Opinion
No. 3096.
February 25, 1936.
Appeal from the District Court of the United States for the District of Massachusetts; Hugh D. McLellan, Judge.
Suit in equity by Warren E. Collins and others against John H. Emerson. From a decree dismissing the bill of complaint ( 10 F. Supp. 885), plaintiffs appeal.
Affirmed in part and reversed and remanded in part.
Nelson Littell, of New York City (Grafton L. Wilson and Hale Dorr, all of Boston, Mass., on the brief), for appellants.
Robert L. Thompson, of Boston, Mass. (Robert Cushman and Roberts, Cushman Woodberry, all of Boston, Mass., on the brief), for appellee.
Before BINGHAM, WILSON, and MORTON, Circuit Judges.
This is a suit in equity for infringement of letters patent No. 1,834,580, applied for July 23, 1929, issued December 1, 1931, to Philip Drinker and others; letters patent No. 1,906,453, applied for November 23, 1931, and issued May 7, 1933, to Philip Drinker and others; and letters patent No. 1,906,844, applied for November 27, 1931, and issued May 2, 1933, to Philip Drinker and others. These patents are now owned by Warren E. Collins, Warren E. Collins, Jr., and Walter G. Chick, individually and as trustees, and the sole licensee is Warren E. Collins, Inc., all of whom are plaintiffs in this action. The issues are validity and infringement.
In the first patent the claims in issue are 1, 2, 3, 6, 7, 8, 9, 10, 11. In the District Court Nos. 1, 2, 3, 6, 8, 9, 10, and 11 of these claims were held invalid for want of invention, unless limited to the preferred pumping means, and, if so limited, they were not infringed, and as to claim 7 calling for "means to control the temperature within said casing including a warming means, cooling means and ventilating means, said housing having gauges whereby said internal conditions may be known" — the gauges being the thermometer 30 and the manometer 34 in figure 1 — the court found that the defendant's apparatus contained a manometer, but no thermometer and no cooling means and did not infringe that claim. The second patent was held invalid as to the claims in issue, for want of invention, and a like holding was made as to the claims in issue in the third patent.
All three patents relate to a device for producing artificial respiration through mechanical means by a method well known since 1876, and are commonly spoken of as artificial respirators. These devices are used for producing artificial respiration in patients unable to breath naturally, as in the case of a sufferer from infantile paralysis whose lung muscles are affected, victims of gas poisoning, electric shocks, drowning, and asphyxia of new born babies. The device is old and simple. It consists of a casing, sufficiently air-tight to hold alternately reduced and increased air pressure, inclosing the body of the patient up to the neck, with his nose and mouth outside the casing exposed to the atmosphere, or having his entire body inclosed within the casing, his mouth and nose being in communication with the outside atmosphere by means of a tube or perforated diaphragm; and an air pump or bellows connected to the casing for producing alternately a partial vacuum (negative pressure) and positive pressure (at least atmospheric pressure) within the casing. When a partial vacuum is produced within the casing, the atmospheric pressure outside forces air through the patient's nose or mouth, expanding his lungs; and when positive pressure is produced the air within the patient's lungs is forced out through the nose and mouth. By alternately creating the condition of negative and positive pressure within the casing at intervals corresponding generally with the normal rate of respiration, artificial respiration is established.
The first patent in suit (No. 1,834,580) is an artificial respirator consisting of a casing 10 having an open end which may be closed by a removable wall section or cover 14, and a sliding bed 18 upon which the patient may be moved into or out of the casing through the opening at the end. The removable wall section or cover 14 has a central aperture in which a rubber collar is fastened by means of a circular wooden rim, and having several bolts inserted through the rim, the rubber collar, and wall section or cover 14. Within the casing is an electric light 32 and a thermometer 30. Outside the casing, but connected with its interior, is a manometer 34 to indicate the pressure within the casing. The pressure within the casing is varied by two pumps 38 and 39, the first being connected to the pipe 40 and the second to the pipe 42. These pipes converge by means of a Y-joint 48 into a single pipe 50 connected with the interior of the casing. One pump creates negative pressure or suction, the other positive pressure. The pumps run continuously at a set speed but are so arranged that they are never both connected at the same time with the Y-joint, and, consequently, alternative negative and positive pressures are created in the casing. The amount of negative or positive pressure is controlled by a valve 70 which varies the size of the air passage in pipe 50. By decreasing the size of the air passage in said pipe, the amount of pressure is decreased. The temperature in the casing "may be raised at any time by means of an electric light 32, or lowered by passing air through a suitable cooling medium."
In the operation of the machine the patient reclines upon the bed with his head extending out through the rubber collar, resting upon a head rest. When negative pressure (partial vacuum) is created in the casing, the patient's lungs are expanded and the atmospheric pressure forces air into his lungs through his nose and mouth. When positive pressure is created within the casing, the patient's lungs are collapsed, forcing the air from the lungs out through his nose and mouth.
The patentees state that the primary object of their invention is "to provide an apparatus which will simulate natural breathing in the patient, both as to the periods of exhalation and inhalation and as to the quantity of air drawn into the lungs of the patient, thus resulting in artificial breathing which will be regular and rythmic." Means are provided for regulating the period of exhalation and inhalation. The quantity of air drawn into the lungs of the patient depends upon the amount of negative pressure in the casing, which may be varied by valve 70. The remaining objects of the device, as stated by the patentees, are to provide an apparatus in which the patient would be at ease during treatment and in which he can breathe naturally, if so inclined, and in which he can talk and sleep without discomfort and without discontinuing the treatment. These objects are accomplished by providing the bed upon which the patient can recline and a rest for his head outside the casing.
Of the claims in issue in the first patent, the plaintiffs state that claim 1 is sufficiently illustrative. That claim reads as follows:
"1. In an apparatus for producing artificial respiration, a casing constructed to receive the major portion of the body of a patient, a bed which is slidable into and out of said casing, said casing having a removable wall section connected to said bed and including an opening through which the head of a patient may be passed, a sealing member positioned in said opening and formed with a restricted opening adapted to fit snugly around the neck of the patient, means for forming a seal between said movable wall and the casing and means for producing alternate variations of pressure within said casing."
The elements embodied in this claim are (1) a casing long and large enough to receive the body of the patient; (2) a slidable bed; (3) a removable end wall section of the casing connected to the bed, the end wall section having an opening for the head of the patient; (4) a sealing member (a pliable rubber collar) positioned in the opening of the end wall section, the sealing member or collar having a restricted opening to fit the neck of the patient; (5) means for sealing the end wall to the casing; and (6) means for producing alternate variations of pressure in the casing.
Claim 2 has all the elements of claim 1, except that the removable wall section or closure is not connected to the slidable bed or carriage. Claim 3 is like claim 1, having its removable wall section or closure connected to the slidable bed or carriage. Claim 6 does not differ from claims 1 and 3 except that it is more restricted and specific. Claim 7 is like claims 1, 3, and 6, except that it is still more restricted by having "means to control the temperature within said casing including a warming means, cooling means and a ventilating means, said housing having gauges whereby said internal conditions may be known." Claim 8 does not differ from claim 2. Claim 9 includes an adjustable head rest co-operating with a flexible rubber sheet or collar. It is broad enough to include an adjustable head rest either separately constructed or attached to the removable wall section. Claim 10 is like claim 9, except that it claims means for clamping the outer edges of the rubber sheet or collar to the removable wall and claims a "head rest." It is broad enough to include a head rest adjustable or nonadjustable, whether constructed separately or attached to the removable wall section. Claim 11 is like claim 10.
In the District Court it was found that claim 7 was not infringed by the defendant's device, as it does not include all of the elements of claim 7. With this conclusion we fully agree.
The District Court also found that adjustable head rests were old, having been in common use in dentists' and barbers' chairs for many years, and inferentially found that nonadjustable head rests were likewise old. The Woillez publication of June 20, 1876, discloses an "appropriate support" (adjustable or nonadjustable) for the head, for use in connection with a respirator for inducing normal breathing; and as adjustable and nonadjustable head rests were old, we do not think it was invention to provide one either for use when attached or not attached to the removable wall of a respirator. See Exhibit 6, Binger and Davis plethysmograph of November 1, 1927; Boyle (1897), U.S. patent No. 592,234; Sauerbruch publication (1904).
The main question in regard to the validity of the claims in issue of the first patent resides in that feature of claims 1, 3, and 6 wherein the removable wall section is connected to the slidable bed or carriage, for unless that feature of those claims involves invention, they are anticipated by the disclosure of the Woillez publication of June 20, 1876. In that publication Woillez disclosed a respirator, which he calls a spirophore, in the following terms:
"This apparatus consists of a zinc or sheet iron cylinder large enough to receive the body of an adult up to the neck. It is equipped with wheels, which permit moving it rapidly to the place where it is necessary. This cylinder set almost horizontal, slightly inclined, is hermetically closed at the foot end and open at the head end. Through this opening at the head end, you slide in the body of the patient by means of a sort of stretcher equipped with rollers, on which he is previously placed; then you close the head opening around his neck by means of a diaphragm that you attach to the edges of the opening. The head thus remaining free rests on an appropriate support. A flexible, impermeable fabric, attached to the cover diaphragm, is secured around the neck or head (from chin to sinciput) to avoid as far as possible the passage of exterior air to the inside of the apparatus, at the moment when the vacuum is produced there.
"The air thus confined in the apparatus around the body of the patient can be partially rapidly withdrawn by means of a powerful aspirator bellows of about 20 litres capacity, situated outside the principal cylinder and actuated by means of a lever. The interior of this pump communicates with the interior of the apparatus through a large tube tightly screwed on.
"Finally, to facilitate experiments, a translucent glass has been placed at the head end of the machine, to allow observation of the functioning of the chest during experimentation; and on top there has been placed a glass tube connecting merely with the inside of the chamber, and in which one can follow the movements of a free rod which should rest perpendicularly on the breastbone of the patient.
"Just as 'in the spiroscope' you were able to see a lung expand when the vacuum was produced inside the tube of fine glass, so the entire chest expands when a corresponding vacuum is created in the cylinder of the new apparatus. The experiments the results of which I am going to show will demonstrate to you very clearly with what ease artificial respiration is produced in such cases and how the respiratory movements take place."
The description of the Woillez apparatus in this publication is so complete that one skilled in the art would have no difficulty in reproducing it. In fact Exhibit 23, here in evidence, is a sketch of the Woillez apparatus from this description, the correctness of which was not questioned at the trial, when it should have been, if at all. The testimony was that it correctly and fairly illustrated the Woillez respirator, and the court so found.
The publication also contained an extended statement of the actual operation of the device when used in connection with the body of a living person, with the body of a dead person where rigor mortis was present, and where rigor mortis was not present, and, in each case, when negative pressure (partial vacuum) was created in the casing, the lungs of the subject were expanded by air rushing through the nose and mouth, and when positive pressure was created in the casing, the subject's lungs contracted forcing the air out, producing artificial inspiration and expiration. It is there pointed out that by depressing the lever of the bellows inspiration is produced and by raising it expiration; that a depression of the lever not only raised the sides and sternum of the dead body but also the epigastrium, which showed that the diaphragm "was involved below and expanded the chest in this direction as it did during life"; and that the rod spoken of in the disclosure as resting upon the middle of the sternum was raised a centimeter at each artificial breath. Furthermore it disclosed that twelve to fourteen litres of air could be introduced within the space of a minute into the depth of the lungs, and that, if artificial respiration was continued ten minutes, 120 to 140 litres of air would be made to circulate in the lungs; that this occurred notwithstanding the lungs of the subject were seriously congested; and that an autopsy disclosed that the "lungs were without the least rupture or tearing, either at their surface or inside." It also described the working of the machine upon the body of a person who died in the last stages of consumption where there was no pronounced rigor mortis and where the air passages were clogged with purulent secretions of mucous as far as the larynx, and states that the respirations produced were powerful enough to make the air penetrate to the depths of the lungs "in spite of the quantity of mucous obstructing the air spaces, since, after three or four breaths, these mucous secretions were driven out noisily in the form of great bubbles."
If the Woillez' description of the collar for preventing the passage of air around the neck (where he says: "A flexible, impermeable fabric, attached to the cover diaphragm, is secured around the neck") is not sufficiently explicit, it is made so in Dr. Knapp's publication in 1904, where he describes the Woillez collar as "rubber material attached around the neck." Woillez' device shows that the iron casing is hermetically closed at the foot end and open at the head end; that there is a sliding stretcher equipped with rollers by which the patient may be slid into the casing up to the neck; that a cover diaphragm or removable wall closing the opening is secured to the casing by suitable means; that a flexible, impermeable fabric of rubber material serves as a collar and seals the opening between the patient's neck and the edges of the hole in the cover diaphragm or removable wall, which produces an air-tight seal; and that alternate variations of pressure within the casing are produced by a powerful bellows.
The Woillez publication discloses every essential element of the claims here in issue, unless it would be invention to connect the cover diaphragm to the sliding bed, as is done in claims 1, 3, and 6; and the only question remaining with relation to the first patent is whether it was invention to connect the diaphragm cover to the sliding bed, as was done in the plaintiffs' device, or whether it was an obvious thing which any mechanic skilled in the art could readily do.
The connection of the diaphragm cover to the bed is of consequence only in getting the patient out of the casing. If that cover is not attached to the sliding bed, on the patient being removed from the casing it would be necessary for the attendant to hold it in position during the period of removal and the obvious thing to do, to relieve the attendant from holding the cover, would be to attach it to the sliding bed. This can hardly be said to be invention and we think it is not.
The Woillez publication was not before the Examiner in the patent office and we cannot believe that the claims here in issue would have been allowed had it been. Furthermore, we agree with the defendant's mechanical expert that no different or other problem is presented in attaching the cover to the sliding bed, so far as getting the patient in and out is concerned, than the problem of mounting the end closure onto the sliding bed or carriage of any therapeutic apparatus, of which a respirator is one. The patent to Boyle, No. 592,234 (1897), disclosed a casing cover attached to a sliding bed to aid in getting the patient in and out of a therapeutic apparatus in exactly the same way as is done in the plaintiffs' device.
The Woillez disclosure cannot be regarded as an abandoned experiment and there is no evidence in the record that it was. On the contrary, Dr. Knapp's publication in 1904 indicates that the Woillez apparatus was then being used, for he describes it as follows:
"Woillez' apparatus consists of an iron cylinder. One end of this is hermetically closed; through the other open end, the apparently dead person is inserted as far as the neck. At this end of the cylinder rubber material is attached around the neck, in such a way that the cylinder becomes as nearly as possible air-tight. The interior of the cylinder is, by means of a hard rubber tube, in communication with a capacious bellows holding 20 litres of air. By powerful pumping movements an alternately reduced and increased air-pressure is obtained in the cylinder. In the former case a strong inspiration results, while in the latter powerful expiration takes place, by which fluid, if present in the bronchi, is forced out through the nose and mouth."
In any event a respirator was disclosed to the world for use and the evidence indicates that it was in fact being used in 1904, some twenty-eight years after the original disclosure and use. But as an anticipation it is immaterial whether the Woillez' apparatus went into commercial use or not before the plaintiffs' device was invented. The printed publication of Woillez' apparatus was given to the public before the plaintiffs' apparatus was invented or discovered and also more than two years prior to the application. This constitutes it an anticipation on the same footing as prior public use and sales. R.S. § 4886, as amended ( 35 U.S.C.A. § 31); Walker on Patents (6th Ed.) § 96, p. 120; Tashjian v. Forderer (C.C.A.) 14 F.2d 414, 415.
The second patent, No. 1,906,453, relates particularly to the collar and its clamping means in any respirator. It provides for "an air-tight seal around the neck of the patient which will be suitably adjustable and comfortable and will not restrict the neck," in connection with a clamping means consisting of a ring and a plurality of clamps. In the specification it is stated:
"The collar 16 is shown as made in the form of a thin flexible sheet which can be drawn taut and held in position by means of a ring 22 and four clamps 23. The clamps may be of any suitable form. As shown, they are supported on headed pins 25 about which they pivot and they are tightened by means of thrust screws 26. A collar of preferred form made of rubber of good quality is shown in cross-section in Fig. 5. The central aperture may be only an inch or two in diameter when the collar is contracted but when it is applied about the neck of the patient it is stretched radially until it fits comfortably but tightly around the neck of the patient and then it is clamped by the ring 22 and clamps 23 in the desired position. By the use of four clamps 23 operating on the ring 22, it is possible to adjust the clamps to hold about three-fourths of the diameter of the collar while still permitting adjustment of the other one-fourth. As shown, the central portion of the collar is relatively thin but not so thin as to be easily torn during the necessary manipulations and the outer portions which are engaged by the clamping ring 22 are somewhat thicker the better to resist damage by the clamping. Immediately around the edge of the central aperture a slight bead 28 is formed to resist tearing. The same type of collar and adjustment thereof is used on adult as well as infant respirators.
"The rubber collar is made large enough so that it can be clamped in eccentric position relative to the aperture in the end wall thereby providing for adjustment along the face of the end wall of the casing to a position corresponding to the position of the body support and head rest."
Of the five claims (1, 3, 4, 5, and 7) relied upon in the second patent, the plaintiffs state that claim 4 is sufficiently illustrative. This claim reads as follows:
"4. In an apparatus for producing artificial respiration, a casing to receive the body of the patient, an opening in one wall thereof through which the head of the patient projects, means to produce periodic variations in pressure within said casing and means to provide a seal for said opening comprising a stretchable rubber collar adapted to fit around the patient's neck and be clamped to the wall of said casing, and clamp means at a plurality of points around the opening in said casing wall to clamp said collar in adjusted and stretched position." (Italics supplied.)
Claim 2 for the beaded edge collar, the thin central portion and the thicker outer margin is not in issue; nor is claim 6 for tilting the body support.
The other claims in issue are of the same general character as claim 4, except that they are not limited to "clamping means at a plurality of points around the opening." They all relate to a stretchable adjustable collar and the means of clamping it to the end wall of the casing and adjusting the opening of the collar comfortably around the patient's neck and relative to the opening in the casing wall.
Before the clamping means provided for in this patent, the plaintiffs found it necessary to supply flexible collars differing in size to fit the necks of different persons, particularly as between adults, children, and babies. Under the first patent they had to supply some three different sizes. Under the second patent they supplied but one, and that having but a comparatively small hole. After that collar is placed around the patient's neck it is stretched to fit comfortably about the neck and clamped in that position by the four point clamping means provided, and thereafter any quarter section of it can be stretched or loosened while the remaining three-fourths of the collar is not disturbed. In this way a single collar was made to fit comfortably about the neck of any patient, and when thus stretched and clamped its center could be made eccentric of the opening in the end wall of the cylinder by successively loosening the clamps and stretching or releasing the stretch in the collar. These are the features to which the claims in issue in this patent, and particularly claim 4, relate.
The question of novelty and patentability in these claims resides in the character of the rubber collar disclosed when used in conjunction with a clamping means having a plurality of clamps likewise disclosed, whereby a collar, having a relatively small opening, may be adjusted to the size of the neck of the patient and also eccentric of the opening in the middle of the removable wall, by releasing a single clamp and readjusting the collar at that point without the necessity of releasing all the clamps and readjusting the collar at all points at the same time.
The District Court appears to have thought that the collar called for in the claims of the second patent was shown and claimed in the first patent, in which the collar is spoken of as a flexible rubber sheet, and in the Drinker and Shaw article of 1929, which shows the collar of the first patent. But the collar of the first patent was not made to meet the problem contemplated by the collar of these claims in its use, adjustment and readjustment to the neck of the patient and eccentric of the opening in the end wall, in connection with the clamping means there called for. The District Court was also of the opinion that the clamping means feature of these claims was anticipated by the locking or clamping means shown in Binger and Davis plethysmograph (Exhibit 6), and held that these claims were invalid.
The plethysmograph, with its collar and clamping means, was unquestionably shown to be in public use as early as November, 1927, something over three years prior to November 25, 1931, when the second patent was applied for. Whether the locking and clamping means there shown anticipated the clamping means disclosed in the second patent (No. 1,906,453) depends upon whether they are in the same or a legal equivalent, that is, function in the same way, and produce the same result, so that the device would be an infringement if later in time. The locking and clamping device shown in the Binger and Davis plethysmograph was not made to clamp a rubber collar to the flat side of an end wall, either in a respirator or plethysmograph. It was made to be used, and can only be used, in connection with a flange made to fit tightly into the opening in an end wall, the flange at its outer end having a groove extending around its circumference into which the tongue on the inner edge of the locking device fits, and locks a flexible collar when drawn over the opening in the flange and between the groove in the flange and the tongue on the locking device. The problem confronting Binger and Davis in providing their clamping means was quite different from that presented to the patentee in the second patent. Furthermore the tests made before the court demonstrated that the Binger and Davis clamping means operate differently from those in issue in this patent. In the case of their clamping means, when the ring is locked, every portion of the rubber collar is locked and held in position and when it is partially unlocked every portion of the collar is equally released at the same time. It is true that when partially released the collar may be stretched, but to do this successfully and hold in position the portion of the collar not desired to be disturbed, it is necessary for the operator to use one of his hands to press the ring of the clamp down against the flange so as to hold it there, and the portion of the collar between the groove in the flange and its tongue on the inner edge of the ring, not being adequately released, is liable to be torn and broken when the operator attempts to stretch it. We do not think that such a device anticipates the clamping means of these claims, or would infringe them.
The idea of using a rubber collar, with a small aperture that could be stretched and adjusted by clamping means to fit comfortably the neck of a large person or small babe, and made sufficiently large so that the center of its aperture could be adjusted eccentrically of the opening in the end wall, with clamping means that could readily be used by any one to make the different adjustments, was new with the patentees of the plaintiffs' device, and we think involved invention.
The defendant has used in his respirator a stretchable, adjustable rubber collar with a small aperture, and made large enough so that it can be clamped eccentrically of the opening in the end wall, and has clamped it in substantially the same way as was done in the plaintiffs' device. In one of the alleged infringing devices the ring is cut into four parts, to each of which a clamping means is attached. This is undoubtedly an infringement of the claims of the second patent. The ring in the other device (Exhibit E) is divided into two parts, each part having clamping means located diametrically opposite one another. We think this also is an infringement.
The alleged invention in the third patent (1,906,844) relates to the use of portholes in the casing of a respirator, the portholes being equipped with rubber collars to make a seal around the attendant's arm when inserted into the respirator, permitting the manipulation of the patient without stopping the operation of the respirator. The claims in issue are 1, 2, 4, 6, 7, 8, 9, 10, 13, 14, and 15. Claims 2 and 4 are said to be sufficiently illustrative and read as follows:
"2. An apparatus for producing artificial respiration comprising a substantially air-tight casing to receive the body of a patient with his head projecting out of said casing, means for periodically varying the pressure within the casing, and means permitting access of an operator's hands for manipulation within the casing comprising a plurality of ports having apertures adapted to receive the arms of the operator, means in said apertures to form an air-tight seal around the operator's arms when inserted therein, and means to seal said apertures in the plane of the casing wall when the operator's arms are not inserted therein."
"4. An apparatus for producing artificial respiration comprising a substantially air-tight casing to receive the body of a patient with his head projecting out of said casing, means for periodically varying the pressure within the casing, and means permitting access of an operator's hand for manipulation within the casing comprising an arm port, sealing means within said port adapted to form a substantially air-tight seal around the arm of an operator when inserted therein, a second port adapted to permit the insertion of a bed pan or other article into the casing, said port being so positioned relative to the arm port as to permit the operator to grasp in his hand an article introduced through the second port, a cover closing said second port to form an air-tight seal and means to seal said arm port when the operator's arm is withdrawn."
The elements involved in these claims differ only in one respect. The fourth claim calls for a second port hole for the insertion of a bed pan or other article into the casing. Both claims call for a plurality of ports for the insertion of the operator's hands and arms into the casing. The arm ports are provided with flexible collars to seal the port when the operator's arms are inserted; they are also provided with a flap or cover or both to close the opening in the collar when the operator's arms are removed. The so-called "second port" in the fourth claim also has a cover for closing that port.
The Binger and Davis plethysmograph of the prior art had an arm port, permitting the operator's arm to be inserted within the casing for manipulating the patient, provided with a collar for sealing the aperture when the arm was inserted, and provided with closing means to seal the porthole when the arm was removed. In fact it had been known for years how to insert a hole in a casing and provide means for sealing the same when the arm of the operator or the neck of a patient was inserted through the hole. It surely could not be invention to provide a plurality of arm ports, or to make an opening for a bed pan. We fail to see wherein there is any invention in this patent.
The decree of the District Court is affirmed as to the first (No. 1,834,580) and third (No. 1,906,844) patents. It is reversed as to the second patent (No. 1,906,453) and the case is remanded to that court for further proceedings not inconsistent with this opinion.
I concur as to the second and third patents. As to the first patent, I reach the opposite conclusion.
The changes from the crude apparatus of Woillez to the practical and useful respirator of the plaintiff's patents seem to me to have involved invention. The important steps in this development were (1) fitting the casing or chamber with a flange adapted to take a gasket, (2) the rigid, separate end piece having in its center portion a flexible neckpiece and adapted on its edges to fit the gasket on the casing and be locked by clamps so that it could be made air tight and quickly fastened and detached, (3) uniting this rigid end piece to the sliding bed so that when the bed was pushed into the chamber the end piece was brought into position to lock quickly, and (4) placing on the rigid end piece a head rest to support the patient's head. Invention is undefinable, a matter of feeling rather than of logic; but it seems to me that these changes were not merely separate mechanical improvements in existing apparatus, but were related steps in the construction of something much better, which the inventor had in mind, and which was certainly a very great improvement over anything that had preceded it, the first practical respirator for regular use. It does not seem to me a sufficient answer to the plaintiffs' claim of invention to say that each of these changes considered separately was obvious. They may be so, and still the combination of changes necessary to bring about the desired result may be far from obvious. It is not without significance that the clever technicians of the Rockefeller Institute in making a similar apparatus did not see the patentee's solution of the problem presented, and made something much clumsier and less efficient which, considering the air pressure necessary in a respirator, probably could not have been used as one in a practical way, because of the difficulty in keeping such long joints sufficiently air tight. I therefore think the first claim of the first patent in suit was valid and infringed.