OPINION
SMITH, Circuit Judge.The Recess Appointments Clause in the Constitution provides that “[t]he President shall have Power to fill up all Vacancies that may happen during the Recess of the Senate, by granting Commissions which shall expire at the End of their next Session.” U.S. Const, art. II, § 2, cl. 3. The central question in this case is the meaning of “the Recess of the Senate,” which is the only time in which the president may use his power to recess appoint officers. Three definitions have been offered: (1) breaks between sessions of the Senate *208(i.e., “intersession breaks”); (2) these intersession breaks as well as breaks within a session (i.e., “intrasession breaks”) that last for a non-negligible time, or (3) any break in Senate business that makes the body unavailable to provide advice and consent on the president’s nominations. This is a difficult question that has never been addressed by our Court or the Supreme Court. We hold that “the Recess of the Senate” in the Recess Appointments Clause refers to only intersession breaks. As a consequence, we conclude that the National Labor Relations Board panel below lacked the requisite number of members to exercise the Board’s authority because one panel member was invalidly appointed during an intrasession break. We will therefore vacate the Board’s orders.
I
New Vista operates a nursing and rehabilitative care center in Newark, New Jersey. On January 25, 2011, a healthcare workers’ union petitioned the National Labor Relations Board (“the Board”) for. certification as the representative for New Vista’s licensed practical nurses (“LPN”). New Vista opposed this certification on the grounds that its LPNs are supervisors who cannot unionize under the National Labor Relations Act (“NLRA”), 29 U.S.C. § 152(3), (11). See NLRB v. Kentucky River Cmty. Care, Inc., 532 U.S. 706, 709, 121 S.Ct. 1861, 149 L.Ed.2d 939 (2001) (explaining that supervisors do not fall within the NLRA’s definition of a bargaining unit). On March 9, 2011, the Board’s regional director determined that New Vista’s LPNs were not supervisors and thus certified the union as well as ordered an election. New Vista appealed to the Board, which affirmed the regional director’s order.
The union won a majority in the ensuing election. New Vista refused to bargain with the union,1 which then filed a charge of unfair labor practices against New Vista before the Board. On behalf of the union, the Board’s general counsel moved for summary judgment against New Vista, which New Vista opposed. The Board unanimously granted summary judgment in favor of the Union and against New Vista in a “decision and order” dated August 26, 2011.
This order was issued by a three-member “delegee group” of the Board. The NLRA establishes that the Board is composed of up to five members, appointed by the president and confirmed with the advice and consent of the Senate. 29 U.S.C. § 153(a). Section 153(b) authorizes the Board to “delegate to any group of three or more members any or all of the powers which it may itself exercise.” Id. § 153(b). These delegee groups must “maintain a membership of three in order to exercise the delegated authority of the Board.” New Process Steel, L.P. v. NLRB, 560 U.S. 674, 130 S.Ct. 2635, 2644, 177 L.Ed.2d 162 (2010).
Importantly, this three-member-composition requirement is distinct from § 153(b)’s quorum requirements. The quorum requirements speak to the number of members who must be present to exercise the Board’s powers for either the *209Board itself or a properly constituted three-member (or more) delegee group. See id. at 2642-43 (explaining that the “group quorum provision” “authorizes two members to act as a ... group of at least three members” but does not “authorize two members to constitute a valid delegee group”); see also id. at 2642 (defining quorum as “the number of members of a larger body that must participate for the valid transaction of business”). To have a quorum, a delegee group must have at least two of its three members present and the Board must have at least three of its five members present. 29 U.S.C. § 153(b).
In contrast, the three-member-composition requirement speaks to how many members are required for a delegee group to be a properly constituted body that can exercise the Board’s powers. These different requirements are certainly related, but this case simply turns on whether the delegee group that issued the August 26 Order and the subsequent reconsideration orders had three members.
On September 7, 2011, New Vista filed a motion with the Board to reconsider the August 26 Order. The company argued that the three-member delegee group acted ultra vires because although the order is dated August 26 — one day before one member, Wilma Liebman, resigned — it was not issued until it was mailed during the week of August 29. This would mean, according to New Vista, that the panel had only two members when the order was issued, thereby violating 29 U.S.C. § 153(d)’s three-member-composition requirement. The company also argued that the August 26 Order was substantively incorrect. Meanwhile, on September 13, 2011, the Board filed with this Court an application for enforcement of the August 26 Order. We granted an uncontested motion to hold in abeyance the filing of the administrative record pending resolution of the motion for reconsideration. This functionally acted as a stay of the proceedings before us.
On December 30, 2011, the Board denied New Vista’s motion for reconsideration. New Vista took two actions. First, it filed a second motion for reconsideration on January 3, 2012. In this motion, the company argued that the three-member December 30 delegee group was improperly constituted and thus without power to issue the order because one of the panelists was recused from the case. The company also argued in a March 14 “further motion for reconsideration” that the December 30 Reconsideration Order delegee group was improperly constituted because one of the panelists was a recess appointee whose term concluded at the end of the Senate’s 2011 session — which New Vista contended was December 17, 2007, thirteen days before the December 30 Reconsideration Order was issued.
Second, on January 9, 2012, New Vista filed a petition for review of the December 30 Reconsideration Order with this Court. We have treated this petition as a cross-petition for review opposing the Board’s petition for enforcement of the August 26 Order. We also granted another Board motion to hold in abeyance the filing of the administrative record for these petitions until New Vista’s second motion for reconsideration was resolved. This, again, functionally acted as a stay of the proceedings before us.
On March 15, 2012, the Board denied New Vista’s second motion for reconsideration. This order did not address the company’s March 14 argument that the term of one panelist had ended on December 17. On March 22, 2012, New Vista filed a third motion for reconsideration. This motion reiterated the company’s March 14 argument that the December 30 delegee group was improperly constituted *210because the Senate’s session had ended on December 17. The motion also argued that the three-member delegee group that issued the March 15 Reconsideration Order lacked three members because two of its members were invalidly appointed to the Board under the Recess Appointments Clause while the Senate was not in “recess.” In sum, New Vista argued that if the Senate’s session had ended when it began using pro forma sessions, then the December 30 panel had only two members because the term of one of its members expired. But if the Senate’s session did not end at that time, then the March 15 panel was improperly constituted because the president’s recess appointments were invalidly made while the Senate was not in recess. The Board denied this motion on March 27, 2012. The Board also filed the administrative record with this Court on that date, thereby stripping itself of jurisdiction. See 29 U.S.C. § 160(e) (“Upon the filing of the record with it the jurisdiction of the court shall be exclusive and its judgment and decree shall be final.”).
On April 4, 2012, New Vista filed a petition for review of the March 15 and March 27 Reconsideration Orders. We granted New Vista’s request that this petition be consolidated with New Vista’s earlier petition for review for all purposes. These consolidated petitions for review are collectively a-cross petition opposing the Board’s petition for enforcement of the August 26 Order.
II
We consider sua sponte whether the delegee group that issued the August 26 Order had jurisdiction. See Bender v. Williamsport Area Sch. Dist., 475 U.S. 534, 541, 106 S.Ct. 1326, 89 L.Ed.2d 501 (1986) (explaining that “every federal appellate court has a special obligation to ‘satisfy itself not only of its own jurisdiction,, but also that of the lower courts in a cause under review,’ eyen though the parties are prepared to concede it” (quoting Mitchell v. Maurer, 293 U.S. 237, 244, 55 S.Ct. 162, 79 L.Ed. 338 (1934))). In their initial briefs, the parties contended that the delegee group had subject-matter jurisdiction under 29 U.S.C. § 160(a), which “empower[s]” the Board (and its three-member delegee groups) “to prevent any person from engaging in any unfair labor practice ... affecting commerce:” We do not doubt that § 160(a) provides one jurisdictional requirement for the Board to adjudicate a case. But that does not preclude others. We have thus inquired whether 29 U.S.C. § 153(b)’s three-member-composition requirement is jurisdictional. We hold that it is.
This Court has previously explained that “the overall authority of the Board to hear [a] case under the NLRA” is a jurisdictional question that “ ‘may be raised at any time.’ ” NLRB v. Konig, 79 F.3d 354, 360 (3d Cir.1996) (quoting NLRB v. Peyton Fritton Stores, Inc., 336 F.2d 769, 770 (10th Cir.1964)); see also Polynesian Cultural Center, Inc. v. NLRB, 582 F.2d 467, 472 (9th Cir.1978). Under § 153(b) and New Process Steel, delegee groups of the Board do not have statutory authority to act if they have fewer than three members. New Process Steel, 130 S.Ct. at 2644; Teamsters Local Union No. 52S v. NLRB, 624 F.3d 1321, 1322 (10th Cir.2010) (holding that a “two-member NLRB group that issued the order in this case lacked statutory authority to act ” (emphasis added)). The three-member-composition requirement is thus jurisdictional because it goes to the Board’s authority “to hear [a] case under the NLRA.” Konig, 79 F.3d at 360.
Nevertheless, the Supreme Court “has endeavored in recent years to ‘bring some discipline’ to the use of the *211term ‘jurisdictional.’ ” Gonzalez v. Thaler, — U.S. -, 132 S.Ct. 641, 648, 181 L.Ed.2d 619 (2012) (quoting Henderson v. Shinseki — U.S. -, 131 S.Ct. 1197, 1202-03, 179 L.Ed.2d 159 (2011)). So there may be reason to believe that Ko-nig’s analysis and the subsequent jurisdictional conclusion for this case are no longer valid. Lebanon Farms Disposal, Inc. v. Cnty. of Lebanon, 538 F.3d 241, 249 n. 16 (3d Cir.2008) (explaining that “[a]n intervening decision of the Supreme Court is a sufficient basis for us to overrule a prior panel’s opinion without referring the case for an en banc decision”). Our review of the Court’s recent clarification shows that Konig remains good law and that the three-member-composition requirement is jurisdictional. The Court has explained that jurisdiction “refers to a court’s adjudicatory authority.” Reed Elsevier, Inc. v. Muchnick, 559 U.S. 154, 130 S.Ct. 1237, 1243, 176 L.Ed.2d 18 (2010) (quoting Kontrick v. Ryan, 540 U.S. 443, 455, 124 S.Ct. 906, 157 L.Ed.2d 867 (2004)). Subject-matter jurisdiction “refers to ‘the courts’ statutory or constitutional power to adjudicate the case.’ ” Id. (quoting Steel Co. v. Citizens for Better Env’t, 523 U.S. 83, 89, 118 S.Ct. 1003, 140 L.Ed.2d 210 (1998) (emphasis in original)).
Although these statements refer to Article III courts, jurisdictional issues are just as important for administrative adjudicative bodies. “It is well settled that an administrative agency,” like an Article III court, “is a tribunal of limited jurisdiction.” Pentheny Ltd. v. Gov’t of Virgin Islands, 360 F.2d 786, 790 (3d Cir.1966). An administrative agency “may exercise only the powers granted by the statute reposing power in it.” Id.; see also 2 Am.Jur.2d Administrative Law § 282 (2013) (“Administrative agencies are tribunals of limited jurisdiction.... As a general rule, agencies have only such adjudicatory jurisdiction as is conferred on them by statute.”). These powers are limited by the scope of the jurisdictional statute in the same way that a federal court’s powers are limited by the Constitution and statute. Compare 2 Am.Jur.2d Administrative Law § 282, with Exxon Mobil Corp. v. Allapattah Servs., Inc., 545 U.S. 546, 552, 125 S.Ct. 2611, 162 L.Ed.2d 502 (2005) (“The district courts of the United States, as we have said many times, are ‘courts of limited jurisdiction. They possess only that power authorized by Constitution and statute.’ ” (quoting Kokkonen v. Guardian Life Ins. Co. of Am., 511 U.S. 375, 377, 114 S.Ct. 1673, 128 L.Ed.2d 391 (1994))). The fact that this case deals with an administrative agency does not eliminate the requirement that a delegee group satisfy all jurisdictional requirements before it may exercise the Board’s powers.
In Henderson v. Shinseki the Supreme Court stated that “a rule should not be referred to as jurisdictional unless it governs a court’s adjudicatory capacity, that is, its subject-matter or personal jurisdiction.” 131 S.Ct. at 1202. As noted, subject-matter jurisdiction is “statutory or constitutional power to adjudicate the case.” Steel Co., 523 U.S. at 89, 118 S.Ct. 1003 (emphasis in original). Furthermore, in Arbaugh v. Y & H Corp., 546 U.S. 500, 126 S.Ct. 1235, 163 L.Ed.2d 1097 (2006), the Supreme Court provided a “readily administrable bright line” rule: “If the Legislature clearly states that a threshold limitation on a statute’s scope shall count as jurisdictional, then courts and litigants will be duly instructed and will not be left to wrestle with the issue.” Id. at 515-16, 126 S.Ct. 1235. “But when Congress does not rank a statutory limitation on coverage as jurisdictional, courts should treat the restriction as nonjurisdictional in character.” Id. at 516, 126 S.Ct. 1235. “Congress, of course, need not use magic words *212in order to speak clearly on this point,” so context can show that a requirement is jurisdictional. Henderson, 131 S.Ct. at 1203.
The Supreme Court’s recent decision in New Process Steel indicates that § 153(b)’s three-member-composition requirement is jurisdictional. In that case, the Board had delegated its power to a three-member delegee group. Three days after the delegation became effective, the term expired for one of the three members of the delegated group. This left the group with only two members. 130 S.Ct. at 2638-39. The Supreme Court held that § 153(b)’s three-member-composition requirement meant that the “two remaining Board members cannot exercise” the authority of the Board. Id. at 2638, 2644 (“We thus hold that the delegation clause requires that a delegee group maintain a membership of three in order to exercise the delegated authority of the Board.”). The presence of three Board members in a delegee group is a necessary condition for the Board to exercise its power to adjudicate a matter before it.
New Process Steel renders the three-member-composition requirement “a threshold limitation” on the scope of the power delegated to the Board by the NLRA: the Board cannot exercise its power through a delegee group if that group has fewer than three members. This statutory mandate is therefore jurisdictional. See Arbaugh, 546 U.S. at 515, 126 S.Ct. 1235 (explaining that “threshold limitation[s] on a statute’s scope” imposed by Congress are jurisdictional); Teamsters Local Union No. 523, 624 F.3d at 1322 (holding that a “two-member NLRB group that issued the order in this case lacked statutory authority to act ” (emphasis added)). By explaining that three members are required “in order to exercise the delegated authority of the Board,” New Process Steel, 130 S.Ct. at 2644, the Supreme Court has in essence declared that the three-member-composition requirement goes directly to the board’s “power to hear a case,” which is exactly what jurisdictional questions relate to. United States v. Cotton, 535 U.S. 625, 630, 122 S.Ct. 1781, 152 L.Ed.2d 860 (2002); see also Noel Canning v. NLRB, 705 F.3d 490, 497 (D.C.Cir. 2013) (“[T]he objections before us concerning lack of a quorum raise questions that go to the very power of the Board to act.”).2
The Board relies on three cases3 as authority providing that “a claim that a *213federal officer was appointed unconstitutionally is not a jurisdictional challenge.” NLRB Ltr. Br. at 2 (Feb. 28, 2013) (citing Freytag v. Commissioner of Internal Revenue, 501 U.S. 868, 111 S.Ct. 2631, 115 L.Ed.2d 764 (1991); Intercollegiate Broad. Sys., Inc. v. Copyright Royalty Bd., 574 F.3d 748 (D.C.Cir.2009); Evans v. Stephens, 387 F.3d 1220 (11th Cir.2004)) (en banc) (emphasis in original). These cases hold that Appointments Clause challenges are nonjurisdietional when brought independently. Freytag, 501 U.S. at 878-79, 111 S.Ct. 2631; Intercollegiate Broad. Sys., 574 F.3d at 755-56; Evans, 387 F.3d at 1222 n. 1. Those holdings are not relevant to the jurisdictional conclusion we reach today. We do not hold that challenges under the Appointments or Recess Appointments Clauses are jurisdictional. We instead hold that the NLRA’s three-member-composition requirement is jurisdictional and must be met before the Board can exercise its power over a case. Because this requirement is jurisdictional, any reason for which the delegee group consists of fewer than three members— including whether one member is invalidly appointed under the Recess Appointments Clause — can be raised by a party or by this Court at any point in litigation as a jurisdictional defect. See Henderson, 131 S.Ct. at 1202.
The jurisdictional nature of the three-member-composition requirement is especially important in this case because it requires us to analyze whether Craig Becker — one of the three-member delegee group that decided the August 26 Order — held a valid appointment under the Recess Appointments Clause. This question is distinct from the recess-appointments question initially briefed by the parties. The parties’ briefs address whether Richard Griffin and Sharon Block — who were members of the delegee group that decided the March 15 and March 27 Reconsideration Orders — were invalidly recess appointed because their January 9, 2012 appointments were made while the Senate was holding so-called pro forma sessions.4 Member Becker was not appointed when the Senate was holding pro forma sessions but, instead, was appointed on March 27, 2010, one day after the Senate “adjourn[ed]” for two weeks. 156 Cong. Rec. S2180 (daily ed. Mar. 26, 2010) (statement of Sen. Kaufman) (reporting Senator Ted Kaufman’s motion for and the Senate’s unanimous consent of the body being “adjourned until Monday April 12, 2010 at 2 p.m.”). As will be seen in Part V, this means that our consideration of Member Becker’s appointment entails evaluation of at least one more definition of “recess” than the evaluation of Members Griffin and Block’s appointments. Before delving into the difficult constitutional task of defining “recess,” however, we must first address two preliminary questions: whether the delegee group that issued the August 26 Order lacked three members as a result of Chairman Liebman’s resignation and whether the definition of recess is a non-justiciable political question.
Ill
‘We have a longstanding practice of avoiding constitutional questions in cases where we can reach a decision upon other grounds.” Egolf v. Witmer, 526 F.3d 104, 109 (3d Cir.2008). That practice leads us first to consider New Vista’s non-constitutional argument that the August 26 Order was issued by a delegee group of fewer than three members. New Vista *214contends that one of the three members resigned before the order was issued. The delegee group that issued the order consisted of Chairman Liebman, Member Becker, and Member Hayes. The face of the order is dated August 26, 2011. New Vista Nursing & Rehab., 357 N.L.R.B. No. 69 (Aug. 26, 2011). The Board docket also reflects August 26, 2011 as the date that the order was issued. New Vista Nursing & Rehab., NLRB No. 22-CA-029988 (Aug. 26, 2011), http://www.nlrb.gov/case/22-CA-029988. On August 27, Chairman Liebman resigned. New Vista argues that the order was actually entered after Liebman resigned because the order “was mailed, received by the Regional Board Agent, and was posted on the Board’s Summary of Decisions Website on August 31, 2012.” Pet’r’s Br. at 31. The Board does not dispute that the order was mailed to interested parties after August 27 but contends that the order was issued on August 26— the date that appears on the face of the order.
“Agency action is entitled to a presumption of regularity.” Frisby v. U.S. Dep’t of Hous. & Urban Dev., 755 F.2d 1052, 1055 (3d Cir.1985). “Acts done by a public officer which presuppose the existence of other acts to make them legally operative, are presumptive proofs of the latter.” R.H. Steams Co. of Boston, Mass. v. United States, 291 U.S. 54, 63, 54 S.Ct. 325, 78 L.Ed. 647 (1934). Here, the act done was the issuance of the August 26 Order, which presupposes that the members listed as having made the decision did in fact make that decision. The issuance of the order creates a presumption that all three members listed on the order decided it. See id. It is New Vista’s burden to rebut that presumption.
New Vista offers only a single piece of evidence in rebuttal: that the order was not mailed until after August 26. This is insufficient, and Braniff Airways, Inc. v. Civil Aeronautics Bd., 379 F.2d 453 (D.C.Cir.1967), demonstrates why that is so. In that case, Braniff Airways argued that the Civil Aeronautics Board lacked a quorum because one of its members had resigned before the order was issued. Id. at 459. The order in that case was issued on June 1, the same day the member in question resigned. The Court found that the Board had a quorum solely on the basis that the order “on its face indicated that it was concurred in and signed on June 1, 1965.” Id. The Court reached that conclusion despite payroll records with conflicting accounts, one of which showed that the member was on the payroll only through May 31, 1965. Id. Notably, the Court also discounted that the order “was not served until June 2,” on the basis that “[i]n [their] view it is plain that once all members have voted on an award and caused it to be issued the order is not nullified because of incapacity, intervening before the ministerial act of service, of a member needed for a quorum.” Id. (emphasis added).
The D.C. Circuit’s reasoning is equally persuasive here. The only evidence New Vista puts forth is that the order was mailed after it was dated and posted on the docket. This falls short even of what Braniff Airways presented. It relied not only on a delay in service but also on payroll records. New Vista presents even weaker grounds to doubt the order’s date than Braniff offered the D.C. Circuit. New Vista cannot overcome the presumption of regularity.
New Vista also argues that it is entitled to seek further discovery into when the members voted on the August 26 Order. The company acknowledges, however, that “the NLRB may not be required to enter for the record the time, place, and content of their deliberations,” Pet’r’s Br. at 53, *215and the Board has stated that the minutes sought do not exist, Resp. Br. at 29. Yet New Vista persists, asserting “that the record of the time of their votes on agency-actions under review is essential to determine” the validity of the August 26 Order. Pet’r’s Br. at 53. The company fails to explain why the date listed on the order itself is not evidence “of the time of their vote.” Absent a reason to doubt the date listed, the presumption of regularity requires that we consider the date as the record of when the delegee group caused the opinion to be issued, which presupposes that they voted on or before that date. Accordingly, New Vista has failed to show that one of the members resigned prior to the issuance of the August 26 Order.
IV
The amicus argues that we should decline to define the word “recess” within the Recess Appointments Clause because it is a nonjusticiable political question. “Questions of justiciability are distinct from questions of jurisdiction, and a court with jurisdiction over a claim should nonetheless decline to adjudicate it if it is not justiciable.” Gross v. German Found. Indus. Initiative, 456 F.3d 363, 376 (3d Cir.2006) (citing Baker v. Carr, 369 U.S. 186, 198, 82 S.Ct. 691, 7 L.Ed.2d 663 (1962)). An issue presents a nonjusticiable political question when one of the following characteristics is “inextricable from the case”:
a textually demonstrable constitutional commitment of the issue to a coordinate political department; or a lack of judicially discoverable and manageable standards for resolving it; or the impossibility of deciding without an initial policy determination of a kind clearly for nonjudicial discretion; or the impossibility of a court’s undertaking independent resolution without expressing lack of the respect due coordinate branches of government; or an unusual need for unquestioning adherence to a political decision already made; or the potentiality of embarrassment from multifarious pronouncements by various departments on one question.
Baker, 369 U.S. at 217, 82 S.Ct. 691. Ami-cus’s principal contentions are that the recess-appointments claim by New Vista is nonjusticiable because (1) “‘the issue is textually committed’ to the president,” Amicus Br. at 4 (quoting Nixon v. United States, 506 U.S. 224, 228, 113 S.Ct. 732, 122 L.Ed.2d 1 (1993)), and (2) there are “no ‘manageable standards’ to solve the partisan argument between the Executive and Congress ... regarding dysfunctional Senate confirmation processes,” id.5 Neither argument is persuasive.
*216Nothing in the language of the Recess Appointments Clause textually commits to the president the task of defining “recess.” The Clause states that “[t]he President shall have Power to fill up all Vacancies that may happen during the Recess of the Senate, by granting Commissions which shall expire at the End of their next Session.” U.S. Const, art. II, § 2, cl. 3. This language lacks the explicit assignment of power to any one branch, such as the assignment found in the Constitution’s Impeachment Trial Clause which states that “[t]he Senate shall have the sole Power to try all Impeachments.” U.S. Const. art. I, § 3, cl. 6 (emphasis added); Nixon, 506 U.S. at 228-35, 113 S.Ct. 732 (concluding that the explicit assignment, along with drafting history indicating that the assignment was intentional, meant that the power to try impeachments was textually committed to the Senate). The Recess Appointments Clause also does not contain an imperative to either branch to craft a rule regarding the meaning of recess — or, more broadly, when the president may use his recess appointments power. The Clause is thus also distinguishable from the Naturalization Clause’s grant to Congress of the authority to “establish an uniform Rule of Naturalization.” U.S. Const, art. I, § 8, cl. 4; New Jersey v. United States, 91 F.3d 463, 469 (3d Cir.1996) (stating that this Clause represents a textual commitment to Congress).6
Finally, the Clause does not provide unqualified power to either the Senate or the president that would suggest it makes a textual commitment to either. It limits the president’s recess-appointment power by requiring that the Senate be in recess, and it limits the Senate’s ordinary advice- and-consent power by eliminating that power while the Senate is in recess. The Clause thus cannot be read to invariably favor one branch’s interests in such a way that it makes a textual commitment to one of them. See Freytag, 501 U.S. at 880, 111 S.Ct. 2631 (“Because it articulates a limiting principle, the Appointments Clause does not always serve the Executive’s interests.”); Ryder v. United States, 515 U.S. 177, 182, 115 S.Ct. 2031, 132 L.Ed.2d 136 (1995) (“The [Appointments] Clause is a bulwark against one branch aggrandizing its power at the expense of another branch, but it is more: it ‘preserves another aspect of the Constitution’s structural integrity by preventing the diffusion of the appointment power.’” (quoting Freytag, 501 U.S. at 878, 111 S.Ct. 2631)); The Federalist No. 76 (Alexander Hamilton) (explaining the Constitution’s rejection of unitary power in either the president or the Senate in favor of one that divides power between them).
The amicus disputes this, arguing that the Clause makes a textual commitment by providing the president “unilateral appointment authority when the Senate *217[is] unavailable to render its advisory consent vote.” Amicus Br. at 12. This argument reveals the tendency of the political-question doctrine “to obscure the need for case by case inquiry.” Gross, 456 F.3d at 377-78 (quoting Baker, 369 U.S. at 210-11, 82 S.Ct. 691). We have cautioned against this tendency, instructing that our inquiry must “avoid ‘resolution by any semantic cataloguing,’ ” and must instead “undertake a ‘discriminating inquiry into the precise facts and posture of the particular case.’ ” Id. (quoting Baker, 369 U.S. at 217, 82 S.Ct. 691). The amicus’s argument runs afoul of our instruction because it merges the issue present in this case (when the president can use his recess-appointments power) with an issue not in this case (how the president can use that power). The amicus’s characterization of the power speaks to both issues: it states how the president can use his recess-appointment power (“unilateral authority”) and assumes the answer to the question in this case of when he can use that power (“when the Senate [is] unavailable to render its advisory consent vote”). The greater power the president has during a recess does not shed light on what the word “recess” means or who decides what it does mean and thus does not provide a reason to conclude that the Clause makes a, textual commitment to the president. Cf. INS v. Chadha, 462 U.S. 919, 940-41, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983) (explaining that Congress’ plenary authority over immigration does not immunize it from judicial review for violations of other constitutional restrictions on its power committed while exercising that authority).
The amicus’s concerns regarding the lack of judicially manageable standards for defining “the Recess of the Senate” are similarly unfounded. There are several judicially manageable standards for defining “the Recess of the Senate” and, correspondingly, for when the president may use his recess-appointments power. The parties present two different standards: according to New Vista, any time after both houses have agreed to adjourn for more than three days, Pet’r’s Br. at 40-41, and according to the Board, any time the Senate is not available to conduct regular business, Resp. Br. at 44. Cf. Zivotofsky, 132 S.Ct. at 1428-30 (relying on the “detailed legal arguments” provided by the parties regarding whether the statute at issue was constitutional to show the existence of judicially manageable standards). The D.C. Circuit has provided another: intersession breaks that follow adjournments sine die of the Senate. Noel Canning, 705 F.3d at 506-07. Of these standards, those provided by the D.C. Circuit and New Vista are judicially manageable because they rely on regular procedures employed in the Senate and found in the Senate’s record. The Board’s more open-ended definition of recess might very well be unmanageable because it does not rely on any particular Senate procedure and would require judicial “exploration] [of] communications between the Senate Minority and the president” in addition to review of the “scheduling schemes of the Senate Minority and House Majority.” Amicus Br. at 20-24 (arguing, after rejecting the standard offered by New Vista, that the Board’s standard is unmanageable). But this only cautions against selecting the Board’s standard rather than showing that there are no judicially manageable standards available.
Of course, if the question is framed — as the amicus has — as a need to derive a judicially manageable standard “to resolve [ ] the underlying cycles of partisan confirmation obstruction payback which caused the NLRB vacancies,” Amicus Br. at 25, then there is likely no judicially manageable standard. See also Evans, 387 F.3d at 1227 (rejecting as nonjusticiable an argu*218ment that the president unconstitutionally-used the recess-appointment power because the appointee had been previously rejected by the Senate and thus constituted a circumvention of the Senate’s advice and consent role). But that is not the question we face. Instead, we must define the phrase “the Recess of the Senate,” which is a question distinct from resolving the “cycles of partisan confirmation obstruction payback.” See id. at 1224-26, 1227 (defining recess to include intrasession breaks despite holding that the political argument made was nonjusticiable).
This task falls within the “ ‘province and duty of the judicial department to say what the law is.’ ” Zivotosky, 132 S.Ct. at 1427-28 (quoting Marbury v. Madison, 5 U.S. (1 Crunch) 137, 177, 2 L.Ed. 60 (1803)). This “duty will sometimes involve the ‘Resolution of litigation challenging the constitutional authority of one of the three branches,’ but courts cannot avoid their responsibility merely ‘because the issues have political implications.’ ” Id. at 1428 (quoting Chadha, 462 U.S. at 943, 103 S.Ct. 2764) (alteration in original). Thus, “the fact that the resolution of the merits of a case would have ‘significant political overtones does not automatically invoke the political question doctrine.’ ” Khouzam v. Att’y Gen., 549 F.3d 235, 249-50 (3d Cir.2008) (quoting Chadha, 462 U.S. at 942-43, 103 S.Ct. 2764). That the issue presented here touches on political events of the day is not dispositive of whether this case presents a nonjusticiable question. Because there are manageable standards and because the Clause does not make a textual commitment to the Senate or the president, we hold that interpreting the phrase “the Recess of the Senate” is a justiciable question.
V
Having determined that the Recess Appointments question is justicia-ble, we now begin our analysis of the recess-appointment issue. Member Becker is the only member of the delegee group that issued the August 26 Order who was recess appointed and thus the only one whose appointment is in question. As noted, he was appointed during an intrasession break that began on March 26, 2010, and ended on April 12, 2010. This break lasted seventeen days and the Senate was indisputably not open for business. His appointment will be invalid if the Recess Appointments Clause does not empower presidents to make recess appointments during these types of breaks.
The Clause provides that “[t]he President shall have Power to fill up all Vacancies that may happen during the Recess of the Senate, by granting Commissions which shall expire at the End of their next Session.” U.S. Const, art. II, § 2, cl. 3. This is understood to allow the president to use his recess appointment power only “during the Recess of the Senate,” thereby rendering the definition of recess, along with its temporal reach, of pivotal consequence to the controversy now before us. See Noel Canning, 705 F.3d at 499-500; Evans, 387 F.3d at 1224. Three possible definitions have been presented. The D.C. Circuit defines the term to mean only intersession breaks, which are “the period between sessions of the Senate when the Senate is by definition not in session and therefore unavailable.” Noel Canning, 705 F.3d at 499-500, 506. The end of a session is typically demarcated by a particular type of Senate adjournment — an adjournment sine die — which is the procedure used to end a Senate session. Id. at 512-13.7 An intersession break is the peri*219od between an adjournment sine die and the start of the next session. David EL Carpenter et al., Cong. Research Serv., R42323, President Obama’s January k, 2012, Recess Appointments: Legal Issues 4 n. 23 (2012).
A second definition, one which the Eleventh Circuit has adopted, is that recess includes intersession breaks as well as some “intrasession” breaks, which are breaks in Senate business during a session. Evans, 387 F.3d at 1224. An intra-session break is demarked by a Senate adjournment of any type — other than adjournment sine die — and lasts until the next time the Senate convenes, which is set by the motion to adjourn. See, e.g., Cong. Rec. S2180 (daily ed. Mar. 26, 2010) (statement of Sen. Kaufman) (reporting Senator Kaufman’s March 26, 2010 motion for and the Senate’s unanimous consent of the body being “adjourned until Monday April 12, 2010 at 2 p.m.”). From 1921 until recently, there was a consensus that an intrasession break was not “the Recess of the Senate” unless the break lasted for a non-negligible number of days. The first attorney general to adopt this view suggested that the minimum duration was ten days. 33 U.S. Op. Att’y Gen. 20, 24-25 (1921) (rejecting the proposition that “an adjournment for 5 or even 10 days can be said to constitute the recess intended by the Constitution,” but advising the president that a break of 28 days is within the meaning of recess). All presidents, at least in practice, followed this ten-day minimum until January 2012. Carpenter et al, supra, at 15 & n. 97 (stating that no presidents until 2012 made a recess appointment during an intrasession break shorter than ten days). Accordingly, the second definition includes only those intrasession breaks that last for a significant duration, which historically has been ten days or more.8
The third and final possible definition is of more recent vintage. In January 2012, President Barack Obama made several recess appointments while the Senate was holding pro forma sessions every three or four days. These sessions are considered recesses under the third definition. Pro forma sessions are formal meetings of the Senate in which usually only one Senator is present to convene the body briefly before adjourning it until the next pro forma session. Id. at 2; see also, e.g., 157 Cong. Rec. S8787 (daily ed. Dec. 20, 2011) (statement of Sen. Warner) (recording Senator Mark Warner’s convening and adjournment of the Senate in a span of thirty-five seconds). Before such sessions are held, the Senate agrees by unanimous consent that there will be “no business conducted” except business that was previously agreed to, such as convening a new session of the Senate. See, e.g., 157 Cong. Rec. S8783-84 (daily ed. Dec. 17, 2011) (statement of Sen. *220Wyden) (recording the schedule of pro for-ma sessions to be held between December 17, 2011 and January 23, 2012). However, these consent agreements can, and have been, subsequently altered to allow initially unplanned business — including the passing of legislation' — during a pro forma session. See, e.g., 157 Cong. Rec. S8789 (daily ed. Dec. 23, 2011) (statement of Sen. Reid) (obtaining unanimous consent that a bill “be considered read three times and passed” if an identical version is passed by the House, which the House subsequently did, during a pro forma session); see also Carpenter et al., supra, at 18 & n. 108. Importantly, these sessions prevent the Senate from being adjourned for more than three or four days at a time, which means the adjournment never reaches the ten-day minimum discussed above. See, e.g., 157 Cong. Rec. S8784 (daily ed. Dec. 17, 2011) (statement of Sen. Wyden) (recording Senator Ron Wyden’s motion, and the Senate’s unanimous concurrence therewith, that the Senate be “adjourned until Tuesday, December 20, 2011, at 11 a.m.”); 157 Cong. Rec. S8787 (daily ed. Dec. 20, 2011) (statement of Sen. Warner) (recording the Senate’s adjournment “until Friday, December 23, 2011, at 9:30 a.m.”).
The third definition of recess, which is offered by the Board, allows the president to make recess appointments while the Senate is holding these proforma sessions. The Board argues that a recess occurs when “the Senate is not open to conduct business” and thus unavailable to “pro-vid[e] advice and consent on nominations.” Resp. Br. at 44. The Board argues that this definition follows from Attorney General Harry Daugherty’s 1921 opinion, which adopted a partially functionalist definition of “the Recess of the Senate”:
[T]he essential inquiry, it seems to me, is this: Is the adjournment of such duration that the members of the Senate owe no duty of attendance? Is its chamber empty? Is the Senate absent so that it can not [sic] receive communications from the President or participate as a body in making appointments?
33 U.S. Op. Att’y Gen. at 25. The Board contends that these criteria decide whether the Senate is open to conduct business and available to provide its advice and consent. Unlike Attorney General Daugherty’s opinion, the Board appears to consider these criteria controlling in themselves, such that there is no requirement for a minimum, non-negligible period of time to pass in order for the Senate to be in recess.9 Id.
Based on these criteria, the Board contends that periods in which the Senate holds pro forma sessions only constitute a recess. This is because during these sessions, the body is neither doing business nor available to provide its advice and consent. This means, per the third definition, that these sessions do not interrupt *221what would otherwise be an intrasession break that begins with the adjournment before the first proforma session and lasts until the next convening of the Senate in a non -proforma session.
In sum, the parties argue that “the Recess of the Senate” has one of three meanings: (1) intersession breaks; (2) intersession and intrasession breaks that last a non-negligible period, which has historically been ten days (“long intrasession breaks” hereinafter); or (3) any time in which the Senate is not open for business and is unavailable to provide its advice and consent.10 We hold that “the Recess of the Senate” means only intersession breaks, and so we conclude that Member Becker’s appointment was invalid.
A. “[T]he Recess of the Senate”
1. The Literal Meaning of Recess
When interpreting the Constitution, “we begin with its text.” City of Boerne v. Flores, 521 U.S. 507, 519, 117 S.Ct. 2157, 138 L.Ed.2d 624 (1997). In doing so, “we are guided by the principle that ‘[t]he Constitution was written to be understood by the voters; its words and phrases were used in their normal and ordinary as distinguished from technical meaning.’ ” District of Columbia v. Heller, 554 U.S. 570, 576, 128 S.Ct. 2783, 171 L.Ed.2d 637 (2008) (quoting United States v. Sprague, 282 U.S. 716, 731, 51 S.Ct. 220, 75 L.Ed. 640 (1931)). The “[njormal meaning may of course include an idiomatic meaning, but it excludes secret or technical meanings that would not have been known to ordinary citizens in the founding generation.” Id. at 576-77, 128 S.Ct. 2783.
The word “recess” lacks a natural meaning that clearly identifies whether it includes only intersession breaks or also includes intrasession breaks, whether they be of a certain duration or a period of unavailability. Dictionaries from the time of ratification provide definitions that can be read to support any of these definitions. Samuel Johnson’s dictionary defines recess to mean “[rjetirement; retreat; withdrawing; secession” as well as “[djeparture” and “[rjemoval to distance.” Samuel Johnson, 2 A Dictionary of the English Language 469 (6th ed.1785).11 *222All of these definitions contain some connotation of permanence or, at least, longevity. “Secession,” for example, means “[t]he act of departing” or “[t]he act of withdrawing from councils or actions.” Id. at 589; see also Merriam-Webster’s Collegiate Dictionary 1121 (11th ed.2003) (defining “secession” to mean “formal withdrawal from an organization”). And “departure” is defined by Johnson to mean “[a] going away,” the “[djeath; ... the act of leaving the present state of existence,” and “an abandoning.” Samuel Johnson, 1 A Dictionary of the English Language 568 (6th ed.1785); see also Memamr-Webster’s at 334 (defining “departure” to mean “the act or an instance of departing,” “a setting out (on a new course)”); 1 William Blackstone, Commentaries *187-88 (defining one method of terminating parliamentary business, the dissolution, as “the civil death of the parliament”). The implication of permanence supports an understanding of recess to mean only intersession breaks because these are followed by an adjournment sine .die, which are adjournments without a set date for reconvening. And the implication of longevity supports the idea that recess includes long intrasession breaks.
Neither of these implications is consistent with the Board’s unavailable-for-business definition of recess, but other entries in Johnson’s dictionary provide some support for that definition. Johnson’s definition of recess includes “[rjemission or suspension of any procedure.” Johnson, 2 A Dictiona'i’y of the English Language at 469. And, of course, words such as “departure” also have less permanent implications than death. Johnson, 1 A Dictionary of the English Language at 568 (defining “departure” as “[a] going away”). The term “recess,” by itself, thus lacks a literal meaning that unambiguously supports one of the three definitions.
2. The Historical Use of Recess
Importantly, though, the Constitution does not say only “Recess.” Rather, it limits the president’s recess-appointments power to the “Recess of the Senate.” The words “of the Senate” provide some context for our analysis: parliamentary procedure at the time of ratification. Deal v. United States, 508 U.S. 129, 132, 113 S.Ct. 1993, 124 L.Ed.2d 44 (1993) (“[T]he meaning of a word cannot be determined in isolation, but must be drawn from the context in which it is used.”).
American colonial legislatures and the first Senate largely derived their parliamentary procedures from the procedures used by the English Parliament. See Henry M. Robert III, et al., Robert’s Rules of Order: Newly Revised xxxiv-xxxv (11th ed.2011) (recounting the migration of English procedures to the American colonies); Thomas Jefferson, A Manual of Parliamentary Practice: For the Use of the Senate of the United States (2d ed.1812) (relying heavily on English precedents in providing procedural rules for the Senate). English parliamentary procedure at the time had three types of breaks: adjournments, which were “continuances of the session from one day to another ... and sometimes a fortnight or a month together”; prorogations, which were “continuances of the parliament from one session to another” initiated by the king; and dissolutions, which were terminations of a Parliament initiated by the king’s order, his death, or a length of time that necessitated new elections before another Parliament could be convened. 1 William Blackstone, Commentaries *186-89; see also Jefferson, supra, § 51 at 164-65; Michael B. Rappaport, The Original Meaning of the Recess Appointments Clause, 52 U.C.L.A. L.Rev. 1487, 1550-51 (2005). The Parliament thus had three *223breaks: adjournments for intrasession breaks and prorogations as well as dissolu-tions for intersession breaks.
At first blush, these three types of breaks appear to correspond with the three mechanisms for breaks referred to in our Constitution. “Adjournment,” or its verbal form “adjourn,” is the same phrase the Constitution uses to denote day-to-day and longer breaks within sessions of either chamber. U.S. Const, art. I, § 5, cl. 1 (allowing a minority of members to “adjourn from day to day”); id. art. I, § 5, cl. 4 (requiring concurrence between both chambers if, “during the session of Congress,” they are to “adjourn for more than three days”).12 The word “dissolution” does not appear in the Constitution, probably because the president does not have the power to dissolve Congress. See id at art. II, § 3 (providing that the president, at most, “may adjourn [Congress] to such time as he shall think proper” if they cannot agree on “the time of adjournment”); The Federalist No. 69 (Alexander Hamilton) (explaining the powers of the president and how they are less than those of the king and even the governor of New York by contrasting the president’s power to “only adjourn the national legislature” with the “British monarche[’s]” power to “prorogue or even dissolve the Parliament”). But the concept of dissolution is still present in the Constitution: Congress is automatically dissolved — and any ongoing session ended — every two years by termination of the terms of one-third of Senators and all members of the House. U.S. Const, art. I, § 2, cl. 1; id. art. I, § 3, els. 1-2. These dissolutions end a session and, following elections, begin another session in a new Congress, see Jefferson, supra, § 51 at 166 (“A dissolution certainly closes one session; and the meeting of the new Congress begins another.”) — just as the king’s dissolution, or the dissolution by the passage of time, did for the English Parliament, 1 William Blackstone, Commentaries *189.
In light of these parallels, it is tempting to say that “Recess of the Senate” corresponds with prorogations and thus must refer only to terminations of sessions and the intersession breaks that follow them. But this argument proves too much. Even though the Constitution uses “adjournment” to mean breaks within a session, it also uses the term to mean breaks between sessions. The Supreme Court held in the Pocket Veto Case, 279 U.S. 655, 49 S.Ct. 463, 73 L.Ed. 894 (1929), that “adjournment” in Article I, § 7, clause 2 of the Constitution is any break in business “that prevents the President from returning the bill to the House in which it originated within the time allowed.” Id. at 680, 49 S.Ct. 463 (internal quotation marks omitted); see also U.S. Const, art. I, § 7, cl. 2 (providing that a bill passed by Congress becomes law after ten days after presentment to the president “unless the Congress by their adjournment prevent its return”). This definition does not distin*224guish between breaks within sessions and those between sessions. See id.; accord Rappaport, supra, at 1551 n. 198 (explaining that “the Framers used the term ‘adjournment’ with a broader meaning than it had traditionally under English law”). This means that the Constitution does not simply adopt “adjournment” as it was used in Parliament and correspondingly suggests that “Recess of the Senate” is not simply prorogation by another name.
Understanding the differences between prorogation and adjournment is helpful, however, to make sense of ratification-era state constitutions.13 Eight of these constitutions use the word “recess.” Six contain the same ambiguity found in the federal Constitution.14 The word “recess” in the Massachusetts and New Hampshire constitutions, however, includes only intersession breaks. See Rappaport, supra, at 1552. These constitutions have similar provisions that provided their respective governors with different powers depending on whether the legislature was in “session” or “in recess.” Mass. Const, of 1780, pt. 2, ch. 2, § 1, art. V; N.H. Const, of 1792 pt. 2, § L. When the legislatures were in “session,” the governors had the power either to prorogue or to adjourn them. See, e.g., Mass. Const, of 1780, pt. 2, ch. 2, § 1 (“The Governor ... shall have full power and authority, during the session of the General Court [i.e., the Massachusetts legislature], to adjourn or prorogue the same to any time the two Houses shall desire”). But when the legislatures were “in recess,” the governors only had the power to prorogue them — or, in simpler terms, extend the duration of the intersession recess, see Johnson, 2 A Dictionary of the English Language 412 (defining “prorogue” as “to withhold a session of parliament to a distant time.”). See, e.g., Mass. Const, of 1780, pt.2, ch 2, § 1 (providing the governor, “in the recess of the said Court,” the power “to prorogue the same from time to time”). These provisions make sense only *225if the legislature is not in “session” when it is “in recess.” Otherwise, the provisions are in conflict, stating that the governors both had and did not have the power to adjourn the legislature during intrasession breaks. These two constitutions thus used recess to mean intersession breaks only.15
There are, however, examples of state executives assuming that a constitutional recess includes intrasession breaks. Vermont and Pennsylvania’s former constitutions, for example, provided their respective executives power to “lay embargoes ... in the recess of the house only.” Vt. Const, of 1777, ch. 2, § XVIII; Pa. Const, of 1776, pt. 2, § 20. Governors of both states imposed embargoes during intrasession breaks,16 which suggests they understood that such breaks were included in the meaning of recess.
The New Jersey governor acted similarly. He relied on the Senate Vacancies Clause in the federal Constitution to appoint a senator on December 19, 1798. 8 Annals of Cong. 2197 (1798). Prior to the Twentieth Amendment, this Clause allowed state executives to make temporary *226appointments of Senators “during the Recess of the Legislature of [that] State.” U.S. Const, art. I, § 3, cl. 2. His appointment of a senator on December 19, 1798, shows that he construed recess to include intrasession breaks because the New Jersey General Assembly was in an intrasession break from November 8, 1798, until January 16,1799.17
This history shows that recess had at least two meanings at the time of ratification: either intersession breaks only' or intersession breaks plus long intrasession breaks. The state constitutions favor the former, while the governors’ actions favor the latter. To be sure, the executive’s actions should be viewed with some skepticism because an expansive definition of recess served their institutional self-interest by expanding their powers. See Steven G. Calabresi & Saikrishna B. Prakash, The President’s Power to Execute the Laws, 104 Yale L.J. 541, 558-59 (1994) (explaining that post-enactment actions by the first Congress must be viewed cautiously because of their institutional interest in limiting the president’s power). But it would be erroneous to reject their understanding on this basis alone. Nothing in the historical record affirmatively rejects their understanding for purposes of the federal Constitution.18 But neither is there anything affirmatively establishing that it adopted this definition of recess in lieu of the definition found in the Massachusetts and New Hampshire constitutions. Standing alone, “Recess of the Senate” is thus ambiguous. Edwards v. A.H. Cornell & Son, Inc., 610 F.3d 217, 222 (3d Cir.2010) (“Words or provisions are ambiguous when ‘they are reasonably susceptible of different interpretations.’ ” (quoting Dobrek v. Phelan, 419 F.3d 259, 263 (3d Cir.2005))).
Importantly, though, neither of these possibilities is similar to the unavailable-for-business definition put forth by the Board. Every example discussed thus far has two common characteristics. First, each break lasted for a considerable period of time. The intrasession breaks in which the governors of Vermont and Pennsylvania used their powers were 57 and 71 days, respectively. See swpra note 10. And the intrasession break in which the New Jersey governor appointed a senator was 69 days. See supra note 11. As far as we are aware, the shortest break referred to *227as a recess lasted 14 days, 2 A Documentary History of the English Colonies in North America 1346-48 (Peter Force, ed., 1839), which conforms with the modern practice equating recess with breaks lasting at least 10 days. These durations suggest that a recess was more than the day-to-day adjournment of a legislature and likely held the connotation of long duration. This is contrary to the Board’s current view that breaks in business need not be of any particular duration to constitute a recess.
The second notable trait of these breaks is that the beginning of each, was determined solely by when the legislature adjourned — rather than by some functionalist definition of when the body was unavailable for business. The Board has pointed to no examples of the word “recess” turning on factors such as whether members were required to attend, the legislative chamber was empty, and the body could receive messages. The examples instead show that recess was tied to the type, or possibly the duration, of the legislature’s self-defined adjournment. Accord Jefferson, supra, at 51 at 165 (explaining that Senate “Committees may be appointed to sit during a recess by adjournment, but not by prorogation”).
In short, the natural meaning of recess does not help us decide between intersession breaks and intrasession breaks of a fixed duration, but the relevant context does undermine the Board’s' current position.19 To resolve the remaining ambiguity, one might argue that the Constitution uses a definitive article: “the Recess of the Senate.” The word “the” might mean that the phrase refers to a specific thing, possibly suggesting that recess refers to the one recess that follows every session, an intersession break. See Noel Canning, 705 F.3d at 499-500. But “the” also can denote a particular class of something as well. Indeed, that is how the D.C. Circuit ultimately interpreted “the Recess,” holding that it means all intersession breaks. Id. But even conceding that “the” is meant to denote a specific class of something, there is nothing in the word “the” itself that necessarily requires that class to be intersession breaks. “[T]he Recess” might, for example, simply refer to times in which the Senate is in a recess. See Evans, 387 F.3d at 1224-25. There is nothing that shows what “the” means in the Recess Appointments Clause, especially because the Constitution uses “the” in several manners. See, e.g., U.S. Const, art. I, § 3, cl. 4 (providing that “[t]he Vice *228President ... shall be President of the Senate”); art. I, § 3, cl. 5 (providing that the Senate shall select a president pro tempore “in the Absence of the Vice President”). Accordingly, we are convinced that use of “the” is uninformative. We must therefore look to the broader textual context in which “the Recess of the Senate” was ratified.
B. Textual Context
1. Constitutional Context and the Unavailable-for-Business Definition
“If, from the imperfection of human language, there should be serious doubts respecting the extent of any given power, it is a well settled rule, that the objects [i.e., the purpose] for which it was given ... should have great influence on the construction.” Gibbons v. Ogden, 22 U.S. 1, 188-89, 9 Wheat. 1, 6 L.Ed. 23 (1824). The purpose of the Recess Appointments Clause is most evident in its relation to the Appointments Clause. The text and structure of the Constitution demonstrate that the Recess Appointments Clause is a secondary, or exceptional, method of appointing officers, while the Appointments Clause provides the primary, or general, method of appointment. The Appointments Clause provides the general rule for appointing officers through presidential nomination and senatorial advice and consent. U.S. Const, art. II, § 2, cl. 2. The Clause lacks any limitation on when this power is operative — the president always has the power to fill vacancies through nomination and the advice and consent of the Senate. See id. (“[The President] shall nominate, and by and with the Advice and Consent of the Senate, shall appoint ... ”).20 This perpetual power stands in contrast to the power given to the president in the Recess Appointments Clause, which explicitly allows him to fill vacancies unilaterally only “during the Recess of the Senate.” Id. at art. II, § 2, cl. 3. The clauses thus reveal a constitutional preference for divided power over the appointments process, which is deviated from only in specified situations.21
Alexander Hamilton echoed this understanding of the Constitution. He explained in Federalist 67 that the Appointments Clause “declares the general mode of appointing officers of the United States.” The Federalist No. 67 (Alexander Hamilton). The Recess Appointments Clause, however, is “nothing more than a supplement to the [the Appointments Clause], for the purpose of establishing an auxiliary method of appointment, in cases to which the general method was inadequate.” Id. Accordingly, the “ordinary power of appointment is confined to the president and the Senate jointly, and can therefore only be exercised during the session of the Senate.” Id. (emphasis in original). But “in [the Senate’s ] recess,” the *229“President, singly,” has power to make temporary appointments. Id. (emphasis in original). This deviation is necessary, Hamilton argues, because it is “improper to oblige this body to be continually in session” and because “it might be necessary for the public service to fill without delay.” Id.
The “main purpose” of the Recess Appointments Clause, therefore, is not — as the Eleventh Circuit held and the Board argues — only “to enable the President to fill vacancies to assure the proper functioning of our government.” Evans, 387 F.3d at 1226. This formulation leaves out a crucial aspect of the Clause’s purpose: to preserve the Senate’s advice-and-consent power by limiting the president’s unilateral appointment power. Accord Noel Canning, 705 F.3d at 505 (explaining that the Eleventh Circuit’s statement of the Clause’s purpose “omits a crucial element of the Clause, which enables the president to fill vacancies only when the Senate is unable to provide advice and consent” (emphasis in original)).
The importance of this aspect of the Clause’s purpose is difficult to understate. At the time of ratification, skepticism in executive unilateral appointments power was firmly established. “ ‘[T]he power of appointments to offices’ was deemed ‘the most insidious and powerful weapon of eighteenth century despotism.’ ” Freytag, 501 U.S. at 883, 111 S.Ct. 2631 (quoting Gordon Wood, The Creation of the American Republic 1776-1787 79 (1969)). But the framers’ skepticism concerning unilateral power was not limited to the executive. They also rejected unilateral legislative control of appointments out of concern for “diversity of views, feelings, and interests, which frequently distract and warp the resolutions of a collective body.” The Federalist No. 76 (Alexander Hamilton). As a consequence of these concerns, the framers sought to “ensure that those who wielded [appointments powers] were accountable to political force and the will of the people” by limiting the power of any one person or body. They did so by dividing that power between the executive and legislative branches. Freytag, 501 U.S. at 883-84, 111 S.Ct. 2631; see also Ryder, 515 U.S. at 182, 115 S.Ct. 2031 (“The [Appointments] Clause is a bulwark against one branch aggrandizing its power at the expense of another branch, but it is more: it preserves another aspect of the Constitution’s structural integrity by preventing the diffusion of the appointment power.”). To ignore this division of power is to neglect a central principle that underlies the two Appointments Clauses.22
*230And therein lies the implausibility of the unavailable-for-business definition. As explained above, the Board argues that a recess occurs any time members of the Senate do not have a duty to attend, the Senate chamber is empty, and the Senate is unavailable to receive communications from the president. See Resp. Br. at 44-45; 33 U.S. Op. Att’y Gen. at 21-22, 25. The1 problem with this definition is that the Senate fulfills these criteria whenever its members leave for the weekend, go home for the evening, or even take a break for lunch. In each of these instances, the senators have no duty to attend, the Senate chamber is empty, and the body cannot receive messages from the president.
Defining recess in this way would eviscerate the divided-powers framework the two Appointments Clauses establish. If the Senate refused to confirm a president’s nominees, then the president could circumvent the Senate’s constitutional role simply by waiting until senators go home for the evening. The exception of the Recess Appointments Clause would swallow the rule of the Appointments Clause.
The Board appears to recognize this difficulty with its definition. Oral Arg. Tr. at 48:6-9 (stating that “[t]he executive branch has not claimed authority to make recess appointments during lunch”). Accordingly, the Board argues that there is a limitation in addition to the three open-for-business criteria: unavailability to provide advice and consent. Oral Arg. Tr. at 49:15-18. But the Board does not clearly define unavailability in a way that distinguishes it from the Board’s discussion of when the Senate is open for business. At times, its brief treats the two requirements as one. Resp. Br. at 44 (“[T]he Clause authorizes appointments when the Senate is not open to conduct business and thus not providing advice and consent on nominations.”).
Perhaps the best indication of what the Board means by unavailability is its reliance on the Senate’s unanimous-consent agreement that established the schedule for the pro forma sessions from December *23120, 2011, through January 23, 2012. This resolution provided that there would be “no business conducted” during the sessions. 157 Cong. Rec. at S8783 (statement of Sen. Wyden). This resolution might be understood to mean that during the pro forma sessions the Senate was open for business but unavailable to provide advice and consent on nominations because of the body’s prior agreement.
The first problem with this argument is that the Senate’s actions under the resolution reveal that it could have provided advice and consent during these pro forma sessions if it had desired to do so. On December 23, 2011, during one of the pro forma sessions stipulated in the unanimous-consent agreement, the Senate passed a bill that provided “a 2-month extension of the reduced payroll tax, unemployment insurance, TANF, and the Medicare payment fix.” 157 Cong. Rec. at S8789 (statement of Sen. Reid). That same day, the Senate also “agree[d] to the request for a conference” from the House in relation to related bills passed by both chambers. Id. If the Senate could pass a bill and agree to a request from the House to create a conference for another bill, then the Senate likely could have provided its advice and consent but chose not to — as they are entitled to under the Appointments Clause.23
Besides this factual difficulty, the Board’s limiting principle has another, larger problem: it still does not foreclose day-to-day adjournments from constituting recesses. The important feature of the Senate’s scheduling agreement that the Board emphasizes is the provision that there would be “no business conducted.” Resp. Br. at 45-47; Oral Arg. Tr. at 49:21-24. This, however, is indistinguishable from a daily adjournment. At the end of *232the day, the Senate adjourns, which represents an agreement that it will do no business until it reconvenes the next day. In fact, when the Senate agrees to adjourn, it agrees that no senator can even be recognized to speak on the floor. See Riddick’s Senate Procedure: Precedents and Practices, S. Doc. No. 101-28, at Adjournment 2 (1992) (“Once the Chair has announced that the Senate stands in adjournment, there is no recourse available to the Senator seeking recognition until the Senate reconvenes.”). The only distinction is formalistic — day-to-day adjournments are embodied in a motion to adjourn (that is often unanimously agreed to) rather than a unanimous consent agreement — but there is no reason to believe that makes an actual difference under the Board’s approach. Therefore, the Board’s limiting principle fails to limit the meaning of recess and must be rejected to prevent the Recess Appointment Clause’s exception from swallowing the rule of divided power.
Now that we have established what “the Recess of the Senate” does not mean, we must establish what it does mean. The Recess Appointments Clause’s preservation of the Senate’s advice-and-consent power does not help us decide between the remaining two possibilities because the requirement that an intrasession break last a certain duration would prevent the exception from swallowing the rule. We must therefore look to provisions of the Constitution.
Several constitutional provisions appear relevant to our analysis, such as those that use the word “adjournment.” See Noel Canning, 705 F.3d at 500. Adjournment, as discussed above, is an instance in which Congress or one of its chambers takes a break of any type or length. See, e.g., Pocket Veto Case, 279 U.S. at 680, 49 S.Ct. 463 (interpreting “adjournment” in the Pocket Veto Clause to include both types of breaks). Thus, if the framers had intended for the president to be able to appoint officers during intrasession breaks, then the Recess Appointments Clause could have been worded differently, allowing recess appointments “during the Adjournment of the Senate.” See Noel Canning, 705 F.3d at 500, 505-06. Because the Constitution uses recess instead of adjournment, we presume that recess has a meaning different from adjournment. Kelo v. City of New London, Connecticut, 545 U.S. 469, 496, 125 S.Ct. 2655, 162 L.Ed.2d 439 (2005) (“When interpreting the Constitution, we begin with the unremarkable presumption that every word in the document has independent meaning, ‘that no word was unnecessarily used, or needlessly added.’ ”) (quoting Wright v. United States, 302 U.S. 583, 588, 58 S.Ct. 395, 82 L.Ed. 439 (1938)).
That the words have different meanings, however, does not necessarily tell us what those meanings are and whether they might overlap. The Eleventh and D.C. Circuits provide two different possibilities. On the one hand, adjournment could mean the act of adjourning (i.e., ending business) for any period of time, while recess could refer to the period of time that follows an adjournment. Evans, 387 F.3d at 1225. On the other hand, adjournment could again mean the act of adjourning for any period of time, while recess might refer to breaks of a more limited nature — whether that be limited by the duration of the break or the type of break. Noel Canning, 705 F.3d at 500. In both instances, adjournment and recess have different meanings but nothing about the dichotomy between the words tells us which meaning was intended.
When these possibilities are considered in light of the purpose of the Recess Appointments Clause, however, the dichotomy must be that adjournment results in *233more breaks than recess does. If the Eleventh Circuit is correct that the sole reason for using recess instead of adjournment was to recognize a difference between the act of adjourning and the period that follows, then recess would mean any break in Senate business regardless of the break’s length. This is a broad definition that no one, including the Board, adopts because it would result in the exception swallowing the rule. So the dichotomy does reveal that recess must mean something narrower than any break that follows an adjournment.
But what this narrower definition is cannot be derived from the dichotomy between adjournment and recess alone. Nothing about the words tells us whether recess is limited by the duration of the break (as the intrasession definition does) or by the type of break (as the intersession definition does). Contra Noel Canning, 705 F.3d at 500, 505-06 (using the dichotomy plus the fact that recess is preceded by “the” as support for its conclusion that “the Recess” must mean intersession breaks only). The dichotomy between adjournment and recess therefore leaves us in the same place as the Recess Appointments Clause’s purpose: rejecting an all inclusive definition of recess but without a basis to decide between the intersession definition and the intersession-plus-long-intrasession-breaks definition.
2. Constitutional Context and the Remaining Definitions
We resolve this uncertainty by first noting what is absent in the Constitution: a link between “the Recess of the Senate” and any particular length of time. Attorney General Daugherty, who first suggested a minimum duration of ten days, did not tie this duration to any constitutional provision. See 38 U.S. Op. Att’y Gen. at 24-25 (“Nor do I think an adjournment for 5 or even 10 days can be said to constitute the recess intended by the Constitution.”). Some have tried to tie the duration to the Adjournment Clause, which requires either chamber of Congress to obtain the consent of the other to adjourn for more than three days, U.S. Const, art. I, § 5, cl. 4.24 See, e.g., 33 U.S. Op. Att’y gen. at 24-25 (invoking the Adjournment Clause to reject the idea that two days may constitute a recess); Edward A. Hartnett, Recess Appointments of Article III Judges: Three Constitutional Questions, 26 Cardozo L.Rev. 377, 419-21 (2005). The argument is that the Adjournment Clause provides a measure of what constitutes a de minimis break — one that should be read into the Recess Appointments Clause to prevent the exception from swallowing the rule. See Hartnett, supra, at 419-21.25 *234The central error in this argument, however, is that “[n]othing in the text of either Clause, the Constitution’s structure, or its history suggests a link between the Clauses.” Noel Canning, 705 F.3d at 504; cf. Atl. Cleaners & Dyers v. United States, 286 U.S. 427, 433-34, 52 S.Ct. 607, 76 L.Ed. 1204 (1932) (demonstrating that the context of individual provisions is important to deciding the meaning of them by explaining that the same words in the Constitution often have different meanings depending on their context). Absent some connection, there is no reason to believe that the Adjournment Clause’s duration requirement controls the meaning of the Recess Appointment Clause. And beyond the Adjournment Clause, nothing in the Constitution establishes the necessary length of an intrasession break that would constitute a recess.26 This is the most significant weakness of the long-break in-trasession argument.
Although there is no constitutional basis for any sort of durational limit on what constitutes “the Recess,” the Recess Appointments Clause does contain a temporal characteristic: the Recess Appointment Clause’s specification that recess-appointed officers’ terms “shall expire at the End of [the Senate’s] next Session.” U.S. Const, art. II, § 2, cl. 3. A session of the Senate, everyone agrees, begins at the Senate’s first convening and ends either when the Senate adjourns sine die or automatically expires at noon on January 3 in any given year. Henry B. Hogue, Cong. Research Serv., RS21308, Recess Appointments: Frequently Asked Questions 1-2 & n. 5 (2012). The expiration of these officers’ terms at the end of the next session implies that their appointments were made during a period between sessions.
This implication follows from the reason for making recess appointments expire at the end of the “next Session.” As discussed, the Recess Appointment Clause provides an “auxiliary” method of appointing officers. The Federalist No. 67 (Alexander Hamilton) (explaining that the Clause is “nothing more than a supplement to the [Appointments Clause]” that “establish[es] an auxiliary method of appointment, in cases to which the general method is inadequate”). The durational provision maintains this by limiting recess appointees’ terms to last for only the time needed for the president and the Senate to have the opportunity to undergo the normal process. As Justice Joseph Story explained, the Clause authorizes the president “to make temporary appointments during the recess, which should expire, when the senate should have had an opportunity to act on the subject.” 3 Joseph Story, Commentaries on the Constitution of the United States § 1551 at 410 (1833) (emphasis added). Limiting the duration to a single opportunity follows from the auxiliary nature of the Clause. After all, the Senate’s decision not to act on a nomination effectively is a rejection of that nomination, as evidenced by the Senate’s routine return to the president of nomina*235tions who have not been acted on. Standing Rules of the Senate XXXI, para. 6 (“Nominations neither confirmed nor rejected during the session at which they are made shall not be acted upon at any succeeding session without being again made to the Senate by the President.”). In fact, a system in which Senate silence would allow for the appointment of officers was explicitly rejected at the drafting convention. 2 The Records of the Federal Convention of 1787 at 80-88 (Max Farrand ed., 1911); see also Adam J. White, Toward the Framers’ Understanding of “Advice and Consent”: A Historical and Textual Inquiry, 29 Harv. J.L. & Pub Pol’y 103, 117-19 (2005) (explaining the drafters’ rejection of a system in which only the Senate had the power to veto nominations); Matthew C. Stephenson, Can the President Appoint Principal Executive Officers without a Senate Confirmation Vote?, 122 Yale L.J. 940, 964-95 (2013). The Clause’s function is thus fulfilled once an opportunity for the Senate to act has come and gone.
So if recess includes intrasession breaks, then we would expect the recess-appointment term to last only until the end of that session. This is because once the Senate returned from its break there would be an opportunity to undergo the normal process. Yet the Constitution provides that the term would last until the end of the next session. This suggests that the dura-tional provision contemplates a meaning of recess that means intersession breaks only.
This is best seen in the process of recess appointments that results under each definition of recess. Under the intersession-only definition, the president would make a recess appointment between sessions of the Senate, which ensures the continued operation of the government even though the Senate has not considered the president’s selection. Once the Senate begins its “next Session” by reconvening, the primary appointments process becomes available and — because the Constitution requires joint appointment authority — must be undertaken by the Senate and the president. However, to allow the operation of government to continue, the Senate has until the end of this session to consider the president’s selection and confirm or deny it. And if the body does not act or denies that appointment, then the recess appointment ends because the constitutional requirement of joint agreement has not been reached. Through this process, the Appointments Clause retains its primacy as the preferred constitutional method of appointing officers, while the Recess Appointments Clause retains its auxiliary role that allows the president to fill positions when the ordinary process is unavailable.
Under an intrasession definition, the Clause would no longer have an auxiliary role. The president would make the recess appointment during a break within a Senate session. But the Senate’s reconvening and first subsequent adjournment — whether that be for a long intrasession break or for the intersession break— would have no immediate effect on the recess appointment because the appointment lasts until the “next Session,” as demarked by adjournments sine die. The appointment would not expire until the Senate reconvened, adjourned sine die, reconvened, and then adjourned sine die a second time. Thus, the appointment would continue even though the opportunity to undergo the ordinary, preferred process had come and gone. This shows that when the intrasession definition of recess is combined with the durational provision, a fundamentally different relationship between the clauses is created: the intrasession definition makes the Recess Appointments Clause an additional rather than auxiliary method of appointing officers.
*236The durational provision thus indicates that the most natural reading of the Clause defines recess to mean intersession breaks only. Cf. Weinberger v. Hynson, Westcott & Dunning, Inc., 412 U.S. 609, 681-32, 93 S.Ct. 2469, 37 L.Ed.2d 207 (1973) (“It is well established that our task in interpreting separate provisions of a single Act is to give the Act the most harmonious, comprehensive meaning possible in light of the legislative policy and purpose.” (internal quotation marks omitted)); Gustafson v. Alloyd, Inc., 513 U.S. 561, 575, 115 S.Ct. 1061, 131 L.Ed.2d 1 (1995) (“[A] word is known by the company it keeps. This rule we rely upon to avoid ascribing to one word a meaning so broad that it is inconsistent with its accompanying words”). This is supported by the fact that the original Senate Vacancies Clause used a different durational provision: “the next Meeting.” U.S. Const, art. I, § 3, cl. 2, superseded, by id. Amend. XVII.27 The original language shows that the durational provision in the Recess Appointments Clause could have been phrased in a manner that would have allowed the Senate and president only one opportunity to undergo the ordinary process if recess instead included intrasession breaks. By setting the duration to the “next Meeting,” it becomes irrelevant what type of break the legislature took because once it convenes, the appointment expires and the legislature must act.28 That the Clause uses “next Session” rather than “next Meeting” thus shows that recess contemplates a particular type of break. And, in light of the competing operations of the definitions, that type is the intersession break.
The Board disagrees with this characterization. It argues that the duration provision conforms with an intrasession definition of recess because if recess appointees’ tenures did not extend until the end of the next session, then the Senate would lack an opportunity to consider a recess appointee when an intrasession break coincides with the end of a session. NLRB Ltr. Br. at 12-13. After all, if the appointment lasted until the end of the Senate’s session, and the intrasession break in which he was appointed lasted until the end of that session, then the appointee’s term would expire at the end of that break and the Senate would not have a chance to consider the appointment. So, according to the Board, fixing the duration to the next session might ensure that the Senate has an opportunity to provide its advice and consent.
This argument is unpersuasive for two reasons. First, the problem arises only if one adopts an intrasession definition of recess. If recess is limited to intersession breaks, then there will never be any doubt *237that the Senate will have its single chance to weigh in: once it reconvenes for its next session. Avoiding this problem is yet another reason to define recess to mean intersession breaks, Cf. Am. Tobacco Co. v. Patterson, 456 U.S. 63, 71, 102 S.Ct. 1534, 71 L.Ed.2d 748 (1982) (explaining that “[statutes should be interpreted to avoid untenable distinctions and unreasonable results whenever possible”).
Second, we acknowledge that the dura-tional provision can be read consistently with an intrasession definition. But the Board’s point does not show that the most natural reading of the Clause’s duration provision supports this definition. Instead, it tends to show the opposite. We doubt that the phrase “next Session” is intended to address an unusual situation— one that the drafters’ of the Constitution were unlikely to contemplate. An intra-session break has extended until the end of one of the Senate’s 296 completed sessions only once, in 1992. (And even if we were to adopt the Board’s contention that pro forma sessions constitute a recess — which we do not — then the number increases to three times, in 2008 and 2011).29 In other *238words, if fixing the duration until the Senate’s next session (rather than the end of that session) is meant only to ensure the Senate has a chance to provide its advice and consent without regard to its effect on the broader framework, then the duration provision’s purpose has only become important one time. And even during this recent instance, the Senate convened before their constitutionally imposed deadline and could have considered nominations if they had chosen to do so. See 137 Cong. Rec. 36364 (Jan. 3, 1992). The complete absence of the problem described by the Board in the last 225 years suggests that the Constitution most likely was not written with such a problem in mind. Cf. Marozsan v. United States, 852 F.2d 1469, 1498 (7th Cir.1988) (en banc) (Easterbrook, J., dissenting) (“The terror of extreme hypotheticals produces much bad law.”). This implies that the durational provision was most likely written simply to reinforce the auxiliary nature of the Recess Appointment Clause by limiting recess appointees’ terms to last only as long as necessary to afford the Senate one opportunity to undergo the ordinary process.
The Constitution thus shows that the more limited definition of recess — that is necessitated by the purpose of the Recess Appointments Clause and the adjournment-recess dichotomy — includes only intersession breaks. Nothing within the broader context of the Constitution supports the Board’s definition. As for the intersession-plus-long-intrasession definition, although it could conform with the relationship between the Clauses, there is no constitutional basis for defining “long” and the definition is unsupported by the other relevant constitutional provisions.30 The relationship between the Appoint*239ments Clauses, the duration of recess appointments, and the Constitution’s use of adjournment elsewhere all show that “the Recess of the Senate” includes only breaks between sessions of the Senate.
C. Historical Practice
Our conclusion is supported by early historical practice. From ratification until 1921, there was a rough consensus that recess appointments could be made only during intersession breaks. See Rappaport, supra, at 1572-73. Before 1867, no president made a recess appointment during an intrasession break of the Senate. Id.; Hartnett, supra, at 408-10. In 1867 and 1868, President Andrew Johnson made several recess appointments during intra-session breaks of the Senate. Hartnett, supra, at 408-10. His use of the appointments powers, however, was a cause of significant turmoil at the time and it served a not insignificant role in his eventual impeachment. Id. at 409; Rappaport, supra, at 1572. Accordingly, it is unclear whether President Johnson’s actions were based on a consensus view of the Constitution. There is evidence that it was not. U.S. Attorney General Philander Knox— the first attorney general to directly address the meaning of recess — advised President Theodore Roosevelt that he could not make a recess appointment during intrasession breaks. 23 U.S. Op. Att’y Gen. 599, 604 (1901). For over one-hundred years following ratification, recess was generally understood to mean intersession breaks only.
To be sure, this practice arose when intrasession breaks were generally no longer than two weeks. Rappaport, supra, at 1572; Hartnett, supra, at 410. But that is no reason to discount the practice. As modern practice has shown, it is sometimes -in the interest of presidents to make recess appointments during breaks as short as two weeks. See, e.g., Evans, 387 F.3d at 1221 (describing President George W. Bush’s recess appointment of Judge William Pryor to the Eleventh Circuit during an eleven-day intrasession break). That presidents did not assert this power for over 100 years — despite this interest— suggests that they do not, in fact, have this power. Cf. Prints v. United States, 521 U.S. 898, 907-08, 117 S.Ct. 2365, 138 L.Ed.2d 914 (1997) (explaining that an absence of examples of Congress “impressing] the state executive into its service ... suggests an assumed absence of such power” (emphasis in original)); see also Noel Canning, 705 F.3d at 502.
Executive practice changed in 1921 when President Warren Harding made an intrasession recess appointment. Michael A. Carrier, Note, When is the Senate in Recess for Purposes of the Recess Appointments Clause?, 92 Mich. L.Rev. 2204, 2235 (1994). As explained above, this act was supported by U.S. Attorney General Daugherty, who reversed the opinion offered by Attorney General Knox a mere twenty years earlier. 33 U.S. Op. Att’y Gen. at 21-22. Attorney General Daugherty explained that “whether the Senate has adjourned or recessed ... is whether in a practical sense the Senate is in session so that its advice and consent can be obtained.” Id. This conclusion was based on a Senate Judiciary Committee report, which argued that practical considerations should prevent a president from using his recess-appointment power during intersession breaks that last mere seconds. Id. at 24. From this report, he drew the practical considerations that the Board urges us to adopt today, explaining that the Senate is not in session when its members have no duty to attend, the chamber is empty, and the Senate cannot receive communications. Id.
*240Importantly, Attorney General Daugherty explicitly rejected the “all recesses” implication of this test. He recognized that the practical considerations identified could allow presidents to use their power for “an adjournment for only 2 instead of 28 days” but rejected the idea that 2 days were sufficient to constitute a recess within the meaning of the Constitution. Id. at 24-25 (answering “unhesitatingly” that two days did not amount to a recess). He explained that not “even 10 days can be said to constitute the recess intended by the Constitution.” Id. at 25. As discussed above, though, this suggestion of ten days is not linked to any text in the Constitution.
Since issuance of Attorney General Daugherty’s opinion, the executive has claimed the authority to recess appoint officers during intrasession breaks. Before World War II, however, the power was used only one other time. Carrier, supra, at 2211-12. After World War II, intrasession appointments remained relatively rare for some time: President Harry Truman made twenty, President Dwight Eisenhower made nine, President Richard Nixon made eight, and President Jimmy Carter made seventeen; but Presidents John Kennedy, Lyndon Johnson, and Gerald Ford made none. Id. at 2212-13. The practice grew dramatically under President Ronald Reagan, who made 73 intrasession appointments, and it has seen significant use ever since: President George H.W. Bush made 37, President Bill Clinton made 53, and President George W. Bush made 141; President Barack Obama made 26 as of January 5, 2012. Id. at 2214-15; Henry B. Hogue et al, Cong. Research Serv., The Noel Canning Decision and Recess Appointments Made from 1981-2013 *4 (2013). Thus, it has been only over the last thirty years that presidents began relying so heavily on such recess appointments.
Notably, this relatively recent practice supports only an intrasession definition that is associated with a long duration. It does not support the Board’s functionalist definition. The executive has maintained from 1921 until 2012, at least in practice, that a certain number of days must pass before an intrasession appointment could be made. See Carpenter et al., supra, at 15 (“The length of the recess may be of great importance, as it appears that no President, at least in the modern era, has made an intrasession recess appointment during a recess of less than 10 days.”); see also 36 Op. O.L.C. *1 (Jan. 6, 2012) (“This Office has consistently advised that a recess during a session of the Senate, at least if it is sufficient length, can be a ‘Recess.’ ” (citation and internal quotation marks omitted)). The Board now seeks to abandon this limitation, which is completely unsupported by modern practice.
More important, however, recent practices cannot alter the structural framework of the Constitution. The Eleventh Circuit relied on a presumption that actions by the president are constitutional. Evans, 387 F.3d at 1222.31 We doubt that *241the presumption applies in separation-of-powers cases. In Clinton v. New York City, 524 U.S. 417, 118 S.Ct. 2091, 141 L.Ed.2d 393 (1998), for example-, the Supreme Court analyzed the constitutionality of the line-item veto without ever expressing the need to defer to the other branches’ constitutional judgments. And in Morris on v. Olson, 487 U.S. 654, 108 S.Ct. 2597, 101 L.Ed.2d 569 (1988), Justice Sca-lia noted in his dissent that one could “not find anywhere in the Court’s opinion the usual, almost formulary caution that we owe great deference to Congress’ view that what it has done is constitutional.” Id. at 704-05, 108 S.Ct. 2597 (Scalia, J., dissenting). The absence of deference is also found in the Supreme Court’s most recent separation-of-powers case, Free Enterprise Fund v. Public Company Accounting Oversight Board, — U.S.-, 130 S.Ct. 3138, 177 L.Ed.2d 706 (2010). There, the Court pointedly explained that “the separation of powers does not depend on the views of individual Presidents, nor on whether ‘the encroached-upon branch approves the encroachment.’ ” Id. at 3155 (quoting New York v. United States, 505 U.S. 144, 182, 112 S.Ct. 2408, 120 L.Ed.2d 120 (1992)). This is because “[t]he Constitution’s division of power among the three branches is violated where one branch invades the territory of another, whether or not the encroached-upon branch approves the encroachment.” New York, 505 U.S. at 182, 112 S.Ct. 2408.
The lack of deference to executive and legislative judgments on these issues follows from the fact that “separation-of-powers jurisprudence generally focuses on the danger of one branch’s aggrandizing its power at the expense of another branch.” Freytag, 501 U.S. at 878, 111 S.Ct. 2631. Giving deference to either branch is inconsistent with this concern because a presumption could prevent us from stopping one branch from “aggrandizing its power at the expense of another branch,” or ensuring that “the carefully defined limits on the power of each Branch” are not eroded, Chadha, 462 U.S. at 957-58, 103 S.Ct. 2764. Our role as the “ultimate interpreter of the Constitution” requires that we ensure its structural safeguards are preserved. Baker, 369 U.S. at 211, 82 S.Ct. 691. It is a role that cannot be shared with the other branches anymore than the president can share his veto power or Congress can share its power to override vetoes. See United States v. Nixon, 418 U.S. 683, 704-05, 94 S.Ct. 3090, 41 L.Ed.2d 1039 (1974). This “requires that [we] on occasion interpret the Constitution in a manner at variance with the construction given the document by another branch.” Powell v. McCormack, 395 U.S. 486, 549, 89 S.Ct. 1944, 23 L.Ed.2d 491 (1969).
The Supreme Court has stated as much in respect to the appointments provisions of the Constitution. In Frey-tag, the Supreme Court explained that the Appointments Clause represents an inde*242pendent restraint on both branches — one that does not exclusively serve either branch’s interests. 501 U.S. at 880, 111 S.Ct. 2631. This is equally true for the Recess Appointments Clause: just as “[t]he structural interests protected by the Appointments Clause are not those of any one branch of Government but of the entire Republic,” id., the structural protections of the Recess Appointments Clause belong to no single branch. Accordingly, “[t]he assent of the Executive to a bill which contains a provision contrary to the Constitution does not shield it from judicial review.” Id. (quoting Chadha, 462 U.S. at 942 n. 13, 103 S.Ct. 2764). This applies equally to the legislature’s assent to executive practice. Any acquiescence between the branches remains subject to the constraints imposed by the Constitution. There is “no statute of limitations for interpreting and enforcing the Constitution.” Evans, 387 F.3d at 1237 (Barkett, J., dissenting).
Furthermore, our analysis of recent practice is “sharpened rather than blunted by the fact that [the practice in question is] appearing with increasing frequency.” Chadha, 462 U.S. at 944, 103 S.Ct. 2764. Our analysis has shown that defining recess to mean intersession plus long intra-session breaks is incompatible with the Constitution. Although this definition is consistent with one possible meaning of “recess” in isolation, it is unsupported by the rest of the Constitution. The Constitution provides no measure of a “long” duration and limits the duration of recess appointees’ terms in a manner that indicates an intersession-only definition. This means that the current practice is contrary to the structural framework set out in the Constitution and must be held unconstitutional.
D. Additional Considerations
Our conclusion that recess includes only intersession breaks is supported by the Supreme Court’s direction that “the doctrine of separation of powers is a structural safeguard” which has as one of “its major feature[s]” the “establish[ment] [of] high walls and clear distinctions because low walls and vague distinctions will not be judicially defensible in the heat of interbranch conflict.” Plaut v. Spendthrift Farm, Inc., 514 U.S. 211, 239, 115 S.Ct. 1447, 131 L.Ed.2d 328 (1995) (emphasis in original). This bolsters our rejection of the Board’s definition because the unavailable-for-business criteria are almost by definition a “low wall” that contain “vague distinctions” which will make them difficult for the Senate and the president to predictably apply. The vagueness of the Board’s definition is perhaps best captured by its argument that the Senate is not available for business during pro for-ma sessions even though there are documented examples of the Senate conducting business during such sessions. Its definition thus falls far short of containing the “major feature” of separation-of-powers structural safeguards.
This is also true for the intrasession definition that limits recess to long breaks. This definition is not “judicially defensible” because whatever duration is selected as long would be based on something other than the Constitution. See Maryland v. Shatzer, 559 U.S. 98, 118-20, 130 S.Ct. 1213, 1228, 175 L.Ed.2d 1045 (2010) (Thomas, J., concurring in part, dissenting in part) (explaining that “an otherwise arbitrary rule is not justifiable merely because it gives clear instruction”). Furthermore, although an arbitrary number of days at first seems to erect a high wall and clear distinction, further review reveals that it is also fraught with ambiguity. For example, if we were to hold that an intra-session break of over ten days constitutes a recess, it is unclear at which point the *243adjournment evolves into a recess. Assume the Senate initially agrees to adjourn for twelve days but provides the majority leader the power to recall the body earlier, as it often does. See, e.g., H. Con. Res. 307, 111th Cong. (2010) (providing the House of Representative’s concurrence with the Senate that the latter would take a month-long recess starting in August 2010 that ended on September 13, 2010 unless “[t]he Majority Leader of the Senate ..., after consultation with the Minority Leader of the Senate, shall notify the Members of the Senate to reassemble at such place and time as he may designate if, in his opinion, the public interest shall warrant it”). Does this adjournment become a recess at the moment the Senate votes for the adjournment or must ten days first elapse? If the former, then assume the majority leader reconvenes the body after eight days of the adjournment. At this point, would a recess appointment made on the first day of the adjournment become invalid because it was not made during “the Recess” of the Senate? The Constitution provides no clear answer to these difficult questions. The long-intra-session break definition thus lacks the clear distinctions required by the Supreme Court, which means that the intersession definition is the only one that provides high walls and clear distinctions rooted in the text of the constitution.
The Board nevertheless argues that the rule we adopt today creates too powerful an opportunity for mischief by the Senate.32 The intersession definition allows the Senate to prevent the president from exercising his recess-appointment powers by manipulating the timing and the types of its adjournments. See NLRB Ltr. Br. at 10-11. This is true. But the opportunity for abuse is present under any possible definition of recess. Under the Board’s definition, the Senate might avoid being in recess by stopping the practice of agreeing by unanimous consent that no business will be done during pro forma sessions; or it might alter its procedures to allow messages to be received during such sessions, thus making itself available for business under the Board’s definition, see Resp. Br. at 44. And under the other intrasession definition, the Senate could very well adopt scheduling orders that prevent a break from lasting longer than whatever duration courts selected — as, in fact, pro forma sessions are designed to do. Yet the potential for abuse is not limited to the Senate, as presidents may also abuse any definition given to recess. Under the interses*244sion definition, as a final example, presidents still could recess appoint (and indeed have so appointed33) officers during intersession breaks that last negligible periods of time — the lack of a constitutional basis for selecting a long duration in defining intrasession breaks is just as absent to define intersession breaks.
All this is to say that the potential for abuse and subsequent gridlock lies not in what recess means but in the Constitution’s framework of divided powers. A division of powers demonstrates that “[c]onvenience and efficiency are not the primary objectives — or the hallmarks — of democratic government.” Chadha, 462 U.S. at 944, 103 S.Ct. 2764. We, as federal judges, are not empowered to regulate, recommend, or comment on how the two other branches of the federal government should use the powers the Constitution allocates between them — not because we can-but-chose-not-to, but because we lack the factual record, institutional tools, and constitutional authority to evaluate which branch is more or less likely to abuse the powers given to them. We can, however, and indeed we must, decide what powers each branch has and when they may use them because “[i]t is emphatically the province and duty of the judicial department to say what the law is.” Marbury v. Madison, 5 U.S. (1 Cranch) 137, 177, 2 L.Ed. 60 (1803). That is all we do today.
VI
Member Becker was invalidly recess appointed to the Board during the March 2010 intrasession break. This means that the delegee group had fewer than three members when it issued the August 26 Order. Consequently, the delegee group acted without power and lacked jurisdiction when it issued the order. Our holding makes it unnecessary to interpret the word “happen” in the Recess Appointments Clause. Accord Noel Canning, 705 F.3d at 515 (Griffith, J., concurring).34 Furthermore, we need not address whether the Board’s substantive decision was correct or whether the delegee groups that decided the subsequent reconsideration orders were properly composed. Dir., Office of Workers’ Compensation Programs v. Sun Ship, Inc., 150 F.3d 288, 291 (3d Cir.1998). Review of the reconsideration orders is also unnecessary because they were based on consideration of an invalid order. We will therefore vacate the Board’s orders.
. Refusal to bargain is a common way to obtain judicial review of representation determinations like the Board’s affirmation of the regional director's March 9, 2011 decision for which direct review is unavailable. NLRB v. Kentucky River Cmty. Care Inc., 532 U.S. at 709, 121 S.Ct. 1861 (explaining that "direct judicial review of representation determinations is unavailable” but that indirect review may be obtained by refusing to bargain and thereby inducing the Board to file an unfair labor’practice claim (citing AFL v. NLRB, 308 U.S. 401, 409-11, 60 S.Ct. 300, 84 L.Ed. 347 (1940))).
. The D.C. Circuit appears to have conflated the quorum requirement with the three-member-composition requirement. See generally Noel Canning, 705 F.3d at 490 (discussing challenge as one based on the quorum requirement); id. at 499 (stating that New Process Steel "holds that the Board cannot act without a quorum of three members” and “[i]t is undisputed that the Board must have a quorum of three in order to take action”). Notwithstanding the semantics, the substance of the D.C. Circuit's conclusion was that when less than three members purport to exercise the adjudicative authority of the Board, it "raise[s] questions that go to the very power of the Board to act.” Id. at 497. We agree.
. The Board also argues that Vermont Agency of Natural Resources v. United States, 529 U.S. 765, 120 S.Ct. 1858, 146 L.Ed.2d 836 (2000), describes the Appointments Clause as nonju-risdictional: Id. at 778 n. 8, 120 S.Ct. 1858. That case, however, states no such thing. Instead, it illustrates the very point we make here. It describes the question in which the appointments issue arose, rather than the Appointments Clause itself, as nonjurisdictional. Id. (stating that "the validity of qui tarn suits” is not “a jurisdictional issue”). And because that question was nonjurisdictional, the appointments issue within the question was not treated as jurisdictional. See id. Our conclusion in relation to the three-member-composition requirement for delegee groups is the opposite: it is jurisdictional. Accordingly, the appointments issue here must be treated as *213jurisdictional because it is one reason that there may not have been three members.
. The characteristics of pro forma sessions are described in Part V.
. The amicus also briefly refers to two other bases for concluding this is a political question: that (1) resolving the issue is impossible " 'without expressing lack of the respect due coordinate branches of government,' " id. at 5 (quoting Baker, 369 U.S. at 217, 82 S.Ct 691), and (2) "the nation's extreme need for finality in the president’s recess appointment practice,” id. (emphasis in original). Neither is persuasive. Defining recess in the Recess Appointments Clause does not express a lack of respect for coordinate branches of government because defining the word is merely an exercise of our judicial authority "to say what the law is,” which sometimes requires an evaluation of whether one branch is aggrandizing its power at another’s expense. See Zivotosky v. Clinton, -U.S. -, 132 S.Ct. 1421, 1427-28, 182 L.Ed.2d 423 (2012); see also Nixon v. Fitzgerald, 457 U.S. 731, 753-54, 102 S.Ct. 2690, 73 L.Ed.2d 349 (1982) (explaining, when discussing appropriate exercise of judicial review of executive action, that "[wjhen judicial action is needed to serve broad public interest — as when the Court acts, not in derogation of the separation of powers, but to maintain their proper balance ... the exercise of jurisdiction has been warranted” (citations omitted)). Nor is the constitutionality of the president’s recess-*216appointments practice the type of question implicating an extreme need for finality that would make it nonjusticiable. Cf. Baker, 369 U.S. at 213, 82 S.Ct. 691 (discussing the need for finality in the context of the president’s war power to end a conflict).
. Even Congress’ plenary authority over immigration and naturalization does not render its actions in this area immune from judicial review under the political-question doctrine. In INS v. Chadha, for example, the Supreme Court held that Congress' plenary authority over immigration did not render any challenge to that authority to be a nonjusticiable political question. 462 U.S. 919, 940-41, 103 S.Ct. 2764, 77 L.Ed.2d 317 (1983). The Court explained that "[t]he plenary authority of Congress over aliens ... is not open to question” except when it is alleged that the means chosen " 'offendfs] some other constitutional restriction’ ” on Congress. Id. (quoting Buckley v. Valeo, 424 U.S. 1, 132, 96 S.Ct. 612, 46 L.Ed.2d 659 (1976)).
. Senate practice also ends sessions automatically through its understanding of the Consti*219tution’s requirement that they "shall assemble at least once in every year” in a meeting that begins "at noon on the 3d day of January.” U.S. Const. Amend. XX. Under this practice, if a session of Congress has not ended by noon on January 3 of a given year, then the session automatically ends and another begins at noon of that day. See Thomas Jefferson, A Manual of Parliamentary Practice: For the Use of the Senate of the United States 166 (2d ed.1812).
. Others have argued that a three-day break is sufficient to constitute "the Recess of the Senate.” See, e.g., Edward A. Hartnett, Recess Appointments of Article III Judges: Three Constitutional Questions, 26 Cardozo L.Rev. 377, 419-21 (2005). This number is drawn from the Adjournments Clause, which requires the Senate and the House to concur on any adjournment lasting longer than three days. U.S. Const, art. I, § 5, cl. 4. The argument is that any intrasession break of less than three days is de minimis and thus not adequate to constitute "the Recess of the Senate.” Hart-nett, supra, at 419-20.
. The Board does note that if pro fonna sessions are ignored, then more than ten days passed during the break in which the president recess appointed two Board members who sat on the March 15 and 27 delegee group. Resp. Br. at 46 (noting that twenty days passed between when the second session of the Senate was convened on January 3, 2012 and when the Senate held its first non-pro forma session). This might suggest that the Board believes a period of time greater than ten days between non-pro forma sessions is still required, but such a contention is absent from its briefs and was not suggested at oral arguments when asked for limiting principles to its definition. Resp. Br. at 43-45 (defining recess in only functionalist terms), 58 (rejecting the relevance of a three-day requirement derived from the Adjournment Clause because nothing shows that it is related to the Recess Appointments Clause); Oral Arg. Tr. at 48:11 to 50:1 (explaining that “unavailability of the Senate to provide advice and consent” is the limiting principle on the functionalist definition of recess).
. We disagree with the dissent that the second and third definitions of recess should be combined into one "intrasession recess” definition. Dissenting Op. at 244-45. Distinguishing between these definitions provides necessary nuance to the analysis. First, as has been discussed, these two definitions have starkly different historical pedigrees: Until 2012, presidents and their attorneys general have always tied intrasession breaks to a non-negligible period of time. See, e.g., 33 U.S. Op. Att’y Gen. at 25. In fact, the Office of Legal Counsel’s 2012 memorandum on President Obama's recess appointments during pro forma sessions begins by emphasizing that the period between the non -pro forma sessions was of sufficient length to be a recess. 36 Op. O.L.C. *4-9 (Jan. 6, 2012). The availability: based definitions of recess that reject any need for a fixed number of days to pass thus represent a significant departure from past practice. Combining the unavailable-for-business definition with the long-intrasession-break definition glosses over important historical differences between the two.
Second, as will be shown, the unavailable-for-business definition has significantly less support than the long-intrasession-break definition from the historical meaning of "recess” as well as the purpose of the Recess Appointments Clause. Accordingly, we reject each definition for somewhat different reasons.
. The entire definition found in Johnson’s dictionary is:
1. Retirement; retreat; withdrawing; secession. 2. Departure. 3. Place of retirement; place of secrecy; private abode. 4. Perhaps an abstract of the proceedings of an imperial diet. 5. Departure into privacy. 6. Remission or suspension of any procedure. 7. Removal to distance. 8. Privacy; secrecy of abode. 9. Secret part.
Johnson, 2 A Dictionary of the English Language at 469.
. The words adjourn or adjournment appear six times in five clauses of the Constitution. U.S. Const, art. I, § 7, cl. 2 ("If any bill shall not be returned by the President within ten days (Sundays excepted) after it shall have been presented to him, the same shall be a law, in like manner as if he had signed it, unless the Congress by their adjournment prevent its return, in which case it shall not be a law.”); id. at art. I, § 7, cl. 3 (“Every order, resolution, or vote to which the concurrence of the Senate and House of Representatives may be necessary (except on a question of adjournment) shall be presented to the President of the United States”); id. at art. II, § 3 ("[The President] may, on extraordinary occasions, convene both Houses, or either of them, and in case of disagreement between them, with respect to the time of adjournment, he may adjourn them to such time as he shall think proper.”).
. The dissent argues that our discussion of state constitutions and early American practice transforms our definition of recess into a technical one. Dissenting Op. at 251-53 & n. 11. These sources are, however, frequently relied on by the Supreme Court to decide the meaning of Constitution. See, e.g., Heller, 128 S.Ct. at 2793-94; Collins v. Youngblood, 497 U.S. 37, 42, 110 S.Ct. 2715, 111 L.Ed.2d 30 (1990). We, too, consider such reliance to be appropriate because the average citizen likely would have understood the Constitution in reference to the state constitutions and practices at the time.
. See Del. Const, of 1776 art. 7; Md. Const, of 1776 pt. 2, art. XIII; N.C. Const, of 1776 pt. 2, arts. XVIII-XX; Pa. Const, of 1776 pt. 2, § 20; S.C. Const, of 1778 arts. IX, XVIII, XXXV; Vt. Const, of 1777 ch. II, §§ XVII-XVIII.
Of these provisions, the North Carolina Constitution’s Recess Appointments Clause has been argued to be the most relevant to the federal Recess Appointments Clause because the federal clause is thought by some to be modeled after the North Carolina one. Noel Canning, 705 F.3d at 501. The North Carolina Constitution gives the governor power to “grant[] temporary commission[s]” of officers “whose appointment[s] [were] by [the North Carolina] Constitution vested in the General Assembly ... during their recess.” N.C. Const, of 1776, pt. 2, art. XX. Recess here is essentially used in the same manner that it is in the federal constitution, which limits the recess-appointment power to “the Recess of the Senate.” Both constitutions thus contain the same ambiguity.
The D.C. Circuit concluded that this ambiguity is clarified for the North Carolina constitution by a North Carolina Supreme Court decision that the D.C. Circuit argues implicitly distinguishes between session and recess. Noel Canning, 705 F.3d at 501. We disagree. The North Carolina Supreme Court opinion is not informative because- — as the Board argues — the question in the case was not the meaning of "recess” but whether a recess-appointed judge’s court had jurisdiction to determine whether he was properly appointed. Beard v. Cameron, 7 N.C. 181, 3 Mur. 181, 184-86 (1819).
. The intersession-breaks-only definition of recess is also seen in a second way. As explained, the governors only had the power to prorogue when their respective legislatures were “in recess”; but they had the power to both adjourn as well as to prorogue the legislatures when they were in session. See Mass. Const, of 1780, pt. 2, ch 2, § 1, art. V; N.H. Const, of 1792 pt. 2, § L. This is telling because if recess included intrasession breaks as well as intersession breaks, then the power to adjourn ought to also be included. Recall that one central difference between adjournments and prorogations is that the former do not end all business such that it need be started anew when the legislature reconvenes while the latter do end business. Jefferson, supra, at 164-65. So in these constitutions, while the legislatures were in session, the governors had the option of either ending business through prorogation or, through adjournment, merely ending their meetings but without ending their business. There is no obvious reason that if recess included intra-session breaks — after which business that was ongoing before the break would continue— the governors would lose their power to end that business. The most plausible explanation of the differing powers in each situation is that recesses were only constituted of intersession recesses, which made it unnecessary to provide the governors the power to adjourn the legislatures because there was no business that could be continued. The Massachusetts and New Hampshire constitutions thus used recess to mean intersession recesses only.
. For the Vermont example, see 3 J. & Proceedings of the General Assemb. of the State of Vt. 235 (P.H. Gobia Press 1924) (recording the Vermont Assembly's "adjourn[ment] until the second Wednesday of June” on April 16, 1781); 2 Records of the Governor and Council of the State of Vt. 164 (E.P. Walton ed., 1874) (recording the May 1781 imposition of an embargo by the executive). This was an intrasession break because the legislature had not adjourned without day, as they often did to end the last meeting of the year. See, e.g., 3 J. & Proceedings of the General Assemb. of the State of Vt. at 31 (adjourning on June 17, 1778 "until his Excellency the Governor commands them to meet”), 73 (adjourning "without day” on June 4, 1779); 123 (adjourning "without day” on March 16, 1780); 271 (adjourning “without day” on June 28, 1781).
For the Pennsylvania example, see J. & Minutes of the Pa. Assembly 212 (1778) (recording the Pennsylvania House of Representatives’ adjournment on May 25, 1778 "to meet on the 9th day of September next” and its subsequent reconvening on August 4, 1778 pursuant to the summons of the "vice-president and [sjupreme executive council”); 11 Minutes of the Supreme Exec. Council of Pa. 544-45 (Theo Fenn & Co., 1852) (recording the August 1, 1778 imposition of an embargo by the executive). The Board has stated that this intrasession break lasted until September 9, 1778. This does not take into account the Pennsylvania House of Representative's being recalled on August 4, however. This discrepancy does not undermine the Board’s general point that the embargo was set by the executive during an intrasession break because the May 25 adjournment was not an adjournment sine die and the August 1 embargo imposition is before the Assembly's August 4 reconvening date.
. Votes and Proceedings of the Twenty-Third General Assemb. of the State of N.J., 1st sitting, 64 (1798-99) (recording the adjournment of the New Jersey General Assembly); J. of Proceedings of the Legis. Council of the State of N.J., 23d Sess., 1st sitting 20 (1798-99) (recording the adjournment of the New Jersey Legislative Council).
. Besides state-executive practice, the Board also points to the Continental Congress’s understanding of the meaning as revealed by its practices. NLRB Ltr. Br. at 6 & n. 3. Under the Articles of Confederation, the Congress could only convene a “Committee of the States” during "the recess of Congress.” Articles of Confederation of 1781, art. IX, para. 5; id. art. X, para. 1. Such a committee was convened during the period that followed the Continental Congress’s adjournment on June 3, 1784 until October 30, 1784. 27 J. of Continental Congress 555 (1784). That this adjournment was until a fixed date suggests that the period after ought to have been an intrasession break because it was not an adjournment sine die, which would be denoted by the absence of a fixed reconvening date. Subsequent proceedings, however, call this understanding into question because the Continental Congress’s journal does not record their reconvening on October 30 but instead shows them convening when the Articles of Confederation required they meet again, 28 J. of Continental Congress 639-41 (1784) (convening "[pjursuant to the Articles of Confederation,” rather than pursuant to the prior adjournment), which is consistent with having adjourned sine die. We decline to rely on this practice one way or another because of the uncertainty.
. The dissent refers to our reliance on state constitutions and contemporary interpretations of recess as a “dubious” method of interpretation. Dissenting Op. at 253-54. To be clear, these historical examples demonstrate that the use of recess at the time of ratification was consistent with either the intersession-break definition of recess or the intersession-plus-long-intrasession-break definition. We discuss these only to show the ordinary meanings of the word "recess” for the founding generation, as demonstrated by their usage. Heller, 554 U.S. at 576, 128 S.Ct. 2783. We do not use them as conclusive evidence that recess means intersession breaks only, which cannot be done because there is not sufficient historical evidence on which meaning was intended in the Constitution.
These historical practices do, however, cast doubt on the unavailable-for-business definition argued for by the Board, a version of which is adopted by the dissent. This is not so much because of what the practices were but what they were not. Namely, the Board and the dissent cannot point to a single example from the period of ratification in which a legislative body or executive defined recess exclusively using a functionalist definition based on availability. If such a definition of recess were a "normal and ordinary” meaning for the "founding generation,” Heller, 554 U.S. at 576, 128 S.Ct. 2783, there ought to be at least one example of its use from that period.
. The Appointments Clause states in full:
He shall have Power, by and with the Advice and Consent of the Senate, to make Treaties, provided two thirds of the Senators present concur; and he shall nominate, and by and with the Advice and Consent of the Senate, shall appoint Ambassadors, other public Ministers and Consuls, Judges of the supreme Court, and all other Officers of the United States, whose Appointments are not herein otherwise provided for, and which shall be established by Law: but the Congress may by Law vest the Appointment of such inferior Officers, as they think proper, in the President alone, in the Courts of Law, or in the Heads of Departments.
U.S. Const, art. II, § 2, cl. 2.
. Besides the exception found in the Recess Appointments Clause, the Appointments Clause also creates an exception for "inferior Officers.” These officers can be appointed either through the ordinary process or, if specified by statute, unilaterally by the President, courts, or department heads. See U.S. Const, art. II, § 2, cl. 2.
. The dissent understands this principle to mean that one purpose of the Recess Appointments Clause is "to provide a check on the Senate’s control over the appointment of officers by sharing the power of confirmation with the executive.” Dissenting Op. at 257. While we agree that the Clause is intended to preserve the balance of power struck in the Appointments Clause, we disagree that it does this by limiting the Senate’s power to provide its advice and consent. The Recess Appointments Clause preserves the balance of power by limiting the instances in which the president has unilateral authority to appoint officers, which is illustrated by its explicit limitation of that power to "the Recess of the Senate.” Nothing in the text of the Clause or the historical record suggests that it is intended to be a type of pressure valve for when the president cannot obtain the Senate's consent, whether that be because it has become dysfunctional or because it rejects a president’s nominations. Cf. The Federalist No. 67 (Alexander Hamilton) (explaining that the Clause is needed because it is “improper to oblige this body to be continually in session” or because "it might be necessary for the public service to fill without delay” rather than because it is a necessary tool to check the Senate’s power).
Our disagreement with our dissenting colleague is rooted in a difference in understanding of the president's and the Senate’s respec*230tive powers. Regarding the president, the dissent contends that we must interpret the president's recess-appointment power broadly because to do otherwise would "evisceratfe] his appointments prerogative” so that he may "be able to surround himself with the people he believed best fit to help him fulfill his duty.” Dissenting Op. at 255. But the president does not have an "appointments prerogative” or the constitutional right to surround himself with those he believes are "best fit to help.” That is exactly what the drafters rejected when they rejected unilateral appointments authority in the executive. The president has a prerogative to nominate whomever he likes, and the Senate has the prerogative to reject or confirm whomever the president nominates. To construe the Recess Appointments Clause as providing presidents these rights is to promote it from an auxiliary appointments device to an additional one, which we know from Hamilton is exactly what it is not. See Federalist No. 67 (Alexander Hamilton).
Regarding the Senate's advice-and-consent power, the dissent analogizes it to the president's veto power. Dissenting Op. at 254-55 & nn. 14-15. This analogy is inaccurate. The drafters of the Constitution rejected an approval mechanism proposed by Madison that gave the Senate only the power to veto presidential nominees by a majority vote in favor of “advice and consent.” 2 The Records of the Federal Convention of 1787 at 80-83 (Max Farrand ed., 1911); see also Matthew C. Stephenson, Can the President Appoint Principal Executive Officers without a Senate Confirmation Vote?, 122 Yale L.J. 940, 964-95 (2013). This means there is no reason to think that the balance of powers created through provisions of the advice-and-consent power to the Senate is anything like the president’s veto power. As we have explained, the balance is much more equitable between the branches and provides each the ability to negate the role of the other.
. The dissent rejects this conclusion on the ground that if the Senate is available any time it could act on nominations "if it had the desire[ 1 to do so,” then the Senate would logically always be available. Dissenting Op. at 267-68. This misses one central feature of pro forma sessions: the Senate has convened. We do not hold that the Senate is available any time when it could confirm nominations if it wanted to. Instead, we are pointing out that the Board cannot distinguish pro forma sessions from ordinary sessions on the basis of the Senate's availability because during pro forma sessions the Senate convenes in a manner that allows it to consent to nominations if it desires to. This is evidenced by the Senate’s passing of legislation during these sessions. Holding that the Senate is unavailable during these sessions requires a definition of availability that allows the counterintuitive situation in which the Senate is available to enact legislation while simultaneously unavailable to provide its advice and consent.
The dissent suggests one possibility, which is that the Senate is not available to provide its advice and consent during pro forma sessions because "business via unanimous consent agreement ... is not the type of business that yields the advice and consent envisioned by the Framers.” Id. at 257-58. Underlying this is the assertion that advice and consent requires a vote by the Senate’s members. Id. at 247. This is a complicated question. See Adam J. White, Toward the Framers’ Understanding of "Advice and Consent”: A Historical and Textual Inquiry, 29 Harv. J.L. & Pub Pol’y 103, 107-08, 147-48 (2005) (collecting sources arguing the Senate is required to act on nominations before analyzing the text and convention debates to conclude that the Senate has no obligation to act on presidential nominees). We are reluctant to express an opinion on it, especially because it has not been briefed.
Assuming that a vote is required to provide the Senate's advice and consent, however, it is also the case that the Senate must vote to “pass” a bill. See Chadha, 462 U.S. at 980-81, 103 S.Ct. 2764 (equating pass with vote). Why unanimous-consent agreements are sufficient to pass legislation, and thus constitute a vote, yet are inadequate to constitute a vote for the purpose of advice and consent is unclear. The dissent's definition thus suffers from the same flaw as the Board's: it cannot provide a principled method of defining availability.
. The Clause states:
Neither House, during the session of Congress, shall, without the consent of the other, adjourn for more than three days, nor to any other place than that in which the two Houses shall be sitting.
U.S. Const, art. I, § 5, cl. 4.
. The Adjournment Clause may be thought to create a problem for the intersession definition of recess. Namely, by requiring that the two chambers of Congress agree on any adjournment lasting longer than three days, the Clause enables the House to prevent the Senate from adjourning sine die. This would be problematic for the intersession definition because, as the argument goes, it inserts the House into the appointments process even though the Constitution purposely excludes it from the process.
The problem is eliminated, however, by Article II, § 3 of the Constitution. This provision allows the president to “adjourn both Houses” only “if the two Houses cannot agree on a date of adjournment.” U.S. Const. Art. II, § 3. Assuming that the Supreme Court would interpret adjourn to be the verbal form of adjournment, which it has said constitutes both inter- and intra-session breaks, Pocket Veto Case, 279 U.S. at 680, 49 S.Ct. 463, this *234provision allows the president to prevent the House from interfering in the appointments process if it prevents the Senate from adjourning for either an inter- or intra-session break.
. Another possible source of a durational limitation on recess is the Pocket Veto Clause, which provides that a bill passed by Congress becomes a law if the President takes no action on it for ten days "unless the Congress by their adjournment prevent its return." U.S. Const, art. I, § 7, cl. 2. The ten-day-duration requirement that might be drawn from this fails for the same reason the three-day-duration requirement fails in relation to the Adjournment Clause. Namely, the context of the Pocket Veto Clause is significantly different from the context of the Recess Appointments Clause, which means there is no reason to believe the former controls interpretation of the latter.
. The Senate Vacancies Clause stated in full: [I]f Vacancies [in the Senate] happen by Resignation, or otherwise, during the Recess of the Legislature of any State, the Executive thereof may make temporary Appointments until the next Meeting of the Legislature, which shall then fill such Vacancies.
U.S. Const, art. I, § 3, cl. 2, superseded by id. Amend. XVII.
. Correspondingly, this could mean that the break before that meeting — i.e., “the Recess of the Legislature” — did not necessarily have to be an intersession break. If this is the case, it is unlikely that recess was used in the same manner in the Senate Vacancies Clause as it is in the Recess Appointments Clause. Some words in the Constitution have different meanings "according to the connection in which [they are] employed” and "the character of the function” in which the word is found. Atlantic Cleaners & Dyers v. United States, 286 U.S. 427, 433-34, 52 S.Ct. 607, 76 L.Ed. 1204 (1932). The different meanings of recess would likely be necessary here to account for varying state procedures that may or may not have had formal sessions similar to the Senate.
. The Official Congressional Directory records fourteen sessions of Congress that have ended within a day of the Constitution’s automatic termination date. See Congressional Directory for the 112th Congress 522-38 (2011). This directory was completed before the end of the 2011 session of Congress, so the inclusion of the session that ended on January 3, 2012, brings the total to fifteen. A session automatically ended the first Monday of December until the Twentieth Amendment changed it to January 3 in 1933. U.S. Const, art. I, § 4, cl. 2; Amend. XX. These are the only relevant terminations of Senate sessions because if the Senate ends their session by convening and then adjourning sine die, then the Senate has a chance to consider nominations while they are convened. For example, in 2003, the Senate had an intrasession break that lasted from November 25 until December 9. 149 Cong. Rec. 31985 (Nov. 25, 2003) (statement of Sen. McConnell). On December 9, they convened and adjourned sine die. 149 Cong. Rec. 32404 (Dec. 9, 2003) (statement of Sen. Frist). The Board points to this as one example of a session ending before the Senate has the chance to consider a president’s recess appointments. NLRB Ltr. Br. at 12-13. But, even though the recess ended on the same day the session did, when the Senate convened to adjourn sine die they conducted quite a bit of business — including the confirmation of fifty-two people as officers of the United States. 149 Cong. Rec. at 32404-OS.
Only in one instance has an intrasession break ended at the same time that a Senate session has. See 137 Cong. Rec. 36362-64 (Nov. 27, 1991 through Jan. 3, 1992) (recording the Senate's November 27, 1991 adjournment until January 3, 1992). Even there, however, the Senate still convened before the session ended and had the opportunity to conduct business if it had wanted to. For example, it received messages from the president regarding nominations, though it did not confirm anyone before adjourning sine die. See 137 Cong. Rec. at 36364.
Two were preceded by a series of pro forma Senate sessions. See 157 Cong. Rec. S8783-84 (daily ed. Dec., 17, 2011) (recording the unanimous consent agreement to a schedule of pro forma session); 154 Cong. Rec. 24802-OS (Dec. 12, 2008; Dec. 12, 2008; Dec. 16, 2008; Dec. 19, 2008; Dec. 23, 2008; Dec. 26, 2008; Dec. 30, 2008; Jan. 2, 2009) (holding a series of pro forma sessions from Dec. 13, 2008 through Jan. 2, 2009).
Eleven were preceded by the Senate conducting business. See 158 Cong. Rec. S8637-68 (daily ed. Jan. 2, 2013) (confirming presidential nominees and completing business from days immediately prior before adjourning pursuant to the Constitution); 141 Cong. Rec. 38549-38608 (Dec. 29, 1995; Dec. 30, 1995; Jan. 2, 1996; Jan. 3, 1996); 116 Cong. Rec. 43999-44129, 44346-44597 (Dec. 30, 1970; Dec. 31, 1970; Jan. 2, 1971) (adjourning sine die one day before the constitutional deadline of January 3 after completing business); 96 Cong. Rec. 17022-17121 (Jan. 2, 1951) (same); 87 Cong. Rec. 10138-10143 (Dec. 26, 1941; Dec. 30, 1941; Jan. 2, 1942) (same); 86 Cong. Rec. 13997-14000, 14003-07, 14011-46, 14058-59 (Dec. 26, 1940; Dec. 30, 1940; Jan. 2, 1941; Jan. 3, 1941) (conducting business several days before the session terminated by function of the Constitution on Janu-aiy 3, 1941); 63 Cong. Rec. 440-48, 450-52 *238(Dec. 2, 1922; Dec. 4, 1922) (conducting business on the first Monday of December, and the days preceding it, before adjourning sine die as required by the Constitution); 50 Cong. Rec. 6030-37, 6041-44, 6050-53 (Nov. 26, 1913; Nov. 29, 1913; Dec. 1, 1913) (same); 37 Cong. Rec. 520-25; 529-31; 542-44 (Dec. 4, 1903; Dec. 5, 1903; Dec. 7, 1903) (same); 6 Cong. Rec. 764-98, 799-805, 816— 17 (Nov. 30, 1877; Dec. 1, 1877; Dec. 3, 1877) (same); 38 Cong. Globe, 40th Cong., 1st Sess. 793-95, 802, 810-11, 816-17 (Nov. 26, 1867; Nov. 27, 1867; Nov. 29, 1867; Dec. 2, 1867) (same).
And one of these terminations of Congress's session was due to continued business by the House, even though the Senate had adjourned sine die earlier. See 125 Cong. Rec. 37605-06 (Dec. 20, 1979) (recording the Senate's sine die adjournment on December 20, 1979).
. The dissent argues that our interpretation of recess reads the modifier "intersession” into the Constitution, contrary to the Supreme Court’s admonition to avoid doing so. Dissenting Op. at 249-50. This misunderstands our reasoning. As we have shown, the ordinary meaning of recess could support any of the definitions asserted, including the intersession definition. Through analysis of historical usage, application of the Recess Appointment Clause’s purpose, and analysis of the relevant constitutional context, we hold that of the ordinary meanings, the Constitution uses the intersession definition of recess. In short, we do not read "intersession'' into the Constitution because — as the word is used in the document — "recess” means only intersession breaks.
This method is also seen in the dissent's reasoning, which defines recess to mean when the Senate is unavailable to provide its advice and consent. Id. at 245. Per the dissent’s logic, Judge Greenaway’s definition would read the Clause to be "the Recess of the Senate [in which it cannot provide its advice and consent].” This is best illustrated by the dissent’s acknowledgement that the Senate recesses when it goes to lunch but that these recesses do not fall within “Recess” as it is meant in the Constitution. Id. at 248-49. Adding "in which it cannot provide its advice and consent” to the Clause is not what we understand the dissent to do. Instead, our colleague argues that recess itself means moments in which the Senate cannot provide advice and consent. While we disagree with this conclusion, both the majority opinion and the dissent are engaged in the same task — • defining the word "recess.”
. The Eleventh Circuit also implicitly derives this presumption from the framework explained by Justice Jackson in Youngstown Sheet & Tube Company v. Sawyer, 343 U.S. 579, 72 S.Ct. 863, 96 L.Ed. 1153 (1952). See Evans, 387 F.3d at 1222. Evans does not discuss Youngstown, but it cites United States v. Allocco, 305 F.2d 704, 713 (2d Cir.1962), as support for the presumption. Evans, 387 F.3d at 1222. Allocco, in turn, relies on Youngstown to defer to executive practice regarding the meaning of "happens” in the Recess Appointments Clause. 305 F.2d at 713-14. Specifically, Allocco relied on Youngstown by using it as support for its interpretation of "happen” since the Second Circuit believed its interpretation as " 'a systematic, unbroken, executive practice, long pursued to *241the knowledge of the Congress and never before questioned,’ ” which “ ‘may be treated as a gloss on 'Executive Power’ vested in the President by § 1 of Art. II.’ ” Id. (quoting Youngstown, 343 U.S. at 610-11, 72 S.Ct. 863). While we are unsure whether the executive practice before the Allocco Court regarding the meaning of "happen” is actually "systematic” and "unbroken,” we are sure that the executive practice regarding the meaning of "the Recess of the Senate” is not. As. discussed, the modem executive practice is contrary to executive practice before 1921 and has only become commonly used in the past thirty years. Furthermore, Congress has questioned presidents’ practices by, for example, holding pro forma sessions in an effort to stop it. We consider the Eleventh Circuit's reliance on Allocco as support for a presumption of constitutionality in separation-of-power cases unpersuasive.
. The dissent makes a form of this argument as well, arguing that the intersession-break definition of recess is "unworkable,” "not judicially manageable,” impracticable, and leads to absurd results. Dissenting Op. at 268-69. We disagree that the definition is unworkable, unmanageable, or impracticable; whether a break is intersession or not is a simple matter of reviewing how the Senate has adjourned. We also disagree that the result of a president’s recess appointment being valid one day and not the next is absurd, id. at 268-69, because this is a necessary result of defining recess. The dissent’s own definition, for example, would have this effect: a presidential recess appointment presumably would be valid on January 22, 2012, because the Senate did not convene at all on that day, see 158 Cong. Rec. Sil (Jan. 20, 2012) (adjourning until Monday, January 23, 2012); but be invalid if made on January 23, 2012, because the Senate became available by convening for a non -pro forma session, 158 Cong. Rec. S13 (daily ed. Jan. 23, 2012). Absurdity is also not clearly shown from the intersession-break definition’s allowance of recess appointments during intersession recesses that last vety short periods. Id. at 54. It is a result that has been rejected only by one 110-year-old Senate Committee Report— no president, court, or scholar has rejected the possibility. Cf. Hartnett, supra, at 406 (“All agree that recesses between sessions ... give rise to the President’s recess appointment power.”).
. Hartnett, supra, at 416 & nn. 176-77 (describing President Theodore Roosevelt’s recess appointment of 160 officers during an intersession break that lasted mere seconds).
. Accordingly, we do not have to address the conflict between the Second, Ninth, and Eleventh Circuits and the D.C. Circuit over the definition of “happen.” Compare United States v. Allocco, 305 F.2d 704, 709-12 (2d Cir.1962) (defining "happen” to mean "to exist”); United States v. Woodley, 751 F.2d 1008, 1009-13 (9th Cir.1985) (same); Evans, 387 F.3d at 1226-27, with Noel Canning, 705 F.3d at 507-14 (defining "happen” to mean "to occur”).
. No party argues that “the Recess” should be limited only to intrasession recesses, and neither do I.