Edwards High Vacuum International Limited, owner, and Norton Company, exclusive licensee of the Power patent No. 2,919,061, filed suit in November of 1968 against the defendant Bendix Corporation, seeking damages and injunctive relief for infringement. Defendant denied infringement, claimed invalidity on various grounds, and counterclaimed for alleged infringement of certain of its patents. The issue raised by the counterclaim was severed and has not yet been tried. After trial without jury, the United States District Court for the Western District of New York, in an interlocutory order, held the Power patent to be valid and infringed. The court granted an injunction and ordered an accounting, which was stayed pending appeal. We have jurisdiction pursuant to 28 U.S.C. Sec. 1292(a) (4) (1964). We reverse on the ground that the patent is invalid for indefiniteness.

The Power patent was issued on December 29, 1959. The patent deals with an aspect of a vapor vacuum pump, such as a diffusion pump, but is not a patent for the pump as such. Pumps of this type had been in use for a considerable period of time before the Power patent was issued. These pumps serve to remove air or other gases from vessels such as television picture tubes or camera lens housings, in order to create a vacuum. The pump produces a jet of vapor which creates a "suction" on the inlet side of the pump; that inlet is connected to whatever vessel is being evacuated and the air within the vessel is "drawn out" by the pump and "swept away" by the vapor jet. The heart of the patent at issue relates to cooled baffles which are placed between the vapor jet and the vessel to condense the overdivergent portion of the vapor jet, i. e., that portion of the vapor jet with a vectoral component of velocity in the direction of the pump inlet, and to prevent it from leaking back (back-streaming) into the vessel and contaminating it. These baffles were used in the prior art; this patent is directed to the placement of a baffle (called a guard ring in the claim), in a particular position. The principal claim of the Power patent, claim 1, reads as follows:

"In a vapour vacuum pump, means for producing a vapour jet, a cooled wall surface in the path of said jet for condensing vapour from said jet, said wall having an inlet and an outlet, said outlet being down-stream of the jet, said jet producing means comprising a nozzle assembly which includes a nozzle, said nozzle having a discharge end for discharging a jet of vapour, said discharge end having a lip, a guard ring surrounding the nozzle in closely spaced relation to the lip, and means for cooling the guard ring, said guard ring extending in the direction of the jet a substantial distance beyond said lip and into the jet of vapour, thereby to condense the undesirably over-divergent portion of the vapour jet as it emerges from under the nozzle lip and also to condense molecules emerging at random from under the nozzle lip, with the result that undesirable constituents of the vapour jet are removed almost completely or to a substantial extent from the jet before it leaves the nozzle assembly." (Emphasis added.)It was the object of this positioning to control back-streaming while at the same time not causing any impairment in the rate of pumping, or speed, such as, it was claimed, resulted from the baffles which were formerly used.

The original application for a patent claimed the use of three water cooled guard rings. The inner ring was to extend to a line drawn through the lip of the nozzle and perpendicular to the pump wall; the other two rings were to be placed further into the pump to condense additional portions of the vapor jet. However, faced with the rejection of his patent because of similarity to a device previously patented (the Vogt patent), Power amended his application by deleting the two outer guard rings, deleting a claim that the inner ring extended "slightly below" the nozzle lip, and substituting in the present claim the language italicized above, stating that the remaining ring was "closely spaced" and "a substantial distance" beyond the nozzle lip. Although defendants contend that the phrases in the claim "in the direction of the jet" and "means for cooling the guard ring" are also indefinite, our holding is based on the indefiniteness of the former two terms. Because of this holding it will be unnecessary to detail the state of the prior art and the contentions surrounding the alleged foreign patent and commercial uses prior to obtaining the patent beyond saying that this patent is in a crowded field long a subject of experimentation and invention and, as such, must be subjected to close judicial scrutiny.

The statute requires that the specification contain "a written description of the invention * * * in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains * * * to make and use the same * * *." 35 U.S.C. Sec. 112 (1964 & Supp. 1971). "Indefiniteness is objectionable because the patent does not disclose to the public how the discovery, if there is one, can be made useful and how its infringement may be avoided." Eibel Process Co. v. Minnesota & Ontario Paper Co., 261 U.S. 45, 65, 43 S. Ct. 322, 329, 67 L. Ed. 523 (1923). Thus, Sec. 112 evinces two objectives: first, that in return for the grant of a patent monopoly the patentee must make sufficient disclosure to enable one skilled in the art to reconstruct the invention; and, second, that the patentee describe his invention with sufficient definiteness to enable others to discern the boundaries beyond which experimentation and invention are undertaken at the risk of infringement. United Carbon Co. v. Binney & Smith Co., 317 U.S. 228, 63 S. Ct. 165, 87 L. Ed. 232 (1942); General Electric Co. v. Wabash Appliance Corp., 304 U.S. 364, 58 S. Ct. 899, 82 L. Ed. 1402 (1937). While the patentee's claims here may meet the first objective, they fall short of the second.

The plaintiffs themselves place great emphasis on the words "closely spaced" and "a substantial distance" in distinguishing the Power patent from the prior art. The claim, however, does not "clearly distinguish what is claimed from what went before in the art and clearly circumscribe what is foreclosed from future enterprise." United Carbon Co. v. Binney & Smith Co., 317 U.S. 228, 236, 63 S. Ct. 165, 170 (1942).

The testimony of plaintiffs' witnesses defines these terms in a vague and even contradictory manner. For example, with respect to "closely spaced," plaintiffs' expert witness, Marsbed Hablanian, testified that a distance of "one eighth or one quarter of an inch would be the best execution" for a particular pump. The following series of questions and answers, however, demonstrates the witness's uncertainty:

"Q. Something beyond that would not be closely spaced, isn't that true? A. Well, this is a matter of trying to interpret the word close in a general way, and that is a matter for a linguist to decide * * *.

******

* * *

A. For that size pump, which is a six inch pump the most optimum design in my opinion would have that spacing, one-eighth to a quarter of an inch. It could be somewhat higher.

The Court: You say it could be somewhat higher; what does it mean when you give us a quarter and an eighth.

******

* * *

A. Well, what I mean to say by that if you make it greater it will function not as well * * *.

******

* * *

The Court: Then you said it could go to three-eighths and it wouldn't make much difference.

The Witness: I [sic] would probably make very little difference if you go to three-eighths.

******

* * *

Q. What about one-half an inch, Mr. Hablanian? A. I would consider this size, one-half an inch, almost a border case. Commercially speaking, I would consider it a border case.

******

* * *

Q. Would your judgment be it is closely spaced guard ring as described in the Power patent? A. What do you mean, the half inch point?

Q. Yes. A. I don't really know. Maybe yes, maybe no.

******

* * *

The Court: If a half inch is a borderline, beyond a half inch would be beyond the border?

The Witness: I guess the whole argument loses its meaning beyond that, because if you go halfway in between, then obviously, it isn't close to either thing. Now I don't know whether you want to go to forty-nine per cent and still call it close. I don't know. It is a dictionary definition now. I don't know what close really means."

Instances of this type of uncertainty are numerous in Hablanian's testimony. The inventor himself, Basil Dixon Power, could give no answer that even approached preciseness. The following colloquy is an example:

"Q. So the distance of two-fifths from the lip of the nozzle to the wall of the pump it would begin to function as a guard ring, is that right? A. I don't know * * *. I know when the guard ring is closely spaced in sensible design. I know when it is not closely spaced in sensible design, but you can progress from one to another by infinite relations and we would get into a hazy region where my answer would be confused.

******

* * *

Q. * * * If it is quite long it could be much further away, is that it? A. The guard ring has a function and the description of the guard ring relates to this function." (Emphasis added.)

Similar uncertainty and, indeed, confusion is found in the testimony concerning the requirement that the guard ring extend "a substantial distance" beyond the nozzle lip. Another of plaintiffs' expert witnesses, Arthur Landfors, testified as follows:

"Q. Let me rephrase the question in terms of the guard ring as used in the patent in suit. How far does that have to extend below? A. The same would be true of the guard ring or the cold cap. In either case, I would want it extending a quarter of an inch down.

Q. So an eighth of an inch wouldn't be enough, is that correct? A. For my purpose, I wouldn't consider it enough."

Power, on the other hand, defined the term as a boundary considerably smaller, and thus easier to infringe, than Landfors' version, so small, indeed, that occasional manufacturing tolerances might inadvertently cause a cap in a pump placed above or at the nozzle lip to fall below.

"Q. Now, how far must we extend that cap twenty below the nozzle lip to constitute a guard ring as defined in your patent? A. I think I have answered the question that we have regarded the lip as a boundary and any extension beyond it technically to come into the category of guard ring.

Q. So if it is a thousandths of an inch below it it is a guard ring, is that correct? A. According to what I have just stated.

Q. Is the answer yes? A. I have stated that the answer is yes, but in practice you would have sensible understanding about small tolerance variations and one would understand.

******

* * *

Q. A thousandths of an inch below would fit your definition of a guard ring. A. It would fit the definition if we take the nozzle lip as a criteria."

These and numerous other discrepancies are fatal to plaintiffs' claim. "We think the facts that the applicant himself and the [plaintiffs'] expert witnesses were in disagreement as to the meaning of the limits of the claims as stated, and that there were several commercial compositions falling within the peripheral area of uncertainty, were significant * * *. Especially in the light of this consideration, we think the holding of indefiniteness is required." Helene Curtis Industries, Inc. v. Sales Affiliates, Inc., 233 F.2d 148, 160 (2d Cir.), cert. denied, 352 U.S. 879, 77 S. Ct. 101, 1 L. Ed. 2d 80 (1956). The claim descriptions, even when considered in the light of the diagrams and the full testimony below, fail to inform the public as to what may or may not be manufactured. Particularly where, as here, great emphasis is placed by the plaintiffs themselves on the language used as distinguishing the invention from the prior art, it is our duty to invalidate the claims. "To sustain claims so indefinite as not to give the notice required by the statute would be in direct contravention of the public interest which Congress therein recognized and sought to protect." United Carbon Co. v. Binney & Smith Co., supra, 317 U.S. at 233, 63 S. Ct. at 168. "An invention must be capable of accurate definition, and it must be accurately defined, to be patentable." Id., 317 U.S. at 237, 63 S. Ct. at 170. This invention is not so defined. We therefore reverse the determination of the district court and hold the patent invalid.