This is an action for a declaratory judgment in which appellants sought to have two patents declared invalid and not infringed, and two other patents, of which appellants were, or stood in the title of, assignors and hence estopped from asserting invalidity, not infringed. Appellants are Swanson and his present company, Progressive Engineering, Inc. Appellees are Cotchett and his connected company, Machinecraft, Inc. At the time the patents were applied for Swanson was an employee of appellees. The district court held the first patent invalid, but found against appellants on the other issues. Progressive Engineering, Inc. v. Machinecraft, Inc., D.C.D.Mass. 1959, 169 F. Supp. 291. The three patents involved in this appeal are No. 2,644,202, issued to appellee Cotchett on July 7, 1953; No. 2,716,780, issued to appellant Swanson on September 6, 1955, and assigned by him to appellees; and No. 2,755,515, issued on July 24, 1956, to Cotchett and Swanson, and assigned by Swanson to appellees. No. 2,716,780 was the last applied for. These patents all related to ball-bearing top rolls to be used on spinning frames employed in the textile industry to manufacture yarn or thread. There was a need for such improvements. Satisfactory operation of top rolls calls for the application of heavy weight through the rolls, and rotation at high speeds, requiring lubrication. Frequent oiling has manifest disadvantages. On the other hand, ball-bearing rolls, although long known in the industry as rarely needing lubrication, require adjustment. Also, they are susceptible to invasion by cotton fibers, or lint, known in its finest form as fly. Both of these factors, if unremedied, cause such rolls to operate unevenly, to the detriment of the product, or even to bind altogther While ball-bearing top rolls had been in use since Campbell patent No. 628,521 of July 11, 1899, because of these failings they had found little acceptance for fine work. Appellees' rolls were designed to correct this.

The district court found, 'This is an outstandingly strong case of a valid invention and a clear infringement * * * The efficiency and economy of Cotchett's invention in overcoming difficulties theretofore experienced was at once appreciated. The most experienced and sophisticated suppliers of machinery saw its novelty and utility. From 1948 to 1958 2,000,000 rolls embodying Cotchett's invention were manufactured.' 169 F.Supp. at page 293. After finding that Cotchett's earlier attempts to solve the problems discussed were unsuccessful, and accordingly holding patent No. 2,525,591 invalid as inoperable, the court stated that the '202 patent 'mastered the theretofore intractable difficulties.' Id., at page 295. It concluded, 'Without consideration of other claims, claims 13, 15, 17, 18, 19, 25, 26 and 32 of patent No. 2,644,202 are adjudicated valid. These valid claims were infringed by plaintiff Progressive.' Id., at page 297. Later in the opinion the court said, 'Progressive has infringed claims 13, 15, and 32.' Ibid. The judgment recited, 'Claims 13, 15, 17, 18, 19, 25, 26 and 32 of the Cotchett patent No. 2,644,202 are valid and claims 13, 15 and 32 have been infringed by plaintiff.'1

The court did not need to deal with the question of validity of the '780 and '515 patents because of the estoppel against appellants. Its only finding with respect to infringement was contained in one sentence. 'The evidence shows that the accused Progressive top rolls infringe all the claims of patent No. 2,716,780 and the fourth claim of Cotchett and Swanson No. 2,755,515.' Id., at page 298. Whatever evidence there may have been on this subject, however, is not to be found in the opinion.

Before discussing appellees' '202 patent, it would be well to describe Campbell's patent No. 628,541 issued in 1899, of which it was an improvement. This consisted of two rolls or cylindrical shells revolving on a fixed shaft from which, to reduce friction, they were radially separated by ball-bearing races, hereafter, because of their shape, called cones. There were two rolls, as all top rolls are made in tandem, with a central shaft space between them. Hereinafter, reference to a roll will mean one of the pair. In Campbell the rolls were kept apart by reducing the diameter of the ends of the shaft on which they turned so that the original diameter remained as a shoulder or butt preventing inward movement of the total assembly. There were inner shoulders within the assembly against which the balls could ride, so that axial movement could be prevented with a minimum amount of friction. If the balls were kept too tight against the shoulders, creating what is sometimes called bearing tension, there would be excessive friction and wear. If they were loose, there would be lateral motion, or end play. Excessive end play may cause the rolls to 'wobble,' which has a deleterious effect on the product.2 Even if originally properly adjusted, excessive end play will eventually develop, requiring further adjustment. To care for this Campbell showed a thread on the end of the shaft, along which the outer cone could be adjusted laterally, and fixed by a set screw, so that the amount of end play could be properly regulated when the rolls were assembled at the factory. In order to exclude lint Campbell had an annular indentation or ring outside the shoulder against which the balls ran. This ring in effect provided a double rim or guard inside the assembly, with a grooved space between the rims in which lint which had passed the first rim could accumulate rather than seek to pass the second.

The Campbell construction was not highly successful. Adjustments after the ball-bearing units wore could be made only at the factory. Also, lint would eventually accumulate in such quantities in the annular ring as either to jam against the shell, or to pass by the second guard or shoulder into the race. This, too, required the return of the rolls to the factory, and even there disassembly and reasembly apparently needed skilled personnel. In order to meet these difficulties Cotchett, after considerable experimentation, designed end caps for the shafts with adjustable groub screws (sometimes called by the parties set screws), moving axially to determine the space inside the assembly for axial movement of the cones, and thereby regulating end play. These end caps were easily removable, and the parts taken out for cleaning or lubrication, after which readjustment by means of the groub screw could be affected.3 Although appellees seek to show that this could be done readily in the mill, the great weight of the evidence, including the recitations in the application for appellees' '780 patent, is that mill personnel were usually insufficiently skilled to make this adjustment. The court's finding on this subject is confusing, because it states that the rolls could be disassembled for 'repair and lubrication in the factory and then reassembled in the textile mill.' 169 F.Supp. at page 297. There is no suggestion in the record that this operation was not intended to be a single one, to be performed in one place, wherever that place might be. Nevertheless we are not prepared to disturb the court's finding of invention so far as the adjustable end caps are concerned.4

The lint trap is another matter. This trap differed from Campbell only in that the shape of the annular ring was changed to a Vee, and there was added an oblique shoulder intended to deflect the direction of any entering lint, and hence referred to as a 'tortuous path' trap. The application states that this shoulder would direct lint downward, causing any entering lint to accumulate in the bottom of the Vee, and that when the Vee filled up there would be a 'sealing ring of felted fibers pressing and wiping against the inside surface of the portion of (the) shell roll, which ring catches or excludes all subsequently entering fibers.' It is quite clear that in actual operation this did not happen. The evidence is uniform that lint continued to give trouble, and is in accord with the statement in the application for appellees' later '515 patent that, 'An acute problem is presented by the accumulation of lint and fly at the bearings of shell type top rolls especially those provided with ball bearings. * * * This is well exemplified by the tortuous path type of lint barrier, wherein the first fiber to lodge therein inevitable collects others, and so on at an accelerating rate; the fibers being unable to enter farther pile up and pack until they stop the rotation of the roll shell completely, even though they do keep the fibers out of the bearings.' There was no evidence that this new shoulder had any actual effect, other than from the pattent attorney who described it, or that the trap as a whole performed any better than Campbell's. While '202, over-all, was a more successful roll, in the absence of specific evidence that the trap worked better, this success could be entirely due to the fact that, because of '202's removable end caps, the accumulation of lint was more readily remediable.

It is true, as already stated, that the court held the lint trap claims valid. However, the weight that might otherwise attach to this finding is dissipated by the court's mistaken statement that 'Campbell * * * does not show * * * a grooved lint trap.' 169 F.Supp. at page 297. We hold on the record as a whole, including the fact that Cotchett's trap fell 'acute(ly)' below the representations of the application, that there was no invention over Campbell. The lint trap claims, 17, 18, 19, 25 and 26 are invalid.

While we are on this subject, we also hold the lint trap claims not infringed.5 Appellants' device, hereafter called Progressive, has neither an annular ring nor an annular Vee. Its lint exclusion device takes two different forms, using varying numbers of washers. While appellees' expert stated that these whshers created a 'tortuous path,' we quite agree with the court's observation made in the course of trial that there was no equivalence.6

Nor do we find infringement of the other claims of this patent. Although there are some additional features in '202 which are not found in Campbell (notably that the shaft is of a uniform diameter, with the rolls kept apart by a sleeve or collar, rather than by the larger diameter of the central shaft),7 we hold that appellees' only significant feature is the adjustable end cap. In asserting invention over Campbell, appellees state, 'Campbell had no End-Caps in the sense of Cotchett patent No. 2,644,202. * * * Campbell obtains his only adjustment from his inner cone-- not from any 'end-cap.' The difference is important.' Or, as the court stated in its opinion, 'What was required Cotchett finally realized were threaded caps whereby ready adjustment of all the component parts on the central shaft could be secured * * *.' 169 F.Supp. at page 295. This was the heart of the invention.

The court adopted appellees' contention that appellants' device corresponded with theirs. Thus, in speaking of Progressive, it stated that 'the end caps threaded on the shaft are used to take up the end play and provide manual adjustment of the bearing tension.' 169 F.Supp. at page 297. This, strictly, defines '202. But the first clause describes Progressive only in the loosest sense, and the second clause becomes accurate only if one inserts the words 'access for' before the phrase 'manual adjustment,' providing for an entirely different concept. It is apparent that Progressive's end caps did not themselves provide adjustment. The end caps were removable, but they contained no grub or set screw, and could be replaced in only one position, viz., screwed up tight. The adjustment was effected only from the inside, or, as appellees in their brief differentiatingly described Campbell, 'through the inner cone' (although not in the same way). In Progressive the entire inner space is occupied by two cones8 (sometimes with a sleeve between, but this is immaterial, and will be disregarded), which are pressed hard against each end of the assembly when the end cap is screwed up, and hard against each other. Each cone houses one or two (the diffeerence is here immaterial) ball races. The cones project beyond the raceway, but one end projects further than the other. Centrally, between the cones, is a continuous beveled shoulder, mounted on the inside of the roll shell. The cone projections meet over this shoulder. The shell's freedom of lateral movement, or end play, if any, is contained by the balls on each side of the shoulder. The distance the ball races are apart is determined by the combined length of the inner projections on each cone. Since the cones are asymmetrical, there are three possible variations, obtained by reversing, one or both cones, end for end, either putting both short projections together, or both long, or one of each. The proper spacing is obtained when the cones are originally assembled, and if, because of wear, this changes, the cones can be taken out, and their positions reversed. The total overall length of the cones remains constant. Consequently, unlike Cotchett, adjustment of the ball races is effected without making any change in the position of the end cap, or in the length of the space inside of the assembly.

There was evidence that Progressive was a superior roll to Cotchett's '202 patent. In view of the fact that manufacture of the latter was abandoned after two years, while Progressive is presently in substantial production, this may well have been true. However, it is immaterial. The question is not whether Progressive was successful in achieving a result, but whether it employed the same or equivalent, or employed different means to do so. From the facts stated it is apparent that the mechanical differences in the way in which adjustment is accomplsihed in Progressive, as against the '202 patent, are more marked than the changes by which '202 effected invention over Campbell. 'Through the cone' adjustment is not to be regarded as an 'important difference' from end cap adjustment when appellees are seeking to show invention over Campbell, but no different when they are seeking to show infringement of their device by appellants.

In 1951, Less than two years after Cotchett had filed for No. '202, he and Swanson applied for a patent on a felt washer, or seal, to keep out dust and lint. This application contained eleven claims. The following year Swanson Left appellees' employ. In 1953, five more claims were added. In 1956 patent No. 2,755,515 issued, allowing three of the original claims, and claim 16, as claim 4. All of these, except 4, specifically provide for the washer to be mounted on the ball race, or cone. Claim 4, which is quite long, merely places the washer 'disposed concentrically of the shaft and between the end cap and the rolling elements, and held by engagement with a stationary race member.' If this is to be taken to include a washer not mounted on the cone, Progressive infringes. Appellants say it is not to be read that broadly.

The introductory portion of the application for this patent, after speaking of lint as constituting an 'acute problem,' as mentioned earlier in this opinion, referred to the inadequacy of existing methods, including 'leather gaskets, and other lint barriers on the shaft.' It then stated, 'We have discovered that the objects of this invention can be attained by a non-metallic barrier, such as a felt washer or other yieldable material, mounted on the cone which forms the stationary inner race member for an antifriction bearing * * *. The felt washer is removably locked in position on the cone * * *. Placing the seal on the inner race results in an increase in the ease and speed of assembly because the tapered shape of the supporting cone * * *.' The description continues, using the word 'cone' four more times. There follow the specifications, geared to the illustrations, in the same vein. No alternative construction is in any way suggested. It is true, as appellees point out, that there is the usual claimer of equivalents but such generalizations add nothing to what the law would imply in any event.

A patent must be read as a whole. In view of the full, exact and repetitious description of this device as having a cone-mounted washer, particularly after having disparaged gaskets and other devices on the shaft, we might in all likelihood construe claim 4 as restricted to cones. However, we do not face that bare question. Contesting a rejection of claim 16, now claim 4, for lack of invention, counsel processing the patent stated, 'McNulty's 'sprung' or disked washer g is not on the cone. Every claim in the case specifies this arrangement which adds so greatly to the ease and pseed of assembling.' This obviously meant that appellees were claiming only 'on the cone.' They now argue that the McNulty structure was irrelevant. That is not the point. There is no assignor's estoppel requiring the admission of what the patent expressly rejects. Appellants' washers are not on the cone. There is no infringement.

The final patent is No. 2,716,780. This was for an entire new assembly in which adjustable end caps to limit end play were done away with. Instead, the proposal was to absorb end play by the insertion of 'one or more resilient members serving as a spring * * *. In the preferred form a simple metallic spring washer is seated * * *, the spring bearing against the cone and the end of the sleeve * * *.' In every claim of this patent the 'resilient means' are described as 'urging' the cone, in one direction or another. In other words, the device would be constructed to such a close fit that during assembly the resilient means would be compressed, and would thereafter expand into any free space sufficiently to press the balls against the shoulders, thereby eliminating end play. Then as, with wear, this space increased, the resilient means would further expand, and continue to fill it. In due course appellees discovered that neoprene washers were sufficiently compressible and expandable to take the place of the original 'preferred' metallic spring disk. At the same time the washers could serve as the lint seal, described in the '515 patent. This is its present construction, employing both patents.

Progressive has two constructions, one with a felt washer and a metal washer, and one with a fiber washer and three metal washers, at each end of the inner assembly. Although the fiber washer is not resilient, the felt washer is. Appellees consequently urge that it constitutes 'resilient means.' However, Progressive does not employ this resiliency for the purpose of absorbing end play, or regulating the position of the balls. In all of appellants; rolls the cap is screwed up as tight as possible, jamming the cones solidly end to end.9 Unlike all of appellees' devices, the position of the balls with relation to the center shoulders is determined exclusively by the variable projections of the cones, and is not, and cannot be, affected by any 'urging,' to use the language of the '780 patent, from the resilient means. Under appellants' construction the washer cannot expand laterally. This lateral inaction is illustrated by the fact that in Progressive's other construction, where there are a number of metallic washers, the compression of the cones is transmitted exclusively through them, and not through the fiber washer. Appellants employ the felt washer's resiliency only radially, for the purpose of the lint seal. They use these alternative forms because of varying requirements with regard to lint. Otherwise, the internal operation of both is exactly the same, and in both it is physically impossible for the resiliency of any washer to affect end play. Appellants, accordingly, do not come within the principle of the '780 invention.

Judgment will enter vacating paragraph 2 of the judgment of the District Court, except insofar as that paragraph held valid claims 13, 15 and 32 of Patent No. 2,644,202, to which extent said paragraph is affirmed, and vacating paragraphs 3, 7 and 8 of said judgment, and the case will be remanded to that court for further proceedings not inconsistent with this opinion. Costs on this appeal to appellants.