(after stating the facts as above). The crucial question in this case is of infringement, and its answer depends altogether upon whether in operation the liquid passes between the faces of the discs elsewhere than at the grooves. If so, it can do so only by the elasticity of the metal in the spindle. We shall show later that, if the spindle through its inherent elasticity does allow the discs to separate, both the language of the claim and the theory of operation of the disclosure have been infringed, but at the outset the question of physical fact must be considered.
We think that the discs are not open at the outset. While the pressure is first rising, the liquid passes through the grooves, and, as appears below, we are disposed to accept Rautenstrauch’s figures upon the equilibrium at that time between the disc and liquid pressure. However, it is quite clear that the liquid pressure could not increase (the delivery of the pumps being constant), unless there was some diminution in the aggregate discharge openings, and that this can only arise, either from some clogging of the grooves or from the expansion of the metal, making their cubic capacity smaller. The clogging of the grooves is proved by Hancock’s experience and by the, test of November 3, 1915, which does not seem to us to have been conducted with unusually dirty milk.
It is furthermore demonstrated by the impracticability of a solid disc with perforations which has been experimentally tried. Such a disc will operate for a while perfectly, but only for a while. Soon the pressures arise beyond the limit of safety and the machine becomes inoperative. Indeed, the fact is not disputed by the defendant in parts of the record that in operation clogs will occur which obstruct the flow of the liquid and must in some way be relieved. The only *320alternative suggestion in place of the clogging of the grooves anywhere in the record is that the expansion by heating of the metal in the discs closes the grooves and diminishes their cubic capacity. This is in our judgment a negligible factor. The milk is at 110° F. when it comes from the pump, and no one suggests that it is pasteurized, say 140° to 160° F., when it issues. The metal of the grooves must quickly take its temperature from the liquid, and no one has attempted to calculate what a change in temperature of at most 30° to 50° F. would have upon monel metal. We may not speculate upon it without some data; prima facie, we have the right to disregard it. Therefore we must assume that the rise in liquid pressure, which always happens, is only from clogging of the grooves.
Now the universal practice is to reduce the pressure on the discs by backing the wheel as the liquid pressure, too, rises, and this repeatedly till the run is through. If this reduction of pressure on the discs does not noticeably increase the cubic capacity of the grooves, it can have no effect upon the clogs', which by hypothesis have partly closed the grooves. Does it enlarge the grooves themselves, according to Bentley’s hypothesis? We think that the plaintiff’s answer is good to this suggestion. It accepts Rautenstrauch’s figures fot the compressibility of the metal in the discs, and shows that the resulting total compression is substantially less than one-tenth of 1 per cent, at maximum pressure. We must remember, however, that the changes in disc pressure do not relieve all of it by any means, how much we do not know. Now it seems obvious that, if the total expansion in the grooves when all the pressure is removed is less than one-thousandth of the groove* it is the merest assumption to think that the expansion due to the relief caused by backing the wheel will have any substantial effect. We are to suppose that the clogs have stuck in the grooves by their own cohesion and the friction upon the sides. The change in diameter would, theoretically, it is true, relieve that friction, but within admissible limits the relief must be in practice imaginary.
Rejecting, therefore, the theory that the clogs are swept out by any change of diameter in the grooves, it seems to us clear that we have left only the opening of the discs to explain the decrease of the liquid pressures. That pressure varies with the velocity of discharge in capillary orifices, and with the square of' the velocity in larger. We must assume, therefore, with Livermore, that when the pressure drops there has been an addition to the aggregate of tire outlets at least pro-portionate to the drop, or the discharge would not be constant. Without attempting any accurate .computation, we can see that a very slight opening of the discs will accomplish this.
Therefore we have only left the question whether the defendant Ira? proved this to be impossible. Rautenstrauch’s calculations are, so far as we can see, subject to only one exception whose importance we cannot tell. In calculating the pressure tending to extend the spindle, he has not allowed for any pressures upon the faces of the discs. Fie does this avowedly because he says that it begs the question to suppose that they open at all, and so it does. At the outset, and even while the discs are held fast, it would nevertheless seem that some allowance *321should in any case be made for the pressure of the liquid within the grooves; but no one appears to have calculated their area, and so we are forced to omit this factor. Nevertheless the total pressure tending to extend the spindle is given by Rautenstrauch as 6,480 plus 3,436, practically 10,000 pounds. He calculates, moreover, that at its extreme tension the pressure on the discs, which tends to contract the spindle, is about 23,000 pounds, and so he insists it is impossible for the discs to separate. This would, of course, be true, if it were certain that the disc pressure were always so high; but the reasoning fails in application because, although the disc pressure starts at its maximum, the wheel is in practice soon “backed.”
The critical equilibrium is therefore between the consequent pressure on the spindle from the wheel and the pressure from the liquid. No one has told us, nor can any one possibly tell us, what the remaining pressure upon the spindle may be after the wheel is “backed,” and we have no basis for speculation. It is true that the rotation of the wheel is said to be through a small angle, but in common experience we all know that as one tightens a screw thread the final increments of necessary pressure enormously increase for a given angle of rotation. It may easily be possible that the wheel, in being “backed” a few degrees, may cause the disc pressure to fall off below 'the liquid pressure and to allow the discs to separate. We are to remember that the slightest separation of the discs, even if they separate only at one side, may make up for the closing of a large number of radial grooves. We must therefore reject the defendant’s demonstration that the discs cannot separate, as based upon assumptions not capable of verification in practice.
The actual photographs strongly corroborate our a priori conclusion that the discs do separate. For example, it is hard to see how one can account for such a picture as “I,” except upon the theory that the discharge is coming out of the upper right-hand half of the space between two of the discs at any rate, perhaps more. If so, they have been separated at one side. It is true that from the photograph we must suppose that some of the opening must be clogged, because nowhere does an unbroken sheet of liquid issue; but, if we assume the width of the opening to be minute, there is no reason to deny the possibility that in parts of its area there may be clogs. “K” is another such photograph ; though the crack has somewhat changed, “No. 4” is again a significant illustration.
Nor do we find any trouble in accounting for the radial marks upon the flat surface of the discs, though it seems to us an exaggeration to call them “scores.” For a portion of the time, as we have seen, the discs are held tight, and the milk issues only from the grooves. During that period the flat side of the disc is being “scored” by the liquid passing through the groove.' Later, after some of the discs have opened, it does not by any means follow that some do not remain in contact and the “scoring” continues between them. Finally, as we have suggested, it seems possible, perhaps even probable, that, when the discs do separate, they do not float upon a film of liquid, but, on the contrary, that at least at one point remain in contact, minutely tilting as *322it were. Now, while it is true that, assuming them to separate at all, •there can theoretically remain only a single point of contact, yet over a large part of the area of each surface the separation may not be enough to allow the liquid to issue.' Over that portion it will continue to leave through the grooves, and the “scoring” will continue there as well. There is, as we have seen, some evidence in the photographs for assuming that this may in effect be what happens at times.
Furthermore, we have Hancock’s testimony that the plain surfaces of the discs show evidence of impurities caught between them; testimony which, if true, indicates that they have been separated in operation. The defendant’s explanation that these impurities may have adhered, after the discs have been freed and are being taken out, seems to us possible, but not so likely as that which accords with the assumption of separation. While we are not inclined to press unduly upon Willman’s answer to the twenty-ninth cross-interrogatory, it is perhaps fair in this connection to allude to it as a corroboration, if not inadvertent. We believe, therefore, that while the subject is not capable of an absolute demonstration, the balance of the evidence makes enough in the plaintiff’s favor to go¡ beyond mere speculation, and we conclude that the discs do open enough to allow the liquid to flow between their plain surfaces during the “normal” operation of the machine.
We have no hesitation in finding that, this fact proved, the defendant’s machine infringes as it is actually used. Gaulin’s disclosure had for its fundamental feature the yield of the two surfaces between which the liquid flowed. That yield was necessary because in practice it was not possible to insure that the cubic capacity of a rigid exit would be kept constant. In commercial practice the milk cannot be secured which is free enough from contamination for that. Hence, to maintain the necessary capacity of the exits, it was essential that there should be some accommodation for the inevitable clogs. ' Now, it may be possible to devise a machine in which a succession of exits may come into operation so that as the earlier become clogged, more will be opened. A modification of Julien’s machine may serve. If these were all rigid, so that the liquid did not open each in some proportional relation to the pressures, Gaulin’s disclosure would not perhaps be infringed. But so long as the organization of the machine allows for the necessary accommodation to increased pressures through the inherent elasticity of the metals, it seems to us of'no consequence where that elasticity may be. The spindle is in that aspect in every sense an equivalent of the spring, verbally under the claim, and functionally under the disclosure.
Nor do we mean to pass upon the question, here academic as we view it, whether Gaulin’s patent was, or was not, a “pioneer.” It is enough that his claim singled out the yield of the surfaces of pressure and that the defendant cannot proceed without such a yield. Nor do we say whether Gaulin’s invention depended more fundamentally upon the relations between the sectional area and the length of the exits. If the defendant were to make a machine in which in practice the discs did not separate, that question might arise; we say nothing about the *323possible interpretation of the claim or its validity in that case.. It is enough now to hold, as we do, that while the machine is put out in such form that it can, and indeed must, be operated to relieve the disc pressure by letting the discs separate, it infringes. The principle applies which was laid down by this court in Parsons Non-Skid Co., Ltd., v. Atlas Chain Co., 198 Fed. 399, 117 C. C. A. 286, except that that case was much weaker than that at bar, because there the infringing use was only the optimum, while here it is necessary to any commercial practice. Under such circumstances there can be no possible escape from infringement, unless the defendant can make the machine such that in practice the discs will not separate. This litigation would not determine that such a machine would infringe.
We do not think that anything in the prior art deserves extended consideration. All of Julien’s patents were clearly based upon fixed and unyielding orifices of discharge. The supposed action of Julien’s piston to uncover the orifices in part is a mere gloss, and would not anticipate, event if the patent operated as supposed.
Finally, it seems to us of no moment whether Gauliri understood the correct theory of homogenization, and what that theory may be, or whether Julien’s machine could homogenize effectively.
The decree is affirmed, with costs.