comparatively small space," we find nothing in it remotely suggesting a solution for the filtering problems solved by the patents in suit. United States patent No. 839,772, issued in 1906 to Krause, similarly has a statement as to relative size of filtering area and filter, but the latter appears to be a large device, containing many inlets and outlets, through which the liquid is introduced, and which are closed while it filters and drains off through a concavity at the bottom. This patent is clearly no anticipation. Winton, United States patent No. 923,747, has already been discussed. Both Morris, United States patent No. 1,137,075, issued in 1915, and Bride, German patent No. 79,614, issued February 13, 1894, show merely means for mounting and distending a filter bag of unknown size within a casing. Neither relates to the present problems and neither has any indication whatever as to size. These six references cannot, either singularly or cumulatively, be deemed an anticipation of Sweetland.

We come then to the ground upon which the District Judge chiefly based his finding of invalidity of these first two patents, namely, prior use as shown by the Royal Tourist filter. The history of this device, always important on an issue of prior use, may be briefly stated. In 1910-11, about eighty Royal Tourist automobiles were sold, equipped with a filter designed by one Jardine, a prominent automobile engineer, who testified for defendants in the present suit. Two of the automobiles were located by defendants at the time of suit and evidence concerning them was introduced. One of these was in such state that it could not be repaired; the filter, however, was removed and fashioned into an exhibit. The other, owned by witness Woodward, had been in continuous operation since 1911; at the time of the trial, it had run over 100,000 miles, and at that time, the owner ran it another thousand miles to Michigan. For present purposes, we need discuss only the construction and operation of the Royal Tourist filter as disclosed by the evidence concerning the one in this Woodward car.

Around this device and its alleged operation, a very considerable part of the present case has been built. The filter, as attached and operating, was exhibited to the court but plaintiffs were denied an opportunity to make an extended practical road test. Accordingly, they constructed replicas and tested them in modern automobiles. The validity of these tests is denied by defendants on the ground that the Royal Tourist device was not intended for operation in a modern car. Consequently, they in turn constructed larger replicas of the Royal Tourist filter which they demonstrated on present day automobiles. Plaintiffs objected to these tests on the ground that the original structure was varied by the introduction of gaskets, and the filtering area was made four and one-half times as great as that of the 1910 structure.

We need not detail the rather complicated construction of the Royal Tourist filter other than to state that it consisted of a series of metal plates, each pair of which embraced a piece of filter paper, and which were grooved so that oil introduced at the axis of the bottom plate might filter through the paper and emerge through V-shaped slots radially cut in the top plate. The eight pairs of plates were clamped together by inserting a threaded bolt into a boss in the bottom of the case and tightening a nut at the top. The stack of filter plates rested on a shoulder formed by the lower portion of the case, leaving a concavity into which the dirty oil was introduced. Thence the oil flowed upward through crevices in the filter stack, into the filter elements, and out into the space between the stack of filters and the outer wall of the case. Concededly this was a filter adapted for an automotive internal combustion engine. Although the filter structure, in construction and operation, is somewhat similar to Sweetland's in that it operated under pressure and had a plurality of filter surfaces through only one of which the oil had to pass, we conclude that the device is not sufficient to show prior use of the Sweetland invention.

It is too well settled to require citation of cases that the burden of proving anticipation by prior use rests with defendants, and that every reasonable doubt must be resolved against them. In the present case while the actual use of a filter on the Royal Tourist cars of 1910 is clearly established, there is grave doubt whether or not the Royal Tourist device was a practicable oil filter. The remarkable longevity of the Woodward car is due in great measure to the care lavished upon it by the owner he always drove it himself, made extensive mechanical repairs himself, being an expert mechanic, and from time to time made improvements. Most important for present purposes, it appears that he changed his oil every *551 five hundred miles, so that at no time was the accumulation of abrasives permitted to reach a harmful stage. Moreover, it affirmatively appears that in 1924, long after Sweetland's original application was filed, Woodward made certain changes in the method of clamping the filter stack to the case so as to avoid leakage at the bottom. This change, defendants claim was immaterial, but in view of the convincing demonstration by plaintiffs that leakage occurred at this point, we are of the opinion that this was not merely a repair but a structural change. A prior use which requires modification to perform the function of the patent attacked is no anticipation. Babcock & Wilcox Co. v. Springfield Boiler Co., 16 F.(2d) 964, 969 (C. C. A. 2d); Electrical Engineers' Equipment Co. v. Champion Switch Co., 23 F.(2d) 600, 603, 604 (C. C. A. 2d); United Shoe Machinery Corp. v. Day Wood Heel Co., 46 F.(2d) 897 (C. C. A. 6th). Although the point is vigorously contested, it seems reasonably clear that in the Royal Tourist device considerable leakage occurred between the filter plates and that such leakage could not have been readily detected. Furthermore, of the 10 per cent. of the oil that went through the filter, hardly more than 1 per cent. seems actually to have been filtered, the remainder leaked through the shoulder at the bottom, through openings caused by the forcing of the oil-soaked filter paper into the V-shaped slots or through punctured filter paper. Finally, as we have said before, the striking commercial success of plaintiffs' filters, after persistent attempts to solve the problem between the date of the Royal Tourist device and the introduction of the Purolator, is satisfying evidence that the earlier device did not offer a filter of commercial utility.

There remains the question whether the A C filters constitute infringements of the first two patents. Defendants' filter comprises a bag rolled spirally and held in a wire frame inside a cylindrical casing. The inner turn or edge of the bag is connected through the top of the casing with the oil inlet pipe. A second pipe through the top of the casing permits filtered oil from the interior of the can to flow back to the crank case. The bag is stitched longitudinally in five or six parallel rows so that when the bag is rolled up and the oil is pumped into it under pressure the channels between these parallel rows of stitching swell up to form a series of spiral oil channels arranged vertically, one on top of the other in the casing; the cross section of each spiral channel being approximately the shape of a lens. The edges of the lens-shaped channels are at the rows of stitching and the distended mid portions of the lens-shaped channels press out against the outer rolls of the bag so as to space the turns of the bag apart and to form similarly shaped channels outside of the bag and between the spiral turns thereof and between each pair of lens-shaped channels. The oil which seeps through the walls of the bag may flow to the periphery of the roll through the diamond-shaped channels and join with the body of clean oil in the casing and finally go out through the outlet pipe in the cover.

The first difference between this device and that shown in the patents is that the oil filters from the inside of the filter bag to the outside, depositing the solid particles of abrasives on the inside, whereas in the patent filter, the oil filters from the outside of each filter element through to the inside, depositing solid particles on the outside. This difference is immaterial since it merely requires the reversal of the cloth so as to expose the nap to the unfiltered oil.

It is next asserted, to show noninfringement, that defendants' filters do not operate so as to build up the slime bed of impurities, specified in the patents, because out of the 600 square inches of surface comprising defendants' filters only 20 square inches, it is said, operate at one time. Defendants contend that at the beginning only the innermost portion of the bag acts as a filter, and as this fills up with dirt, the oil entering the bag travels farther along the unused portions. While it may well be, as certain of the testimony indicates, that more impurities collect on the inside surfaces nearest the inlet, it does not necessarily follow that this portion of the bag alone serves to filter in the beginning and that successive portions alone filter thereafter. For it is apparent that both plaintiffs' and defendants' filters operate under pressure from the oil pump. In both, the oil under pressure necessarily fills the entire case before being forced out through the outlet pipe. In plaintiffs' device, the dirty oil fills the case, submerging the filter elements and seeping at slow velocity through the filter fabric to the clean oil space within the elements. In defendants' filter, the operation is reversed but is not different; the dirty oil is forced into the tubular passages of the filter bag, expands them throughout their length and then similarly seeps through the fabric, filling the case with clean oil. Various tests conducted by defendants' expert *552 with an open unfolded filter bag do not in any way affect this conclusion.

Nor is it any answer to the charge of infringement that a patent was granted to an employee of defendants for its device, or that plaintiffs themselves adopted the cheaper spiral bag form. The record clearly indicates that plaintiffs were compelled to adopt the simpler and cheaper form of filter in order to meet defendants' competition, and that the original form is still manufactured for use on the higher priced cars. Even if the subsequent patent (United States No. 1,651,400, issued in 1927 to McKinley and assigned to defendant A C Spark Plug Company) shows a patentable improvement of the Sweetland device, as to which we express no opinion, we are clear that it embodies the disclosures of the latter. The differences in construction are not material. The patentee secures a large area of filter surface within the small enclosure by breaking the area up into units, which operate, separately but simultaneously, upon fractions of the oil; the defendants get the same concurrent fractional operation by dividing their filter area into units of channel form, wound spirally. There is a superficial dissimilarity but it is not substantial. It is not necessary to resort to a wide range of equivalents, for it appears that the claims of the first two patents in suit, as explained by and read in the light of the specifications, read upon the defendants' device. We are in accord with the lower court in finding each of these claims infringed. Economy Baler Co. v. Solar Sturges Mfg. Co., 29 F.(2d) 656 (C. C. A. 6); Cincinnati Cadillac Co. v. English & Mersick Co., 18 F.(2d) 542 (C. C. A. 6).

4. The third patent, No. 1,624,689, issued to Sweetland April 12, 1927, on an application filed November 19, 1920, discloses a method of using the filter of the first two patents in the lubricating system of the present day internal combustion engine, especially those employed in "automobiles, aeroplanes, motor-boats and the like." It describes the filter of the first two patents in considerable detail and adds various directions as to proper size and relation of filter area to engine speed and pressure. These need not be reviewed, for the invention claimed does not depend upon them.

Inasmuch as the drawings accompanying this patent embody a confusing variety of accessory devices not claimed or here involved,^[8] it seems desirable to describe the elements shown and actually claimed, without reference to any drawing. The unfiltered oil is withdrawn from the crank case by the pump, and the supply line from this pump is divided. Several branch lines go directly to the main bearings and carry an assured supply of unfiltered oil to them. A single additional branch line comprises the filter circuit. The first effect of this splitting or dividing the oil stream from the pump is that a supply of oil, under full pump pressure, to each bearing is assured unless the pressure in the branch filter circuit falls and bleeds, that is, draws oil, from the other branch lines. This scheme of course did not permit all of the oil from the crank case to be filtered, but Sweetland was satisfied in repeatedly circulating a portion of this crank case oil so as to keep its average abrasive content below a harmful range. Accordingly, he placed his filter, not in the main line to the bearings, but in a separate branch line. By sacrificing the theoretical possibility of complete filtration, he was enabled to dispense with a separate pump for the filter circuit and yet secure an assured supply of oil to the bearings. This is the first characteristic of the third patent.

But, as we have said, the maintenance of pressure in the main lines to the bearings was dependent upon the maintenance of pressure in the branch line filter circuit. To maintain this pressure below a permissible maximum, Sweetland first shows a relief valve in his filter circuit provided with a return line to the crank case. Thus, if the filter or filter circuit became clogged, the oil might escape and return unfiltered to the crank case. This feature is not particularly important in the present suit. But it is vital that the pressure in the branch line filter circuit should not fall below that necessary to maintain the supply to the bearings fed by the main line. At the beginning of filter operation, before the slime bed would offer enough back pressure, or in the event of a punctured filter or other leakage, there was danger that the pressure in the branch filter circuit would fall and oil would be lost to the bearings. To guard against and control this possibility, Sweetland inserted in his branch filter circuit, on the outlet side of his filter, a valve which could be used to restrict or stop the flow of oil through the filter circuit. This is the second important characteristic of the so-called Divided Stream patent. Both it and the use of a branch line filter circuit, supplied by a single pump, are clearly set forth in the nine claims in suit of *553 which typical claims are set out in the margin.^[9]

We are of the opinion that this third patent discloses no advance over the prior art. First, the scheme of splitting the oil stream, after it leaves the pump, and employing a separate branch line to obviate the necessity of using more than one pump, is clearly shown in prior patents. The use of such a branch line for a separate lubricating and filter circuit is specified in British patent No. 11,727, issued in 1914 to Cloudsley, the specifications of which in part read: "According to another feature of the invention, the lubricant is supplied to the parts (a) and parts (b) by a pump common to both of them and the lubricant which passes through (a) is dispersed or collected in a separate sump and that which passes through (b) is returned to the sump and re-circulated."

It is true that the filter shown in the branch line in this patent is a simple wire filter, but this does not detract from the clear indication of the use of a branch oil filter line. Divided streams in which filtering or oil purifying devices are inserted are also shown in United States patent No. 1,227,481, issued in 1917 to Morris on an application filed in 1912, and in United States patent No. 1,318,086, issued in 1919 to Kennedy. In the latter invention, which was designed primarily for filtering systems on vessels, duplicate pumps are provided for safety, but a split stream containing a branch line in which an oil purifying device is located may be operated by either pump. Similarly, divided oil streams and branch lines for other purposes are shown in United States patent No. 1,207,067, issued in 1916 to Nugent (overflow tank); United States patent No. 1,090,773, issued in 1914 to Winton (by-pass to reservoir); French patent No. 393,934, issued in 1908 to Daimler Motoren Gesellschaft (indicator). Finally, in the Royal Tourist car, hereinbefore described, the oil stream from the pump is split and a branch line therefrom goes to the fan belt and returns directly to the crank case. Although we are of the opinion that the filter described in the Royal Tourist system was not practicable, there was a prior use of oil distributing pipes of undoubted efficiency even if the filter inserted therein was ineffective.

The second characteristic of Sweetland's Divided Stream patent, the use of a restrictive pressure control or by-pass valve in the separate branch so as to insure the maintenance of pressure in the main line, likewise appears to be old in the motor lubricating and other arts. Plaintiffs' expert admitted that the use of such control valves was a common expedient to prevent branch circuits from bleeding or lowering the pressure in other lines. Such devices, in the form either of a fixed restriction or of a controllable valve, are clearly shown in Nugent, supra; Winton, supra; Kennedy, supra; and Cloudsley, supra. A citation from the last is typical: "According to yet another feature of the invention a pressure-reducing valve is introduced into the supply-system at a point beyond the parts (a) and in front of the parts (b); the use of this valve renders it possible to supply the parts (a) with lubricant at a higher pressure than the parts (b); in practice this is found desirable since the parts (a) generally include the main bearings of the engine and other parts where oil under higher pressure is desirable. The pressure-reducing valve is preferably made adjustable so that the proportional quantities of lubricants delivered to the parts (a) and to the parts (b) may be varied by suitably altering the setting of the valve, or may be kept constant in spite of wear in the bearings and consequent increased flow or leakage of lubricant at some points."

Thus it appears that aside from the specification of the use of his filter covered by the first two patents in suit, there is nothing in this third patent which warrants additional protection. It is true that in detailing the particular filter to be used in the divided stream arrangement for the lubrication of an internal combustion engine, Sweetland in his third patent does repeat, and supplement in some particulars, the disclosures of the first two. But the filter itself is covered by *554 the first two patents, and the claims of the third purport to extend not only to the use of the particular filter but also to the arrangement of pump, main line, branch line, and necessary relief and restriction valves. Since these latter elements were not novel, each of these claims is invalid.

5. The Sealed Cartridge Patents.

The fourth and fifth patents in suit, No. 1,646,377 and No. 1,646,378, relate to the sealing of the can which contains the filter elements, so that it cannot be removed and cleaned but will have to be discarded as a unit when full of impurities. These patents describe substantially the filter specified in the first two patents and utilized in the third, but such filter is sealed within a leak-proof casing and means are provided for connecting or disconnecting the casing in the oil circulating system, so that when a filter has served its mileage and has become filled with sludge, it may be conveniently removed without the difficulty or messiness incident to cleaning, and a new sealed cartridge inserted in its place. It is alleged by plaintiffs that the provision of this convenient and rapid method of replacement constituted a patentable invention.

The fourth patent specifically purports to disclose and claim a process or system for collecting and discarding the dirt in lubricating oil, and as elements in this system there are specified the sealed can and detachable elements. The fifth patent purports to cover the renewable filter unit itself as an article of manufacture. The drawings of the two are identical, and various methods of attaching the cartridge filter to the connections are shown. In one, the filter is clamped between abutments which are tightened and set in place by a thumb screw. In another, a coupling is shown somewhat similar to that employed in connecting the links of a common garden hose. In a third, a hollow nut or bolt, drilled to meet corresponding openings in the inlet pipe and sealed cartridge is used.

We are of the opinion that neither of these patents discloses invention. Sealing the can containing the filter elements in itself was not invention. Nor was the provision of standard methods of fastening pipe connections a patentable contribution to the earlier disclosures of the first two patents. There is no need to recite the many patents offered to show invalidity. For we are entirely clear that the use of an easily detachable filter unit is old in the art. The oldest type appears to be the common water filters which were attached to a faucet and unscrewed and discarded when no longer effective. There is nothing novel in the means of attaching and detaching the filter unit, and there was certainly nothing novel in designing the filter for easy replacement. Ease of substitution of new parts is a feature of many automobile accessory devices, and ordinary provision for this is not invention. Finally, nothing in the fact of sealing the filter or making it detachable contributes to its actual functioning.

The decree dismissing the bill of complaint in No. 5327 is affirmed. The decree in No. 5326 is affirmed so far as it dismisses the bill as to patent No. 1,624,689, but is reversed as to patents Nos. 1,594,334 and 1,594,335, and the cause as to the latter patents is remanded for further proceedings.

[1] The Purolator was also adopted as standard by General Motors of Canada, Studebaker, Nash, Peerless, Moon, and several taxicab companies in 1925. The following year it was adopted as regular equipment on the Auburn, Elkhart, Flint, Franklin, Gardner, Hupmobile, Jordan, Packard, Rolls Royce, Stearns, Lincoln, Minerva, Fiat, and many others.

[2] Chevrolet Instruction Book, October 1, 1927; Oakland Instruction Book, 13th Ed.; Buick Reference Book of 1927; Cadillac Operator's Manual, Ed No. 314-3; La Salle Operator's Manual, Ed. No 3032, and others.

[3] The terminal end of this pipe in the crank case reservoir, indicating that the filter circuit is separate from the main line supply to the engine bearings, is a vital feature of the third patent in suit. This feature does not, however, appear in the first two patents. See infra, pp. 552, 553.

[4] The method of preventing the accumulation of solid particles in the oil in the lubricating system of an automotive internal combustion engine, normally containing a substantially constant volume of oil which comprises providing a filter casing of a size small enough to be carried beneath the hood, on an automobile, providing a filter surface therein of a density to remove the deleterious solids from the oil passing therethrough and of an effective area substantially in excess of prior practice, and correlating the size of the filter area in said casing to the volume of oil in the system and the oil pressure and rate of oil flow from the engine oil pump, so that the oil content of the system may be circulated through said filter under normal engine operation with sufficient frequency to remove the solid particles substantially at the rate they appear in the oil.

[5] The third patent, however, contains the definite direction that "the filter area should not be less than the total piston area of the engine, and for the best results I employ six to ten square feet of filter surface for the average automobile."

[6] "A pressure filter for use in the lubricating system of an automotive engine comprising a casing of sufficiently small size to be carried as an accessory on an automobile, inlet and outlet openings to said casing from the lubricating system, a filter surface in said casing of relatively fine mesh fabric, which offers a resistance to the passage of oil therethrough, necessitating a pressure in excess of gravity to effect filtration, said filter surface being of relatively large area as compared with the size of said casing whereby all of the oil in said lubricating system may be filtered with sufficient frequency at normal oil pressures to maintain the oil substantially free of deleterious solids."

[7] This simplicity of correlation for a skill mechanic is perhaps indicated by the fact that when defendants' experts came to increase the size of the Royal Tourist filter, described below, to cope with the oil capacity of a modern engine having many cylinders, they had practically less data than is given by Sweetland in his first two patents. Nevertheless, a filter which defendants assert was of adequate size, was readily constructed.

[8] E. g., oil pressure gauges, sight glasses, means for heating the filter by exhaust gases, etc.

[9] "5. The combination of an internal combustion engine having a crank case, a filter adapted to offer resistance to the flow of oil therethrough and having its outlet communicating with said crank case, a pump for withdrawing oil from said crank case, and means for delivering the withdrawn oil from said pump in part to said filter and in part to the engine parts."

"9. The combination of an internal combustion engine having a crank case, means for withdrawing oil from said crank case, a pressure filter having its outlet communicating with said crank case and delivering it under pressure in part to said filter and in part to the engine parts, said means being constructed and arranged to maintain the thus delivered oil at a pressure above predetermined value at normal engine speeds and means operative to prevent such pressure from exceeding a predetermined value."

"28. In combination with an internal combustion engine, an oil reservoir, a lubricating circuit, a filter circuit, pump means for forcing oil from said reservoir through said lubricating circuit, a filter, conduits leading from said pump means to said filter and from said filter to said reservoir, one of said conduits having provision for presenting such resistance to the flow of oil therethrough that the desired flow of oil through said lubricating circuit is continuously maintained."