The Porter patent also relates to a V-type eight-cylinder engine with a transverse passage for exhaust gases providing a hot spot for the inlet manifold. Each end of the passage communicates directly with a pair of siamesed exhaust valves.

The Welch patent discloses a transverse passage for forming a hot spot between the banks of cylinders of a V-type engine. That passage communicates directly with only a single exhaust valve in each bank of cylinders.

The appealed claimed did not originate in appellant's application. Its history is set forth in the brief for the Commissioner as follows (with a footnote 1 omitted):

*

Claim 11 originated in application Serial No. 407,476, filed by Harold L. Welch on February 1, 1954, now Patent No. 2,926,647 * ** granted March 1, 1960. In that application, the claim was twice rejected * * as "fully met" by the patent to Moseley * Then the Examiner withdrew his rejection and declared an interference * * * with an application of Boris J. Mitchell with that claim as a count. The Welch application was assigned to Chrysler Corporation *** and the Mitchell application was assigned to appellant's assignee, General Motors Corporation ***. Mitchell's assignee moved to substitute appellant's instant reissue application for the Mitchell application in interference ***. This motion was denied on the ground that the count was not supported by the original disclosure of appellant's patent Mitchell, in effect, won the interference 1 * * * whereupon his assignee's attorneys continued the prosecution of the claim in appellant's instant reissue application and cancelled it in the application of Mitchell * * *.

The Examiner held that the disclosure of appellant's reissue application and the original patent disclosure do not support claim 11 * * *. He further reverted to his original position in successively holding the claim "fully met," "met," and "completely anticipated" by the patent to Moseley * *.

*

The Board of Appeals unanimously reversed the Examiner's rejection for lack of supporting disclosure * * * one member later indicating in his dissenting opinion on reconsideration that he had resolved a doubt on the matter in favor of appellant *

The Board unanimously affirmed the Examiner's rejection on prior art of claim 11, agreeing that the claim "reads upon" the Moseley patent * * *. On reconsideration, two members of the Board adhered to the affirmance and one member changed his position and dissented *** in favor of reversal.

The sole issue on appeal is whether the majority of the board erred in affirming the examiner's rejection of the appealed claim as "fully met" and "completely anticipated" by Moseley. The majority, on reconsideration, adhered to its original position on that matter, which position was expressed as follows:

Appellant's essential argument is that, at the time of the Moseley development, the exhausts were normally siamesed, so that, although the illustration is of communication with but a single port in each head, the convention at the time was to have two adjacent exhaust ports communicate with the same transverse exhaust passage such as passage 33 in each head, as shown in Figure 1.

We can concede, arguendo, that such may have been the majority practice at such time. We are concerned here with the disclosure of the Moseley patent. The illustration is clear and clearly teaches communication with a single valve only in each of the exhaust passages 33.

Turning to the specification of this patent we quote the significant part thereof as to the particular question under consideration from column 2, starting with line 4:

"The engine block, generally shown at 25, is of the V-type having separate banks of cylinders 26 and 27 forming the legs of the V. Intake manifold 22 is of general H-shape and lies above the engine block in the crotch of the V. A pair of exhaust valves are shown at 30 and 31 controlling exhaust ports each leading exhaust passages 32 and 33, leading oppositely. In an eight-cylinder engine, there will be four cylinders in each bank and an intake and exhaust valve for each cylinder. The intake connections to the cylinder are not shown, in detail. Upwardly leading exhaust passages 33 connect with an exhaust cross passage 34 which extends beneath and abuts intake riser 17 to form a 'hot spot' for heating intake mixture. Outwardly disposed exhaust passages 32 connect with longitudinal exhaust manifold branches 35. Branches 35 connect with the exhaust ports of the respective banks." [Emphasis added.]

The first emphasized portion has as its clear meaning that each exhaust passage 33 in each head communicates with a single exhaust valve-controlled port, since it recites a pair of valves 30 and 31 connected to a pair of passages 32 and 33. The immediately following sentence (not emphasized above) reinforces the meaning stated for the emphasized sentence, since it mentions the additional exhaust ports and valves not shown. At the end of the above quotation the second emphasized portion brings out the fact that exhaust manifold branches 35 connect with the exhaust ports of the respective banks. Thus, the description in this reference is consonant with the illustration bringing out the fact that the exhaust ports each communicate with a manifold and that the cross passage communicates, as shown, with two exhaust ports.

Thus, although there may have been widespread practice at the time the Moseley patent issued, of having siamesed exhaust ports communicating with the same exhaust channel in the head, which channel in turn communicates with the exhaust manifold, the Moseley disclosure is clear and runs contrary thereto on the basis of either the drawing alone or the drawings and specification read together.

The dissenting member of the board did not find that Moseley shows or describes his transverse passage as communicating with more than one exhaust port in each bank. Rather, he stated:

*** The illustration and description of the Moseley patent create some inference that a connection to one and only one exhaust port is disclosed and taught to those skilled in the art, since generally, patent disclosures are very complete, so that a lack of description of a part gives rise to the inference that the part is not actually present.

He referred to the fact that the examiner allowed the claim in the Welch application after having previously rejected it on Moseley. The dissenting member took the view that the invention of Moseley "has nothing to do with the detailed relation of the cross passage with the exhaust ports" and hence regarded that patent as showing only what was conventional. He further held that it was usual in the

art to connect a cross passage directly to two exhaust ports in each bank of cylinders, referring to the Porter patent, and that a person skilled in the art would therefore understand the Moseley disclosure to embody such an arrangement. He also stated:

The illustration of Moseley and the accompanying description no more anticipate a device having a cross passage communicating with one and only only [sic] exhaust port in each bank than would the side elevational view of a conventional automobile showing two wheels and an accompanying description of the automobile saying that a pair of wheels is shown anticipate an automobile with two and only two wheels.

Appellant bases his arguments on substantially the same reasoning adopted by the dissenting board member. He urges that the Moseley drawing does not limit the cross passage to connection to only a single exhaust port in each bank of cylinders and he disagrees with the majority as to the interpretation of the specification. He further argues that Moseley is most reasonably interpreted as connecting the cross passage between two exhaust ports on each head, apparently because he regards that construction as the usual one in the prior art. Appellant also argues that his interpretation of Moseley is consistent with the structure of the 1950 Cadillac engines, referring to a letter by patent counsel representing Moseley's assignee, submitted in the prosecution of the Welch patent, which letter states that the Moseley drawing was intended to illustrate that engine.

The solicitor contends that the statement in Moseley that a "pair of exhaust valves are shown at 30 and 31 controlling exhaust ports each leading to exhaust passages 32" indicates that each of the two valves 30 and 31 is connected by an exhaust manifold branch 35 and points out that there is no statement or suggestion that more than one exhaust port is connected to branch 35 by any one passage 32. He further urges that it is what Moseley does show that matters and not what he may have intended to show. The solicitor additionally points out that the material relied on to prove the alleged intent of Moseley to show a 1950 Cadillac engine is not supported by affidavits.

It is also urged by the solicitor that, if appellant intended to rely on the state of the prior art to demonstrate that Moseley did not actually disclose the claimed construction, it was incumbent upon him to provide a verified showing by an expert in the art that, at the time of Moseley's filing of his application, the only known V-engine manifold arrangement of the type illustrated by Moseley involved a cross over connected at each end to two exhaust ports.

With reference to the Moseley patent, the majority of the board was of the opinion that the drawings clearly teach that there is a single valve only in each of the exhaust passages 33. We do not agree with that conclusion. The disclosure can be interpreted differently. It is

true that figures 1 and 2 of Moseley could be interpreted to give the impression that each bank contains only one exhaust valve communicating with the transverse passages 33 and 34. On the other hand, if figure 1 is supposed to represent a cross section through the center of the cylinder block, figure 2 in some way should indicate that it represents a construction which contains one valve only in each bank connecting to the transverse passages 33 and 34, or else it is without significance insofar as the determination of this issue is concerned. This is especially true since we have no other drawing in the record to substantiate the Board's position.

Even if figure 1 is not supposed to be a center cross section, its showing still is not inconsistent with there being two valves at each end of the transverse passage, one immediately behind the other.

Assuming, arguendo, that figure 1 of Moseley is a true representation of a section along line 1-1 of figure 2, no valve at all should be shown on the left side of figure 1, which corresponds to the small portion of the cross over passage to the right of the line in figure 2 if there is only one exhaust valve at each end. However, if appellant's contention is correct, a valve should be revealed at this position in figure 1 in this situation, and so it is. Again, we see the board's position, at best, to be doubtful.

It is interesting to note that Porter shows in his plan view two exhaust valves in each bank communicating with his transverse passage and in his figure representing a cross section he shows only one valve per bank, similar to Moseley. It should be remembered that Moseley does not give us the benefit of a view similar to Porter or Welch, which would make certain how many valves communicate with each end of passage 34.

Coming now to Moseley's specification, the significant part reads: A pair of exhaust valves are shown at 30 and 31 controlling exhaust ports each leading to exhaust passages 32 and 33, leading oppositely.

We find that this wording is subject to different interpretations. It could mean, as the board says, that there is one valve at 30 and one at 31 of figure 1, but it might also convey the idea that there are two valves at each port.

[1] The preceding discussion all adds up to the fact that we cannot, with any degree of certainty, ascertain whether Moseley discloses one or two exhaust valves in each bank communicating with the transverse passage. Under these circumstances we cannot agree with the majority of the board that Moseley anticipates applicant's appealed claim. It is well established that an anticipation rejection cannot be predicated on an ambiguous reference. As this court stated, quoting with approval from another decision, in In re Cramblet, 20 CCPA 755, 62 F. (2d) 358, 16 USPQ 74: