dissenting.
The majority in this case affirms the holding by the International Trade Commission (Commission) that claims 1-5, 7, and 9 of U.S. Patent No. 5,372,392 ('392 patent) would have been obvious over the SMC prior-art clamp. I dissent because I conclude that the claims would not have been obvious to one of skill in the art in light of the SMC prior-art clamp. In particular, I cannot agree that the SMC prior-art clamp is “a four-sided, generally rectangular clamp.” The Final Initial Determination by the Administrative Law Judge in this case held that “An examination of [the SMC prior-art clamp] shows that the connector is not generally rectangular, and is not even four sided.” J.A. 321. I agree completely with the determinations by this ALJ and find that the Commission’s findings to the contrary are not supported by substantial evidence.
Claim 1, the only independent claim of the '392 patent, recites “a generally rectangular ported flange ” and “a four-sided, generally rectangular clamp adapted, in its operative clamping position, to engage, in parallel relationship with one another, [a] pair of ported flanges.... ” '392 patent claim 1 (emphasis added). In the prior appeal we held that SMC’s ported flange was generally rectangular despite having sloped corners. The picture below depicts the SMC ported flange that we held was generally rectangular:
[[Image here]]
Intervenors’ Br. 13.
The claim language at issue on remand was not “generally rectangular ported flange”; it was “a four-sided, generally rectangular clamp.” The ALJ was correct that the focus must be on “the precise claim language”, J.A. 320, namely “a four-sided, generally rectangular clamp.” It is clear from figure 1 of the patent what a four-sided generally rectangular clamp is:
*1329[[Image here]]
'392 patent fig.l (red lines added). As illustrated in the patent, the clamp is generally rectangular and has four sides. The sides labeled la, lb, and lc define three sides of the clamp. The side labeled 3 is the fourth side of the clamp and, when the side is pivoted into operative clamping position, the clamp is generally rectangular. The same is true of the commercial embodiment of the '392 patent, the Excelon Quikclamp®:
[[Image here]]
See, e.g., Norgren Excelon Quikclamp, http://store.norgren.com/us/en/list/airlinepreparation-frl/accessories/frl-connectors/ (red lines added).
We must look to the precise claim language and be ever vigilant not to strip away patent rights by eliminating claim limitations. This claim requires the clamp to be both “generally rectangular” and “four-sided” in structure. Hence, while something with six sides (like the SMC ported flange pictured above) might be generally rectangular, that does not mean it is “four-sided.” The claim does not require that the flange be four-sided, but it does require that the clamp be “four-sided.” The claim also recites the functional limitation that the clamp must be “adapted, in its operative clamping position, to engage ... the pair of ported flanges.”
Hence the claims require that the clamp is: (1) four-sided; (2) generally rectangular; and (3) adapted in its operative position to engage the flanges. These are three distinct requirements, three distinct claim limitations. The ALJ clearly understood this. The majority, however, im*1330properly combines the structural and functional claim limitations to conclude that only the operative portion of the clamp must be four-sided and generally rectangular. This interpretation conflicts with the plain language of the claim, which requires that the clamp as a whole be “four-sided.” The majority recognizes this requirement: “this court recognizes — consistent with our claim construction in Norgren I — that the clamp as a whole, and not merely portions of it, must be four-sided and generally rectangular.” Maj. Op. at 1323. In application, however, the majority does the opposite. It holds: “the four sides of the old-style SMC connector are those required for its function and that the shape formed by those sides is generally rectangular.” Maj. Op. at 1324-25 (emphasis added); see also Maj. Op. at 1325 (“The four outer ‘sides ’ of the oldstyle SMC connector are labeled A-D.... ” (emphasis added)). The plain language and our prior case are clear — the whole clamp must be four-sided and generally rectangular, not just a portion of the clamp. It is not fair for the majority to cherry-pick four sides on a sixteen-sided clamp and then call it a four-sided clamp: how can this be clear and convincing evidence of invalidity?
As seen in the photo below (adapted from Intervenors’ Br. 13), the SMC prior-art clamp has sixteen sides:
[[Image here]]
The ALJ found that “An examination of RPX-002 shows that the connector is not generally rectangular, and is not even four-sided.” J.A. 321. The Commission rejected the ALJ’s findings, instead concluding that the SMC prior art clamp was “generally rectangular.” Though it is certainly true that the claims do not require a perfect rectangle, I do not agree with the Commission that the SMC prior art clamp is generally rectangular. However, given that I review the Commission fact findings for substantial evidence, if the claim limitation was only a generally rectangular clamp, I would not reverse on this point. But the claim limitation is NOT “a generally rectangular clamp.” It is a “four-sided, generally rectangular clamp.” I cannot agree that a sixteen-sided clamp is four-sided. It defies logic. And the fact that an expert testifies that a sixteen-sided clamp is four-sided doesn’t make it so. I refuse to strip a patentee of his patent rights based on such testimony.
*1331The “four-sided, generally rectangular clamp” limitation is different from both the “generally rectangular flange” limitation and the “adapted, in its operative clamping position” limitation. We should not treat the four-sided limitation as superfluous or read it out of the claim. Because the Commission’s fact finding that the SMC prior art clamp meets the four-sided limitation is not supported by substantial evidence, I would reverse the determination of obviousness.