dissenting.
I respectfully dissent. The district court correctly construed the term “re-movably attached/secured” to mean “re-movably,” not its opposite, as the panel majority holds. The Graco carseat is a permanent assembly whose molded parts are permanently screwed together with six “one-way” screws. Upon unscrewing and disassembly, the Graco seat becomes a collection of loose parts, the cup holder incapable of its function of supporting a cup. The Graco carseat is clearly not a base and seat assembly as in the Dorel patents, whereby the seat can be reversibly slipped from its mooring to carry its infant passenger. Removal of the top part of the Graco carseat (for example, to carry the infant into grandma’s house), would require removing the upholstery, upending the carseat (requiring removal of the entire assembly from its crash-proof installation), and removing the six unremovable screws with the special tool needed for removing unremovable screws — and then repeating the operation in reverse upon returning to the car with the top part and its infant cargo. The panel majority’s redefinition of the patented invention to cover such a carseat is devoid of support.
The Dorel patents include drawings of the patented carseat, showing the separate base carrying the cup holders:
*1048[[Image here]]
The specifications show that the seat and base are distinct structures that are designed to be manually released and reattached, as the district court found. The seat 12, which includes a “seat bottom” and a “seat back,” is described as “releas-ably coupled” to the base 16, which includes a “front,” a “rear,” a “pair of side walls,” and a “cupholder receptor 38.” The term “releasably” is also used to describe the manually adjustable cupholder 18, as “releasably held” and “releasably locked” by tabs 22.
In striking contrast, the Graco carseat is a unitary structure without a removable seat portion:
[[Image here]]
The panel majority has misperceived the Dorel structure, perhaps because the seat of the TurboBooster is molded in two segments, whose engineering drawings are shown in the record as follows:
[[Image here]]
These molded parts are not “removably attached”; they are permanently assembled, using six permanent screws set in recessed indentations. Graco’s “one-way” *1049screws can be unscrewed only with a special tool called the “Un-Do-It” tool. Although the majority states that the tool “can be purchased in hardware stores,” that is not the general experience, and Dorel’s expert testified that he could locate such a tool only after searching the internet for the term “one-way screw.”
The panel majority errs in construing the Dorel claims to a removably attached seat and base, to encompass a unitary structure that does not have a separate and separable seat and base. The Dorel prosecution history itself negates such a construction. The original claims, broadly referring to any “juvenile vehicle seat” with a retractable cup holder, were rejected as anticipated by O’Brien U.S. Patent No. 3,637,184, which describes a unitary seat with a retractable cup holder. In response, the claims were amended to “further comprise” a “seat” that is “remov-ably secured” or “removably attached” to a “base.” The district court correctly construed the claims as requiring a removable seat and separate base. See Phillips v. AWH Corp., 415 F.3d 1303, 1317 (Fed.Cir. 2005) (en banc) (“the prosecution history can [make] the claim scope narrower than it would otherwise be”). Indeed, Dorel in its motion for summary judgment observed that the O’Brien patent “discloses a single, indistinct piece, in which no seat could be identified separately from a base, much less a seat removably attached or removably secured to a base.” Neither Dorel nor the panel majority can recapture what Dorel surrendered during prosecution: a single indistinct piece without an identifiable, removable seat portion.
The district court, referring to the prior art, correctly rejected “the one piece, stand alone base/seat” that is adopted by the panel majority as covered by the Dorel claims. The district court construed “base” to mean a structure that not only “props up the seat” and “upon which the seat sits,” but also is “separate from the seat.” The district court correctly construed “seat” to mean not only a structure “intended to be sat in or on,” but also used “in conjunction with but not as part of the base.” The district court further construed “removably attached/secured” to require that the base and seat be “designed at some time or another to come apart.” On these definitions and the undisputed structure of Graco’s carseat, the district court correctly recognized that no reasonable jury could find that the Graco seat device has a seat that is removably attached to a separate base. As the district court observed, the bottom piece of the Graco assembly “is an integral part of the seat,” not removably attached or secured to the seat.
The panel majority ignores this central issue and remands for findings on a non-issue: whether “the top structure is capable of functioning as a ‘seat’ upon being removed from the bottom structure.” Maj. op. at 1047. That aspect is irrelevant. The question is not whether the top structure is “capable” of being sat upon; the question is whether there is a seat removably attached to a separate and distinct base.
The panel majority also holds as a matter of law that the “removably attached” and “removably secured” limitations are met, based on its construction requiring only that the seat and base “be capable of separation.” Maj. op. at 1046. That is not the description of the invention. Of course even “one-way” screws can be laboriously removed and the structure disassembled; indeed, it is difficult to imagine any assembly not “capable” of being disassembled. The district court correctly required that the seat and base be “affixed in a manner *1050that contemplates” removal; that is, the parts must be “designed at some time or another to come apart.” Summary Judgment at 6-7. The record contains no suggestion that Graco’s carseat is designed to come apart. The user manual for Graco’s product negates disassembly, admonishing the user not to modify the seat or use the seat without all of its parts, warning that failure to properly use the seat “increases the risk of serious injury or death.” Even on the panel majority’s speculation that the Graco structure might be disassembled for cleaning, the “removably attached/secured” requirement of the Dorel claims is not met. The specification describes the base as being “releasably coupled” to the seat; this is not compatible with the need to purchase and use a special tool to unscrew one-way screws covered by upholstery.
The panel majority complains that the patent does not discuss the advantage of separating the Dorel seat from its base, stating that “there is no indication that the patent is directed to a product comprising a base that, during everyday usage, remains in the automobile while the user removes the seat and the child.” Maj. op. at 1046. However, such separable car-seats were in the prior art at the time the Dorel applications were filed; the patents are directed primarily to the cup holders, not the seat and base structure. See ’862 patent, col. 1, ll. 9-10 (describing the invention as a retractable cup holder in a “conventional juvenile vehicle seat” that is “generally known” and in “widespread use”); see also Figures 1 and 2 of Dorel’s patents, supra. The record contains other patents by the inventor of the Dorel patents, such as U.S. Patent No. 6,554,358, which is directed to a seat removably attached to a base, and which cites eleven references to seat and base combinations that are manually separable. In addition, Graco points to Department of Transportation regulations, with which all carseats must conform, that describe carseat components that “can only be removed by use of a tool” as “permanently attached.” 49 C.F.R. §§ 571.213 S5.9, 571.225 S15.1.2.1(f). The panel majority’s interpretation of “removably attached” and “re-movably secured” as requiring only that the structure be “capable” of disassembly, however laborious the mechanics, is contrary to the prior art, contrary to usage in the field, and contrary to federal regulation.
The majority’s approach to claim construction strains this court’s attempts to restore consistency of analysis to patent claims by placing the claims in the context of the specification. See Phillips, 415 F.3d at 1321 (“The risk of systematic over-breadth is greatly reduced if the court instead focuses at the outset on how the patentee used the claim term in the claims, specification, and prosecution history, rather than starting with a broad definition and whittling it down”). The panel majority construes the Dorel claims as encompassing a different invention from that described and prosecuted. From this incorrect claim construction, leading the court to an incorrect decision, I must, respectfully, dissent.