This case arises out of the devastating tragedy that occurred at the World Trade Center ("WTC") in lower Manhattan, New York, on the morning of September 11, 2001. At issue in this case is the amount of insurance that is recoverable for the total destruction of the WTC that occurred after the buildings were struck by two fuel-laden aircraft that had been hijacked by terrorists. The appellants are numerous entities that have varying property interests in the WTC, including the Port Authority of New York and New Jersey (the "Port Authority"), which owns the property in fee simple, and Silverstein Properties, Inc. and several related entities ("Silverstein Properties"). In the spring of 2001, Silverstein Properties was the successful bidder on a 99-year lease for the property from the Port Authority. In July 2001, Silverstein Properties obtained primary and excess insurance coverage for the WTC complex from about two dozen insurers (most of which constitute the appellees and other counter-defendants in this case) in the total amount of approximately $3.5 billion "per occurrence." Because Silverstein Properties is the party that actually obtained the insurance coverage at issue in this case and was the primary insured, for ease of reference all appellants will hereafter be referred to collectively as the "Silverstein Parties."

The parties do not dispute that the destruction of the WTC resulted in a loss that greatly exceeded $3.5 billion. The broad question presented in this case is whether the events of September 11, 2001 constituted one or two "occurrences." The answer will determine whether the Silverstein Parties can recover once, up to $3.5 billion, or twice, up to $7 billion, under the insurance coverage. Complicating the resolution of this question is the fact that as of September 11, 2001, only one of the many insurers that bound coverage on the WTC had issued a final policy, necessitating an individualized inquiry to determine the terms of the insurance binders issued by each insurer.

This litigation began on October 22, 2001 when one of the WTC insurers, plaintiff-counter-defendant-intervenor SR International Business Insurance ("SR International"), filed suit against the Silverstein Parties "seek[ing] a judicial declaration of its rights and obligations to all of the insureds under the policy" and a "declaration that the damage to the World Trade Center is one insurance loss." The Silverstein Parties subsequently filed counterclaims against the other WTC insurers, seeking a declaration "that the events of September 11th constituted more than one occurrence under the coverage that the counterclaim-defendant[s] agreed to provide to the Silverstein Parties."1 After an initial assignment to another judge, the action was assigned to District Judge John S. Martin Jr. of the United States District Court for the Southern District of New York for all purposes.

In the first of these related appeals, the Silverstein Parties appeal from three judgments, made final and appealable pursuant to Fed.R.Civ.P. 54(b), granting summary judgment in favor of appellees Hartford Fire Insurance Company ("Hartford"), Royal Indemnity Company ("Royal"),2 and St. Paul Fire & Marine Insurance Company ("St.Paul"), respectively, in which the district court held that (a) the binders they issued were governed by the insurance policy form circulated by Silverstein Properties' insurance broker, and (b) under the definition of "occurrence" in that form, the destruction of the WTC was one occurrence as a matter of law. In the second appeal, the Silverstein Parties appeal an interlocutory order of the district court, certified to this court pursuant to 28 U.S.C. § 1292(b), denying the Silverstein Parties' motion seeking summary judgment against appellee Travelers Indemnity Company ("Travelers") based on the argument that the events of September 11, 2001, constituted two occurrences as a matter of law under the undefined term "occurrence" contained in Travelers' insurance policy. This court issued orders granting the Silverstein Parties' petition for leave to appeal under 28 U.S.C. § 1292(b) and their motion to have the § 1292(b) and Rule 54(b) appeals heard in tandem.

For the following reasons, the judgments of the district court are affirmed.

As a condition of its 99-year lease of the WTC, Silverstein Properties was required to obtain first-party property insurance on the property. Silverstein Properties engaged Willis of New York ("Willis"), an insurance broker, to set up a multi-layered insurance program, which consisted of a primary insurance layer and 11 excess insurance layers providing a total of approximately $3.5 billion insurance on a "per occurrence" basis. In soliciting insurers for the program, Willis circulated a Property Underwriting Submission (the "Underwriting Submission") containing information regarding the proposed placement, including descriptions of the property and the insureds, desired coverage terms and conditions, estimated property values, engineering information, and a property loss history. With respect to at least the four insurers involved in these appeals, the Underwriting Submission also included a specimen copy of Willis's own "broker" form (the "WilProp form").3 Section VIII of the Underwriting Submission states: Policy Form and Contract between Silverstein and the [Port Authority] are attached. DRAFT WilProp for Real Estate Risks is attached. We anticipate that this form will ultimately require amendment to comply with the Contract between Silverstein Properties, Inc. and the [Port Authority]. In the meantime, we provide this document as a starting point.

Of the four insurers in these appeals, Travelers was the only insurer to submit its own specimen policy form (the "Travelers form") during the course of negotiating the terms of coverage.4 Whereas the Travelers form did not define the term "occurrence," the WilProp form defined occurrence as follows:

"Occurrence" shall mean all losses or damages that are attributable directly or indirectly to one cause or to one series of similar causes. All such losses will be added together and the total amount of such losses will be treated as one occurrence irrespective of the period of time or area over which such losses occur.

SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props. LLC, 222 F. Supp. 2d 385, 398 (S.D.N.Y.2002).

As we will explain in greater detail, each of the appellees negotiated separately with Willis concerning its participation in the insurance program and all had bound coverage on various layers as of July 20, 2001. During the course of the next several weeks, Willis negotiated with Travelers over the terms of its final policy, but the Silverstein Parties have presented no evidence to indicate that any of the other appellees participated in or were aware of the details of those negotiations. As of September 11, 2001, none of the appellee-insurers had issued a final policy form, nor had Willis issued the WilProp form as a final policy form, although at least one other participating insurer, Allianz Insurance Company ("Allianz"), had issued a final policy. SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props. LLC, No. 01 Civ. 9291(JSM), 2002 WL 1163577, at *2 & n. 3 (S.D.N.Y. June 3, 2002); see also SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props. LLC, No. 01 Civ. 9291(JSM), 2003 WL 192487, at *4 (S.D.N.Y. Jan.29, 2003). On September 14, 2001, three days after the WTC was destroyed, following discussions between Willis and Travelers, Travelers issued its final policy form.

As a threshold matter, we address the basis for federal subject matter jurisdiction over these exclusively state law claims involving insurance coverage and contract interpretation. "An inquiry respecting [jurisdiction] is one we always have the power to undertake, and where jurisdiction is questionable we are obliged to examine it sua sponte." Petereit v. SB Thomas, Inc., 63 F.3d 1169, 1175 (2d Cir.1995). Although we are satisfied that jurisdiction exists here, we discuss the issue because of some unusual circumstances presented in this case.

The district court did not specifically address jurisdiction in its opinion denying summary judgment. It apparently was of the view, however, that the action against Travelers was separate from the SR International litigation and related counterclaims, and noted that Travelers is a Connecticut company with its principal place of business in Connecticut, while the entities comprising Silverstein Properties are all either Delaware or New York companies and have their principal places of business in New York. We need not decide if there was jurisdiction over the separate action initially filed by Silverstein Properties against Travelers, however, because in March 2002, prior to the district court's decision denying summary judgment, that action was withdrawn without prejudice by stipulation of the parties and Travelers was added as a counterclaim defendant to the action brought against the Silverstein Parties by SR International. See supra n. 1. Accordingly, we address only whether there exists federal jurisdiction over the latter action.

According to the Silverstein Parties, federal jurisdiction exists in this case because there is complete diversity between all of the defendants and plaintiff SR International and supplemental jurisdiction applies to defendants' counterclaims against Travelers and the other insurers. A review of the parties' citizenship5 appears to support this assertion. Plaintiff SR International is a foreign corporation (a citizen of the United Kingdom), but this does not preclude diversity because 28 U.S.C. § 1332(a)(2) provides federal jurisdiction over actions between "citizens of a State and citizens or subjects of a foreign state."

Defendant Wells Fargo is a national bank (i.e., not incorporated in any one state) and by statute is deemed to be a citizen of every state in which it has offices. 28 U.S.C. § 1348 ("All national banking associations shall, for the purposes of all [ ] actions by or against them [other than a few enumerated in the statute], be deemed citizens of the States in which they are respectively located."); see United Republic Ins. Co., in Receivership v. Chase Manhattan Bank, 315 F.3d 168, 169 (2d Cir.2003) (per curiam) (recalling mandate, vacating prior summary order, and remanding to district court to determine whether diversity jurisdiction existed under 28 U.S.C. §§ 1332 & 1348 in light of fact that defendant banks had branch offices in plaintiff's home state); but see Lerner v. Fleet Bank, N.A., 318 F.3d 113, 124-25 (2d Cir.2003) (implicitly assuming that national bank was a citizen of New York, presumably its principal place of business, for purposes of diversity). The fact that Wells Fargo may be a citizen of several unidentified states does not affect diversity here, however, because SR International is a foreign citizen.

Defendant Port Authority presents a closer question on jurisdiction because it is a state-created body, thereby raising the possibility that it is a not a "citizen" of any state, the effect of which would be to destroy diversity. The Port Authority is a body "corporate and politic" established in 1921 pursuant to a bi-state compact between New York and New Jersey, see N.Y. Unconsol. Law § 6407 (McKinney 2000), and assented to by Congress, see 42 U.S. Stat. 174 (1921). The Port Authority's mission was, and remains, the development of public transportation, terminal, and other facilities of commerce within the statutorily defined Port Authority district, which includes the area in and around New York City harbor. See N.Y. Unconsol. Law § 6403. The Port Authority is governed by a board of commissioners, see id. §§ 6405-06, whose resolutions are essentially legislative acts of the bi-state entity that must be approved by the governors of both states. See id. §§ 7151-52.

Baron v. Port Auth., 271 F.3d 81, 83 (2d Cir.2001). We have found no case addressing the question of whether the Port Authority can be considered a "citizen" for purposes of diversity jurisdiction, and courts that have addressed this question with respect to other port authorities and similar agencies in other states have reached different conclusions based largely on the level of autonomy enjoyed by the agency. Compare Indiana Port Comm'n v. Bethlehem Steel Corp., 702 F.2d 107, 110 (7th Cir.1983) (holding that "[Indiana Port Commission] is not the `alter ego' of the state of Indiana, and is, in fact, an independent corporate entity seeking to assert its own rights in this action," and, therefore, there was diversity jurisdiction), and C.H. Leavell & Co. v. Bd. of Comm'rs of Port of New Orleans, 424 F.2d 764, 767 (5th Cir.1970) (holding that "[Louisiana] Dock Board is a sufficiently separate entity from the State of Louisiana to sustain diversity jurisdiction") with Tradigrain, Inc. v. Mississippi State Port Auth., 701 F.2d 1131, 1134 (5th Cir.1983) (finding, based on statutes and other factors, that "Mississippi State Port Authority is merely the alter ego of the State of Mississippi, and as such is not a `citizen' for purposes of diversity jurisdiction").

Although whether the Port Authority is a "citizen" of New York for diversity purposes is apparently an issue of first impression in our circuit, case law has established that it is a political subdivision of the state and, therefore, is not entitled to sovereign immunity. See Feeney v. Port Auth. Trans-Hudson Corp., 873 F.2d 628, 631 (2d Cir.1989) (holding that Eleventh Amendment immunity was inapplicable to "PATH [in part because of] the fact that the compact between New York and New Jersey describes the Port Authority as a `municipal corporate instrumentality,' N.Y. Unconsol. Laws § 6459 (McKinney 1979) and N.J. Stat. Ann. § 32:1-33 (West 1963), language consistent with its being a political subdivision"); see also Japan Airlines Co. v. Port Auth., 178 F.3d 103, 110-12 (2d Cir.1999) (holding, after extensive analysis, "that the Port Authority was not entitled to [either sovereign or] governmental immunity [for case arising out of proprietary conduct], and the district court properly instructed the jury to treat the Port Authority as it would any private corporation"). In light of these rulings, we hold that, as is the case with political subdivisions in general, the Port Authority is not so closely aligned with the two state governments that created it to foreclose its being treated as a citizen of both New York and New Jersey for diversity purposes. See Illinois v. City of Milwaukee, 406 U.S. 91, 97, 92 S. Ct. 1385, 31 L. Ed. 2d 712 (1972) ("It is well settled that for the purposes of diversity of citizenship, political subdivisions are citizens of their respective States."). Accordingly, we agree with the Silverstein Parties that complete diversity exists between plaintiff SR International and all of the defendants in this action.

We also agree with the Silverstein Parties that supplemental jurisdiction provides jurisdiction over the Silverstein Parties' counterclaims against the other insurers even though some of these insurers may be non-diverse. See 28 U.S.C. § 1367 (setting forth requirements for supplemental jurisdiction); Viacom Int'l, Inc. v. Kearney, 212 F.3d 721, 726-27 (2d Cir. 2000) (observing that § 1367 permits defendants to assert claims against non-diverse third parties in diversity cases); David Siegel, Practice Commentary, "The 1990 Adoption of § 1367, Codifying `Supplemental' Jurisdiction," printed in 28 U.S.C.A. § 1367 at 832-33 (West 1993) (same); see also Owen Equip. & Erection Co. v. Kroger, 437 U.S. 365, 374-77 & n. 18, 98 S. Ct. 2396, 57 L. Ed. 2d 274 (1978) (holding that whereas plaintiff's claims against non-diverse parties destroyed diversity jurisdiction, ancillary (now supplemental) jurisdiction supported defendant's counterclaims and third-party claims against non-diverse parties); Herrick Co. v. SCS Communications, Inc., 251 F.3d 315, 325 n. 8 (2d Cir.2001) ("The effect of Kroger was therefore `to limit ancillary jurisdiction primarily to claims asserted by parties in a defensive posture, or who did not choose the federal forum. Therefore, at least in diversity cases, ancillary jurisdiction usually is not available for claims asserted by the plaintiff.'") (quoting 13 Charles Alan Wright, Arthur R. Miller and Edward H. Cooper, Federal Practice and Procedure: Jurisdiction § 3523) (emphasis added).

The Silverstein Parties also assert that there is federal jurisdiction under the Air Transportation Safety and System Stabilization Act of 2001, Pub.L. No. 107-42, § 408(b)(3), 115 Stat. 230, 241 (Sept. 22, 2001), as amended by the Aviation and Transportation Security Act of 2001, Pub.L. No. 107-71, § 201, 115 Stat. 597, 645 (Nov. 19, 2001).6 That Act, originally passed by Congress to limit the liability of air carriers for any claims arising from the September 11th attacks, was later amended to extend its protection to, inter alia, any "person with a property interest in the World Trade Center." Id. at § 408(a)(1). The purpose of the Act is to cap the liability of various entities for damages and contribution claims to the limits of their liability insurance coverage. The Act also creates a federal cause of action for any damages claims arising out of the September 11th attacks. Id. at § 408(b)(1). In addition, the Act grants to the District Court for the Southern District of New York "original and exclusive jurisdiction over all actions brought for any claim (including any claim for loss of property, personal injury, or death) resulting from or relating to the terrorist-related aircraft crashes of September 11, 2001." Id. at § 408(b)(3).

While the other parts of § 408 apply only to liability claims brought against air carriers, World Trade Center property interest holders, and similar entities, the jurisdictional grant of § 408(b)(3) is considerably broader, and its plain language would appear to encompass the instant claims filed by the World Trade Center property interest holders against their insurers — claims that are clearly "related to" the September 11th attacks. The very breadth of this jurisdictional grant, however, raises the question of whether it exceeds the constitutional limitations on Congress's authority to grant jurisdiction to federal courts. See Verlinden B.V. v. Cent. Bank of Nigeria, 461 U.S. 480, 491, 103 S. Ct. 1962, 76 L. Ed. 2d 81 (1983) (noting that "Congress may not expand the jurisdiction of the federal courts beyond the bounds established by the Constitution," such as to purely state-law claims). In fact, in the only opinion in this litigation to discuss jurisdiction, the district court, while addressing a different issue, noted that there is "a serious question whether the grant of jurisdiction in the Act applies to this case," and that to avoid the constitutional question presented, it "would be inclined to construe the Act's grant of jurisdiction as not extending to these claims between the insurers and their insureds." SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props. LLC, 2002 WL 1905968, at *2 (S.D.N.Y. Aug.19, 2002) (holding, after analyzing the statute, that the parties' contractual appraisal procedure for determining the amount of loss was not preempted by the Act). The district court ultimately concluded, "[h]owever, [that it] need not decide whether Congress either intended to or could vest this Court with exclusive jurisdiction over an action between the Silverstein Parties and their insurers." Id. at *3.

We recently acknowledged these same constitutional concerns in a case between foreign reinsurers involving a breach of contract claim arising out of the September 11th terrorist attack, in which we observed that to construe the statute "to encompass all claims for economic loss" relating to the September 11th attacks would raise "serious doubts as to its constitutionality." Canada Life Assurance Co. v. Converium Rückversicherung (Deutschland) AG, 335 F.3d 52, 59 (2d Cir.2003). We declined to "delineate the precise contours of Section 408(b)(3)'s jurisdictional grant," finding it clear that the case then before us, as to which there was no claim or defense that would "require adjudication of any issue of law or fact that concern[ed] the events of September 11," fell outside the statute. Id. at 57.

Because, as discussed above, this federal action is supported by diversity and supplemental jurisdiction, we need not and do not decide whether there would also be jurisdiction under the Act and thereby avoid unnecessarily addressing these constitutional concerns. Cf. Edward J. DeBartolo Corp. v. Fla. Gulf Coast Bldg. & Constr. Trades Council, 485 U.S. 568, 575, 108 S. Ct. 1392, 99 L. Ed. 2d 645 (1988) ("[W]here an otherwise acceptable construction of a statute would raise serious constitutional problems, the Court will construe the statute to avoid such problems unless such construction is plainly contrary to the intent of Congress."); United States v. Arrous, 320 F.3d 355, 360 (2d Cir.2003) ("[W]e avoid interpreting statutes in a way that may create constitutional problems. . . .").

This court has jurisdiction to hear the Silverstein Parties' appeal from the grant of summary judgment in favor of Hartford, Royal, and St. Paul, notwithstanding the lack of a final judgment disposing of all claims against all parties, because the district court entered judgment pursuant to Fed.R.Civ.P. 54(b).7 We also have jurisdiction over the Silverstein Parties' interlocutory appeal from the denial of their motion for summary judgment against Travelers because the district court certified its decision for interlocutory appeal under 28 U.S.C. § 1292(b).8 We granted the Silverstein Parties' petition for leave to appeal the denial of summary judgment against Travelers pursuant to 28 U.S.C. § 1292(b) and their motion to expedite and consolidate the appeal with the Silverstein Parties' Rule 54(b) appeal.

II. RULE 54(B) APPEAL — HARTFORD, ROYAL, AND ST. PAUL

"We review the grant or denial of summary judgment de novo." Gibbs-Alfano v. Burton, 281 F.3d 12, 18 (2d Cir.2002) (quoting Republic Nat'l Bank v. Delta Air Lines, 263 F.3d 42, 46 (2d Cir.2001)). Summary judgment is appropriate only where "there is no genuine issue as to any material fact and . . . the moving party is entitled to a judgment as a matter of law." Fed.R.Civ.P. 56(c). Thus, even where facts are disputed, in order to defeat summary judgment, the non-moving party must offer enough evidence to enable a reasonable jury to return a verdict in that party's favor. See Anderson v. Liberty Lobby, 477 U.S. 242, 248, 106 S. Ct. 2505, 91 L. Ed. 2d 202 (1986). When ruling on a motion for summary judgment, a court is required to construe the evidence in the light most favorable to the non-moving party and to draw all reasonable inferences in its favor. See id. at 255; Maguire v. Citicorp Retail Servs., Inc., 147 F.3d 232, 235 (2d Cir.1998).

The district court granted the motions by Hartford, Royal, and St. Paul seeking partial summary judgment on the grounds that each of the insurers had issued a binder that incorporated the terms of the WilProp form and that under the WilProp form's definition of "occurrence" there was only one occurrence on September 11, 2001. See SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props. LLC, 222 F. Supp. 2d 385, 393-95, 398-99 (S.D.N.Y.2002) ("Hartford Dec."). The district court found that the only relevant question was what were the terms to which the parties had agreed to be bound pending issuance of a final policy, and thus deemed irrelevant evidence indicating that the parties might have agreed to ultimately issue policies tracking the Travelers policy. Id. at 389. The district court then reviewed each of the negotiating histories between Willis and Hartford, Royal, and St. Paul and analyzed the language of their binders. In part because the only policy form before the parties during these negotiations was the WilProp form furnished by Willis, the district court concluded that as a matter of law each of the three insurers had bound coverage on the basis of the WilProp form, rather than, as the Silverstein Parties contended, the Travelers form. Id. at 393-95, 398. The district court further held that, as a matter of law, under the WilProp form's definition of "occurrence" the damage caused on September 11th was the result of one occurrence, entitling the Silverstein Parties to no more than a single policy limit on each of the insurers' policies. Id. at 399.

As noted by the district court, "[a]n insurance binder is a unique type of contract." Id. at 388.

It is a common and necessary practice in the world of insurance, where speed often is of the essence, for the agent to use this quick and informal device to record the giving of protection pending the execution and delivery of a more conventionally detailed policy of insurance. Courts, recognizing that the cryptic nature of binders is born of necessity and that many policy clauses are either stereotypes or mandated by public regulation, are not loath to infer that conditions and limitations usual to the contemplated coverage were intended to be part of the parties' contract during the binder period.

Employers Commercial Union Ins. Co. v. Firemen's Fund Ins. Co., 45 N.Y.2d 608, 412 N.Y.S.2d 121, 384 N.E.2d 668, 670 (N.Y.1978) (footnote omitted) (quoted in Hartford Dec., 222 F.Supp.2d at 389). Thus, a binder is "a short method of issuing a temporary policy for the convenience of all parties, to continue until the execution of the formal one." Lipman v. Niagara Fire Ins. Co., 121 N.Y. 454, 24 N.E. 699, 700 (1890) (quoted in Hartford Dec., 222 F.Supp.2d at 388). As the New York Court of Appeals has explained,

[i]t has long been settled in this State that an insurance binder is a temporary or interim policy until a formal policy is issued. A binder provides interim insurance, usually effective as of the date of application, which terminates when a policy is either issued or refused.

Springer v. Allstate Life Ins. Co., 94 N.Y.2d 645, 710 N.Y.S.2d 298, 731 N.E.2d 1106, 1108 (2000) (internal citations omitted). While not all of the terms of the insurance contract will be set forth in the binder, a binder is nevertheless a fully enforceable "present contract of insurance." Ell Dee Clothing Co. v. Marsh, 247 N.Y. 392, 160 N.E. 651, 652 (1928).

On appeal, the Silverstein Parties argue that in construing the binders issued by the appellee insurers, the district court erred in rejecting evidence of the insurers' agreement to "follow the [Travelers] form" and that this evidence creates material factual issues in dispute as to whether the WilProp definition of "occurrence" applies to these insurers. The Silverstein Parties explain the practice of "following the form" as follows:

Whether or not the broker includes a sample policy form with the submission, the industry practice in layered placements is for a lead insurer to act as the negotiator of policy terms on behalf of the participating insurers. In an effort to achieve concurrency (uniformity in coverage terms provided by the participating insurers), the other participating primary and excess insurers customarily agree to "follow the form" of the lead insurer, i.e., to accept the terms and conditions of the program policy. The lead insurer typically emerges from among the carriers on the primary coverage layer; insurers participating solely in excess layers generally do not negotiate the program policy wording.

Appellants' Br., Dkt. No. 02-9279 ("Appellant's Rule 54(b) Br."), at 20 (internal citations omitted).9

According to the Silverstein Parties' theory of the case, appellees Hartford, St. Paul, and Royal agreed to "follow the [Travelers] form" and, therefore, are bound by the terms of the policy that Travelers issued on September 14, 2001, three days after the World Trade Center was destroyed. In support of this theory, the Silverstein Parties submitted a substantial amount of custom and usage evidence from both expert and fact witnesses concerning the practice of "following the form." The Silverstein Parties further assert that there is evidence demonstrating that each of the three appellees knew the WilProp form was only a starting point for negotiations and that they learned before binding that the Travelers form would be the lead policy. Accordingly, they maintain, appellees' obligations are to be determined under the final policy issued by Travelers on September 14, 2001 (the "September 14 Travelers Policy").

Contrary to the arguments so vigorously pressed by the Silverstein Parties, however, the question in this case is not whether appellees are bound by the September 14 Travelers policy rather than the WilProp policy circulated by Willis in its Underwriting Submission. As of September 11, 2001, none of the insurers could be bound by either policy because neither one had been issued as the final policy by that date.10 See, e.g., Springer, 710 N.Y.S.2d 298, 731 N.E.2d at 1108 (noting that binder and final policy are "two distinct agreements"); Rosenblatt v. Washington County Coop. Ins. Co., 191 A.D.2d 883, 594 N.Y.S.2d 456, 459 (3rd. Dept. 1993) (holding in case where loss occurred prior to the issuance of final policy that "the dispositive issue . . . concerns the risks covered under the binder agreement . . . and this is unaffected by any changes regarding insurance policies issued subsequent to the loss"); Del Bello v. Gen. Accident Ins. Co., 185 A.D.2d 691, 585 N.Y.S.2d 918, 918-19 (4th Dept. 1992) (holding that because binder identified plaintiff as an insured, plaintiff was covered for fire damage even though not so listed in final policy issued post-loss); see also Cardinal v. Mercury Ins. Co., 242 A.D. 98, 273 N.Y.S. 487, 490-91 (3rd.Dept. 1934) (reforming policy to reflect terms of original oral binder for purposes of pre-policy loss), rev'd on other grounds, 266 N.Y. 448, 195 N.E. 148 (1934).

While the Silverstein Parties may well be correct that in many instances an excess insurer will voluntarily bind itself to another insurer's policy form that has been issued but that it has chosen not to look at despite the opportunity to do so, and that courts will presume the insurer knows and assents to the terms of the unseen document, see, e.g., Progressive Cas. Ins. Co. v. C.A. Reaseguradora Nacional De Venezuela, 991 F.2d 42, 45-47 (2d Cir.1993), it is an entirely different matter for a court of law to impose liability on the basis of another insurer's policy form that has not been issued and, therefore, that the excess insurer has never had the opportunity to review. See, e.g., Designcraft Jewel Indus., Inc. v. Rampart Brokerage Corp., 63 A.D.2d 926, 406 N.Y.S.2d 97, 98 (1st.Dept. 1978) (refusing to reform excess policy or hold excess insurer liable for excess loss arising from primary policy's disaggregation of building's primary insurance limits on per-floor basis where excess insurer was neither informed about disaggregation term nor shown underlying primary policy prior to issuance of binder), aff'd by, 46 N.Y.2d 981, 415 N.Y.S.2d 991, 389 N.E.2d 472 (N.Y.1979). The Silverstein Parties acknowledge as much when they state in their brief:

Once the final primary policy is issued, it is customarily sent to all participating insurers for their review. Unless it has waived its agreement to policy language, if an insurer that has agreed to "follow the form" has good-faith objections to the form that has emerged from that negotiation and those objections cannot be resolved consensually, industry custom allows such an insurer to cancel its coverage prospectively after reasonable prior notification to the insured so that the insured — if it is unwilling to acquiesce in the change — can arrange for a replacement insurer. The participating insurer may not, however, cancel retroactively if a loss occurs before the final policy form is issued and presented for its approval.

Appellant's Rule 54(b) Br. at 26-27 (internal citations and footnote omitted). This passage makes manifest the counterintuitive nature of the Silverstein Parties' position that an insurer who agrees in a binder to "follow form" is thereby bound sight unseen to a policy that has not been finalized and issued, but only until the insurer is actually given an opportunity to see the policy and review it, at which point it can cancel its coverage and no longer be bound. The Silverstein Parties' further observation that the insurer cannot cancel retroactively, although offered as support for their claim that the insurer is bound by the final policy form with respect to any losses that occurred before the final policy issued, is gratuitous because it is well settled that the insurer is bound retrospectively not by the final policy form, but by the binder, a distinct agreement the terms of which the insurer negotiated for itself. Despite the Silverstein Parties' arguments to the contrary, therefore, we conclude that the September 14 Travelers policy, issued three days after the loss at issue here, has no bearing on the Rule 54(b) appeal.

The only question we must decide is what the term "occurrence" means under the specific binders that appellees issued and that were in force when the planes destroyed the WTC on September 11, 2001. Should we infer that the parties to the binder intended to: (1) incorporate the specific definition of "occurrence" contained in the WilProp policy; (2) forgo a specific definition (as is the case in the customary Travelers form); or (3) treat "occurrence" in some other fashion? The evidence offered by the Silverstein Parties to demonstrate that the appellees agreed that they would "follow the [Travelers] form" at such time as it might issue is relevant only if, and to the extent that, the facts of the parties' pre-binder negotiations can support a finding that the parties intended the insurer's binder, the policy that was to be in effect until the Travelers form was issued, to incorporate the terms, not of the September 14 Travelers policy ultimately issued, but of Travelers' customary or specimen form (the "Travelers form").

As we have explained, a binder, while an enforceable contract in its own right, is necessarily incomplete in some respect (otherwise a subsequent formal policy would be unnecessary). Thus, as the New York courts have long recognized, terms must often be implied to determine the obligations to which the parties intended to be bound. Hicks v. British Am. Assurance Co., 162 N.Y. 284, 56 N.E. 743, 744 (1900) ("The law reads into the contract the standard policy, whether it be referred to in terms or not."); Lipman, 24 N.E. at 700 (construing binder as necessarily incorporating terms of a standard policy). "To determine the contents of a binder, New York courts generally look to (1) the specific terms contained in the binder or incorporated by reference, and (2) to the extent necessary as gap-fillers, the terms included in the usual policy currently in use by the insurance company" or those required by statute. Lapenta v. Gen. Acc. Fire & Life Assurance Corp., 62 A.D.2d 1145, 404 N.Y.S.2d 182, 184 (4th Dept. 1978); see Sherri v. Nat'l Sur. Co., 243 N.Y. 266, 153 N.E. 70, 71 (1926) (noting that a binder incorporates all the obligations "according to the terms of the policy in ordinary use by the [issuing] company") (internal quotation marks omitted); see also Ell Dee Clothing, 160 N.E. at 653 ("[W]hether the binder is to be interpreted by itself or with the addition of implied conditions, the minds of the parties meet. And in the absence of state regulations, it is for the assurer to show that conditions are implied and what they are."). We agree, therefore, with the district court's observation that "[t]he terms of a binder are not left to future negotiation. . . . The law of New York with respect to binders does not look to the negotiations of the parties to see what terms might ultimately have been incorporated into a formal policy." Hartford Dec., 222 F.Supp.2d at 388-89. Rather, the negotiations are examined to determine what terms the parties intended to incorporate into the binder.

In deciding which terms are to be implied in a binder, reliance may be placed on the extrinsic evidence of the parties' pre-binder negotiations. In particular, we believe that any policy form that was exchanged in the process of negotiating the binder, together with any express modifications to that form, is likely the most reliable manifestation of the terms by which the parties intended to be bound while the binder was in effect. In the absence of such a policy form underlying the negotiations or sufficient extrinsic evidence of the negotiations to determine the parties' intentions, the terms to be implied would likely be the customary terms of the insurer's own form, see Sherri, 153 N.E. at 71; Hicks, 56 N.E. at 744; Lipman, 24 N.E. at 700, unless there is evidence indicating that an understanding existed between the parties that a different policy form would apply to the binder and that the insurer was aware of its terms.

Therefore, with regard to the case at hand, whether an insurer in the WTC program agreed to "follow the [Travelers] form" is largely irrelevant to the inquiry of what terms should be implied in that insurer's binder unless it can be shown that the insurer was provided with a Travelers form (or some other form omitting a definition for occurrence) prior to issuing its binder. In the absence of such evidence, we believe that the fact that an insurer agreed to follow the lead of Travelers and demonstrated an intention to be bound by the final policy form as ultimately negotiated by Travelers would be relevant only to the parties' post-binder relationship, which is of no import to this case. Such an agreement or understanding, whether explicit or derived from custom and usage, would not provide a basis for incorporating into the binder the terms contained in the Travelers form.

Applying this analytical framework to the evidence presented on summary judgment, we agree with the district court that there can be no genuine dispute here that the binders issued by Hartford, Royal, and St. Paul were issued on the basis of negotiations involving the WilProp form, a copy of which had been provided to each insurer by Silverstein Properties' insurance broker, Willis, and that the parties intended and understood the binders to incorporate the terms of the WilProp form except as expressly modified.

Our conclusion is supported by the fact that until the total destruction of the WTC on September 11th, it was in Silverstein Properties' interest to incorporate into their insurance coverage a definition of "occurrence" that would minimize the number of "occurrences" in order to minimize the number of deductibles that would apply in the event of a loss or series of losses. This goal was accomplished by the WilProp form's inclusive definition of "occurrence." When Travelers held out for using its own form in its negotiations with Willis in August 2001, Timothy Boyd, a vice president of Willis, reported that fact to a co-broker, stating, "Although other players have signed binders based on WilProp, Travelers is insisting we use their form and this is under review." Apart from its potential as a party admission, the statement that "players [other than Travelers] have signed binders based on WilProp," made by the Silverstein Parties' agent on August 3, 2001 — after the binders were in place and before the WTC was destroyed — is consistent with our review of the binder negotiations, to which we now turn.

On June 7, 2001, Boyd of Willis sent John Gemma of Hartford an Underwriting Submission that included the WilProp form. Gemma acknowledged at his deposition that when he received these materials, he understood the WilProp form was only a draft and would ultimately require amendment. Gemma responded to Boyd with a proposal in which he offered to provide insurance with a limit of $50 million in excess of $75 million per occurrence. With respect to the applicable form, Gemma specified, "Manuscript Forms Submitted With Attached Amendments," enclosed several pages of amendments to the WilProp form and additional exclusions, and specifically noted that the policy was subject to additional exclusionary endorsements. Gemma made no change to the WilProp definition of occurrence. Boyd stated at his deposition that he understood that Gemma's reference to the manuscript form was to the WilProp form.

Boyd later asked Gemma to participate in a different layer of excess insurance. On July 9, Gemma sent Boyd a revised quote in identical form to the first for $50 million in excess of $50 million. On July 12, however, Gemma e-mailed Boyd that Hartford's participation would be limited to $25 million. At his deposition, Boyd testified that when he called to complain about the decrease in coverage, he informed Gemma that Travelers was insisting on the use of their policy form. According to Boyd, Gemma did not object when Boyd mentioned Travelers; he simply asked for a copy of the Travelers form once it was finalized. Gemma testified that Boyd never told him that the Travelers form was going to be the operative form. In response to Boyd's complaint about the decreased coverage, Gemma obtained approval from his superiors to raise the Hartford limit to $32 million. The e-mail from Gemma confirming the increased limit contained no other terms. Boyd e-mailed back that Willis was pleased to bind participation at $32 million, adding that "[w]e will issue formal documentation soonest." No specific terms other than premiums and the extent of coverage were included in the e-mail.

On July 19, 2001, Gemma sent Boyd a form he referred to as the "outline of our property BINDER." The binder outline specified that the policy form would be the "Manuscript Forms Submitted With Attached Amendments." Except for the coverage limits and certain other changes not relevant here, Boyd understood that Gemma was making the same offer he had made before. On July 20, 2001, Boyd sent Gemma an e-mail attaching a binder. The e-mail said, "We are working diligently with primary carriers to refine policy form. Attached is binder per specs." In a section denominated "Property and Time Element Covered," the binder stated, "And as incorporated into the manuscript form, in conjunction with the contract between the Port Authority of New York and New Jersey as ultimately agreed." Gemma made changes to the binder, initialed and signed it, and returned it to Boyd. He did not change the language in the "Property and Time Element Covered" section, but did change a portion of the exclusions section from "Per Policy Form as to be advised," to "Per Policy Form as quoted." Gemma acknowledged that by failing to strike out the language in the "Property and Time Element Covered" section, he agreed to that language, but he did not state that he understood the phrase to be referring to the Travelers form.

Hartford's New York underwriting office, which was located in the WTC, was destroyed on September 11th. Following the attack, Gemma asked Boyd to provide him with copies of all the documents and e-mails that he had sent to Boyd during the course of negotiations so that Hartford's files could be reconstructed. Gemma stated at deposition that he specifically asked Boyd to send him a copy of the marked-up WilProp form upon which he had based his quote and that Boyd did not question the relevance of the WilProp form to Hartford's coverage. Boyd testified, however, that he again mentioned he was working with the Travelers form and that Gemma asked for a copy as soon as possible. On October 1, 2001, Willis forwarded a number of e-mails to Gemma and his supervisor including an e-mail that attached the original June 7, 2001, Willis submission, including the WilProp form. Willis did not, however, send Gemma's marked-up copy. On the same date, a Willis representative sent Hartford another e-mail that contained as an attachment a document identified in the e-mail as the "underlying policy," which may have been a copy of the Traveler's policy; the e-mail further stated that "[t]he excess following form will be out very shortly."

On October 9, 2001, Boyd sent Gemma a fax, which read:

Attached is your original authorization culled from our files. As you know, we evolved from there, both on a participation level and on a forms level.

Ultimately your participation was amended and when we discussed forms, I informed you that the form we would use would be Travelers.

Accordingly, these forms [have been] sent under separate cover through claims channels. I have sent you the participant list and the agreed form via electronic mail. . . .

Boyd testified that he never received a response to this fax from Gemma.

Despite the fact that each of the documents generated during Willis' negotiations with Hartford prior to the issuance of a binder referenced a particular form that was understood by the participants at the time to be the WilProp form together with Gemma's amendments, the Silverstein Parties argue that a reasonable jury could find that Hartford bound coverage on the basis of the Travelers form in light of: (1) expert testimony that it was insurance industry practice for excess carriers bidding in a layered program to agree to "follow the form" and that the insurer must specify in the binder if it does not want to "follow the form"; (2) expert testimony that an agreement to follow form is typically evidenced by words in the binder such as "wording to be agreed," "as ultimately agreed," "to be advised," or "subject to review and acceptance [of a primary policy]"; (3) Boyd's testimony that he informed Gemma that Travelers was insisting on the use of its form and that Gemma did not object, and instead simply asked for a copy of the Travelers form; (4) the presence in the Hartford binder of the words, "[a]nd as incorporated into the manuscript form, in conjunction with the contract between the [Port Authority] as ultimately agreed"; and (5) Gemma's failure to object when Boyd said to him after September 11th that he had previously informed him that Travelers was the governing policy.

None of these arguments has merit. Boyd's testimony that he informed Gemma on July 12 that Travelers was insisting on the use of its own policy, while perhaps interesting news, is legally irrelevant because Gemma reiterated in his binder outline on July 19, 2001 that he would be bound by the manuscript form submitted, i.e., the WilProp form. Therefore, at least for purposes of the binder, Gemma plainly rejected any proposal to switch from the WilProp form to the Travelers form. Boyd's similar claim concerning his post-September 11th conversation with Gemma is also irrelevant because any such conversation could not alter the terms of the binder that was in force on September 11th. That Gemma failed to respond to Boyd's October 2001 fax is irrelevant for the same reason. Moreover, it is a well settled rule that assent cannot be read into a party's silence in response to another party's assertion unless silence would have a tendency to mislead. See, e.g., Albrecht Chem. Co. v. Anderson Trading Corp., 298 N.Y. 437, 84 N.E.2d 625, 626 (1949); see also Hellenic Lines Ltd. v. Gulf Oil Corp., 340 F.2d 398, 401 (2d Cir.1965). By the time Boyd sent his fax in October 2001, the WTC buildings had long since collapsed and the coverage was what it was. Boyd obviously could not have been misled prior to September 11th by Gemma's silence in October.

The Silverstein Parties urge that the Hartford binder's language, "And as incorporated into the manuscript form, in conjunction with the contract between the Port Authority of New York and New Jersey as ultimately agreed," creates a disputed issue of fact that defeats summary judgment because it reasonably can be interpreted as referring to, and hence incorporating by reference, the as-yet-to-be-agreed Travelers form. The district court construed this language, which appeared in the Hartford binder under the heading "Property and Time Element Covered," as "indicat[ing] only that the parties were agreeing that the property to be insured would include all the property covered in the Silverstein Parties' contract with the Port Authority as that contract might be amended in their negotiations." Hartford Dec., 222 F.Supp.2d at 392.

We agree with the district court's interpretation, particularly given that the binder language, which was drafted by Willis, appears to have been culled from the section in the Willis Underwriting Submission, also entitled "Property and Time Element Covered," that described in detail the specific types of property and time-related interests for which coverage was sought, and that concluded with the statement "And as incorporated into the manuscript form, in conjunction with the contract between the Port Authority of New York and New Jersey as attached." The only difference was the change in the binder from "as attached" to "as ultimately agreed." This difference does nothing to undermine the district court's interpretation because the "as ultimately agreed" language can reasonably be read to refer to the scope of property subject to coverage (which would seem to be the only reason to refer to the lease), and not to any final form of insurance policy. From all these references to the ultimate agreement between the Port Authority and Silverstein Properties, the most that can be gleaned is that the precise parameters of the property covered by the insurance would have to await the finalization of that contract. No factfinder could reasonably find that these references related to following the form of Travelers as the lead insurer.

Indeed, the district court's interpretation of the language is echoed in Section VIII of the Underwriting Submission, which states,

Policy Form and Contract between Silverstein and the [Port Authority] are attached. DRAFT WilProp for Real Estate Risks is attached. We anticipate that this form will ultimately require amendment to comply with the Contract between Silverstein Properties, Inc. and the [Port Authority]. In the meantime, we provide this document as a starting point.

Further support for the district court's interpretation is found in the June 14, 2001 e-mail sent by Boyd to Royal Indemnity to solicit its participation in the WTC insurance program. In the e-mail, Boyd explained, "We have included [the WilProp form] as a guideline form [in the attached Underwriting Submission], although ultimate form must meet property definitions as contained in the contract with PA and Silverstein" (emphasis added).

Finally, the Silverstein Parties make much of the fact that at deposition, Gemma, when asked to review the original language contained in the Underwriting Submission, stated that he understood the phrase "manuscript form . . . as attached" to refer to the WilProp form that had been attached to the Underwriting Submission. But Gemma's statement does not equate to an acknowledgment that the revised phrase contained in the Hartford binder referred to the as-yet-to-be-agreed-upon Travelers form. And even interpreting the language in the light most favorable to the Silverstein Properties for purposes of summary judgment, this language does not create a genuine dispute because, as we have explained, even if this language meant that Hartford agreed to "follow form," and specifically, the Travelers form, "as ultimately agreed," it simply does not follow that Hartford thereby incorporated into its binder the terms of the as-yet-to-be-agreed Travelers form.

In short, we find nothing in the Hartford binder language — including the change of the Underwriting Submission language "as attached" to "as ultimately agreed" — to support the Silverstein Parties' claim that the parties intended to alter Hartford's repeated reservation of its right to use the WilProp form as amended by Hartford as the basis upon which its binder was issued. Indeed, the opposite is apparent in the July 19, 2001 e-mail Boyd sent to Gemma together with the draft binder, in which he stated, "We are working diligently with primary carriers to refine policy form. Attached is binder per specs." The first sentence in this e-mail suggests that no policy form (or "lead insurer" for that matter) had been agreed to by anyone as of July 19, 2001. The second sentence supports the view that the binder Boyd sent to Gemma did not alter any of the terms contained in Gemma's quote, including Gemma's insistence that the binder issue on the basis of the specifications and the WilProp form submitted with the Underwriting Submission together with Gemma's amendments.

Accordingly, we agree with the district court's conclusion that there can be no genuine dispute that Hartford bound coverage on the basis of the WilProp form.

On June 14, 2001, Boyd sent an e-mail to Mike Koenig of Royal containing the Underwriting Submission and the WilProp form as attachments. Koenig forwarded the Underwriting Submission to another Royal underwriter, Larry Stapp, who advised Koenig by e-mail on July 9, 2001, that

[b]ased on our conversation of today and review go ahead and offer [Willis] maximum of $100MM participation excess of $500MM at the $100K pricing we talked about. Make the participation contingent on receiving and reviewing the primary policy wording. You can use the Willis Form Review I sent you when you get the policy and just pick out the key items. Being excess of $500MM we will probably not want many form limitations.

The "Willis Form Review" referred to in the e-mail was a detailed review of a Willis policy form that pre-dated the WilProp form and did not include a definition of occurrence.11

Koenig submitted an underwriting proposal on behalf of Royal to Willis on July 9, 2001. In an accompanying fax, he stated:

Attached please find our authorization for the above risk. As discussed, Royal Indemnity Co. can offer $100MM (20%) part of $500MM xs $500MM at a layer price of $500,000 net. I have included some form changes that we would be looking for; however, this authorization would be subject to review and acceptance of the finalized manuscript form.

Under the heading "Policy Form," the Royal authorization specified, "Willis manuscript policy form as submitted except for the changes noted in the addendum to this quote. Final policy form wording is to be determined subject to review and acceptance of the final primary policy form wording." (The only Willis manuscript policy form that was submitted was the WilProp form.) Similarly, under the heading "Covered Perils," the authorization qualified the perils covered with the condition, "as per the Willis manuscript policy form with the changes described below. Subject to review and acceptance of the primary manuscript policy form," and, under the heading "Additional Conditions," stated, "This authorization is subject to review and acceptance of the finalized form being used by the primary insurers." In an addendum to the authorization, Royal specified certain revisions to the "manuscript form," which clearly pertained to the WilProp form. One such revision required the deletion of a paragraph in a section of the WilProp form entitled "Participation." The deleted paragraph provided a space for the designation of a lead underwriter and would have required Royal to "abide by and accept decisions of the Lead Underwriter with respect to underwriting, policy administration, and claims settlement."

While the language of the authorization demonstrates that Royal contemplated that the wording of the final policy form was tentative, it is clear from the evidence that prior to issuing its authorization, Royal anticipated that the final policy form would be based on the WilProp form. An internal Royal memorandum analyzing the WTC program that was written before Royal issued its authorization states:

Willis property form is very broad and would need substantial revisions. However based on our high attachment point, we would have little opportunity to dictate form changes. We should insist on a Y2K exclusion and delete all computer virus coverage. We should also try to limit the ingress/egress and civil authority coverages. Underlying deductibles will be [ ] $500,000 or $1,000,000 our attachment point is truly exposed only to catastrophic losses.

Elsewhere in the memorandum, it is stated that the "coverage form" will be "Willis manuscript."

On July 12, 2001, Koenig informed Boyd that Royal had altered some of the financial terms of its authorization. In an e-mail confirming the change, Koenig stated that "[a]ll other terms and conditions would remain as per [Royal's] original authorization." On July 17, 2001, Boyd e-mailed Koenig, telling him that Willis had been unable to assign to Royal the full amount of coverage that had been authorized. Koenig responded on July 19, providing Boyd with a policy number and requesting that Boyd "forward a copy of the finalized version of the manuscript form at [his] earliest convenience." On July 20, Boyd sent Koenig a binder containing largely the same language as the binder he had sent to Hartford's Gemma. The cover note stated, "Attached is a copy of the binder. We are working diligently to refine policy with primary carriers." Boyd stated in his deposition that sometime "during the middle weeks of July" he had a conversation with Koenig in which he informed Koenig "that the program would be based upon the Travelers form." Koenig memorialized that conversation in a handwritten note dated July 20 as follows:

Per discussion with Tim Boyd of Willis, terms and conditions of policy are likely to change, becoming more restrictive as form continues to be negotiated with the primary carriers. End result will most likely be a modified version of the Travelers policy form. I told Tim that we would bind subject to the policy form changes and coverage terms per our authorization (with the exception of the revised layer). Tim agreed and told me that this binder was a formality, and it will be revised in our favor once the primary policy form is finished.

On that same day, Koenig faxed Boyd the signed Royal binder after making several changes to the form he had received from Boyd. The fax cover sheet explained that Koenig "made some corrections to the binder in order to make it in accordance with the terms [Royal] authorized." Among the handwritten changes to the binder, Koenig added at the end of the binder the qualification, "Bound as amended and per our authorization." In addition, next to the statement in the "Property and Time Element" section, "And as incorporated into the manuscript form, in conjunction with the contract between the [Port Authority] as ultimately agreed," Koenig wrote, "Subject to Form revisions as described in our authorization." Willis did not qualify its acceptance of Royal's binder. According to Boyd's deposition, he told Koenig at some point in August, after Royal issued its binder, that he was "working very diligently to try and finalize the Travelers form." Koenig asked for a copy of the form but expressed no objection to its use in either the July or the later conversation.

We reject as frivolous the Silverstein Parties' arguments that Royal's authorization and binder contain ambiguities that create material issues of fact precluding summary judgment. As the district court noted in its decision granting summary judgment to Royal, "[i]t is hard to imagine a case in which it could be more certain that an insurer's binder was based on the WilProp form than that of [Royal]." Hartford Dec., 222 F.Supp.2d at 395. The Silverstein Parties nevertheless contend that the fact that Koenig did not eliminate the words "as ultimately agreed" from the phrase referencing the manuscript form in the "Property and Time Element" section and, instead, only added the phrase "Subject to form revisions as described in our authorization" demonstrates that Royal understood the phrase to refer to the final policy form that would be negotiated between Willis and the lead underwriter rather than to the agreement between the Port Authority and Silverstein Properties. Even if the Silverstein Parties' strained interpretation is accurate, the phrase simply sets forth the conditions under which Royal agreed to be bound by the final policy form when it issued in the future; it does not support a finding that the terms upon which Royal issued its presently enforceable binder were anything other than the WilProp form as amended by Royal in its authorization.

The Silverstein Parties also claim that even though the binder expressly binds "as previously authorized," the prior authorization itself is ambiguous because it stated, "Final policy form wording is to be determined subject to review and acceptance of the final primary policy form wording," and "[t]he insurance provided by Royal & SunAlliance will not be broader than the terms and conditions provided by any other participating insurer." This argument too, misses the mark, because the assertedly ambiguous statements refer to Royal's reservation of its right to review the wording of the final policy form before being bound by it, and therefore has no bearing on the terms that were intended to be incorporated into the binder that would govern Royal's coverage until that time. The Silverstein Parties' reliance on the phrase "subject to review and acceptance of the final primary policy form wording" once again ignores the basic tenet that the binder and the policy to be issued are two separate contracts of insurance, containing two separate sets of terms. See Springer, 710 N.Y.S.2d 298, 731 N.E.2d at 1108 (noting that binder and final policy are "two distinct agreements"); Rosenblatt, 594 N.Y.S.2d at 459 (noting that "the dispositive issue here concerns the risks covered under the binder agreement . . ., and this is unaffected by any changes regarding insurance policies issued subsequent to the loss").

The Silverstein Parties argue that there is another ambiguity, this one in the binder's section on exclusions, in which Koenig wrote that a Y2K exclusion was to be included, but otherwise left unchanged the binder's statement that exclusions were "Per Policy Form as to be advised." At most, however, this phrase creates an ambiguity only with respect to the binder's terms concerning exclusions. Because no other part of the binder contains similar language, the applicability of the amended WilProp form would remain intact, including its definition of occurrence.

Finally, the Silverstein Parties argue that Royal's conduct after September 11th demonstrates that it understood that the Travelers form applied to Royal's coverage. Specifically, the Silverstein Parties point out that on September 20, 2001, RSUI, the other Royal division that issued coverage in the WTC program, issued a Travelers form with an RSUI declarations page attached as its primary policy. According to Boyd's deposition, sometime in the fall, after RSUI issued its policy, he asked Koenig if Royal "would follow the same procedure" as RSUI and issue an excess policy based on the Travelers form, and Koenig indicated to him that what he would "like to do is sign and return the [Travelers] form and if you have to make some changes to it we do it by endorsement." Although this evidence may demonstrate equivocation on the part of Royal as to whether it should issue a final policy based on the Travelers form, it sheds no light on the only question before us, which is whether the July 20 binder issued by Royal was based on the WilProp or Travelers form. Moreover, other actions by Royal occurring after September 11th only confirm the district court's conclusion that the Royal binder was issued on the basis of the WilProp form. For example, on October 3, 2001, Koenig sent Brian Doyle, a Royal property executive, an e-mail attaching a copy of his "initial authorization, the signed binder and the Willis manuscript form that our quote was based on." Koenig added, "[o]ur quote was based on the Willis form and was subject to Best Terms. I also stated that our authorization would be subject to review and acceptance of the final primary policy wording." On October 15, 2001, Boyd e-mailed to Koenig an excess policy form drafted on the basis of the Travelers primary policy. On October 19, Koenig forwarded that e-mail to his superiors, stating:

I received this email from Willis regarding the excess form. The excess form still has not been finalized. I did express to the broker my discomfort of having the insured involved in drafting the form after the loss (and I reminded him that we quoted and bound coverage based on the Willis form). I have not issued anything yet or responded in writing to this email, and per my discussion with Bob Medeiros today, I wanted to get your input and guidance before I issued the policy.

On October 26, 2001, Doyle responded to Koenig's e-mail, stating, "Assume that we won't be signing on to this form since our authorization was based around the Willis manuscript policy and we were only presented with the Traveler's form post 9/11. I also see here that the policy issuance date is listed as 9/14."

Viewing this evidence together with the language of the Royal authorization and binder and the parties' pre-binder negotiations, we agree with the district court that there can be no reasonable dispute that the Royal binder was issued on the basis of the WilProp form.

On July 3, 2001, Harry Tucker of Stewart Smith, Willis's wholesale brokerage affiliate, e-mailed Carol Springett-King of St. Paul an abbreviated version of the Underwriting Submission. Although the submission referred to the "manuscript form," it did not annex or reference by name the WilProp or Travelers forms. Tucker, however, provided Springett-King with a copy of the WilProp form on July 9, 2001. The form was accompanied by a message stating that the policy was the "World Trade Center draft policy wording." At his deposition, Tucker could not identify any writing notifying Springett-King prior to September 11th that the Travelers form or any form other than the WilProp form was to be used for the final policy.

Springett-King testified at deposition that she did not do a form review of the WilProp form

[b]ecause it wasn't the final form. . . . Meaning that they were working on the language [in the WilProp form], and it had not been finalized. And they needed to bind coverage fairly quickly, so the understanding was that we would negotiate the form language, and we would get a copy of the final approved form, and we would be able to review it then.

Springett-King also testified that it was her understanding that she would be following Willis's broker's manuscript form as the primary form, but that she would be allowed to review it and make changes by endorsement when she received a final form.

On July 11, 2001, Springett-King called Tucker and quoted a rate of $1500 per million for $30 million of coverage in the layer excess of $250 million. On July 18, Tucker sent Springett-King an e-mail memorializing her quote and asking Springett-King to "confirm coverage bound with an assigned policy number by return e-mail." Springett-King sent the policy number and acknowledged at deposition that by doing so she bound coverage. In her e-mail to Tucker she stated that she would "send a formal binder shortly," but she never did so. On July 23, 2001, Stewart Smith sent Springett-King a Confirmation of Insurance, which, under the heading "Policy Form," stated "Manuscript Form to be agreed." In the accompanying cover letter, Michele Smith of Stewart Smith asked Springett-King to review the form carefully and to advise her if the confirmation did not accord with her records. Smith added, "We look forward to receiving the policy in due course." Tucker testified that he drafted the Confirmation of Insurance, and that his reference to "Manuscript Form to be agreed" meant that negotiations were ongoing with St. Paul as well as with all the other insurers concerning what the final form would be. He further testified that, as of July 23, he did not know what form or whose form that would be.

St. Paul's situation differs from Hartford's and Royal's in two significant respects. First, Springett-King's binder, which was merely a policy number furnished after receiving a bare-bones request from Tucker for coverage, did not indicate that she was binding based on the WilProp form. Second, as noted by the district court, "[t]here is no evidence that St. Paul was informed of Travelers' participation in the World Trade Center insurance program at any time prior to September 11th." Hartford Dec., 222 F.Supp.2d at 396.12

In granting summary judgment in favor of St. Paul, the district court held that, although it might be inclined to deny summary judgment on the ground that Springett-King had not even read the WilProp form before binding coverage, it was not "empowered to impose its own conception of what the parties should or might have undertaken." Hartford Dec., 222 F.Supp.2d at 397 (internal quotation marks omitted). Recognizing the well settled principle that a party is bound to a contract it signs even if it has not read it, see, e.g., Caloric Stove Corp. v. Chemical Bank & Trust Co., 205 F.2d 492, 495 (2d Cir. 1953) (L.Hand, J.) (noting that "the law of New York in this regard is the same as the general law of contracts: i.e., if a party to a written contract signs it, he is bound by its terms, whatever these may be," even if he has not read it) (citing Pimpinello v. Swift & Co., 253 N.Y. 159, 170 N.E. 530, 531 (N.Y.1930)); see also Pimpinello, 170 N.E. at 531 ("If the signer could read the instrument, not to have read it was gross negligence; if he could not read it, not to procure it to be read was equally negligent; in either case the writing binds him."), the district court found that St. Paul bound coverage on the basis of the WilProp form because it was a part of the WTC program solicitation, the terms of which Springett-King accepted in issuing the St. Paul binder. Hartford Dec., 222 F.Supp.2d at 397.

The district court also rejected the Silverstein Parties' argument that the Confirmation of Insurance's statement, "Manuscript Form to be agreed," evidenced Springett-King's agreement to follow form, reasoning that an agreement to follow form without knowing which form or who would be drafting it would be an unenforceable agreement to agree. Id. at 398. While we are inclined to agree with the Silverstein Parties that an insurer's agreement to follow form without knowing which form would ultimately be adopted does not render the insurance binder it issues an unenforceable "agreement to agree," the issue here is not whether St. Paul would have been bound to follow the Travelers final policy form when it issued. Rather, it is which policy St. Paul understood to be the policy form upon which it based its binder. Here the testimony of both Tucker and Springett-King makes plain that Springett-King had no reason to think that the policy form would be anything other than a modified version of the WilProp form. Accordingly, we agree with the district court's sound reasoning that because the only policy form presented to Springett-King was the WilProp form and because she was not informed about Travelers' participation in the WTC program, there can be no genuine dispute that the binder she issued on behalf of St. Paul incorporated the terms of the WilProp form.

Our conclusion that each of the three appellees in the Rule 54(b) appeal bound coverage on the basis of the WilProp form leaves only the Silverstein Parties' claim that there are issues of fact as to whether there were one or two occurrences on September 11th under the WilProp form's definition. As noted earlier, the WilProp form contains the following definition:

"Occurrence" shall mean all losses or damages that are attributable directly or indirectly to one cause or to one series of similar causes. All such losses will be added together and the total amount of such losses will be treated as one occurrence irrespective of the period of time or area over which such losses occur.

Hartford Dec., 222 F.Supp.2d at 398. Although the Silverstein Parties attempt to argue that this definition is ambiguous, we agree with the district court that no finder of fact could reasonably fail to find that the intentional crashes into the WTC of two hijacked airplanes sixteen minutes apart as a result of a single, coordinated plan of attack was, at the least, a "series of similar causes." Accordingly, we agree with the district court that under the WilProp definition, the events of September 11th constitute a single occurrence as a matter of law.

The Silverstein Parties' appeal from the denial of their motion for summary judgment against Travelers raises a different set of issues from those just discussed. This motion was based chiefly on the argument that where an insurance policy uses the term "occurrence" without defining the term, then, as a matter of law, the term's meaning is not ambiguous and must be decided by reference to well established New York legal precedent. The Silverstein Parties further argue that under the definition of "occurrence" established by New York law, the events of September 11th constituted two occurrences as a matter of law.

As was the case with the Rule 54(b) appeal, we review the district court's denial of the Silverstein Parties' motion for summary judgment against Travelers de novo. See Gibbs-Alfano, 281 F.3d at 18 ("We review the . . . denial of summary judgment de novo.") (internal quotation marks omitted). However, whereas in the Rule 54(b) appeal, we construed the evidence in the light most favorable to the Silverstein Parties, in this § 1292(b) appeal, we must construe the evidence in the light most favorable to non-movant Travelers. See Anderson, 477 U.S. at 255, 106 S. Ct. 2505; Maguire, 147 F.3d at 235. Before turning to the Silverstein Parties' arguments, we briefly review the history of the Travelers binder and its treatment by the district court.

On June 7, 2001, Boyd sent an e-mail to James B. Coyle, III, an underwriter at Travelers, that included both the Underwriting Submission for the WTC and the WilProp form. In an affidavit submitted in opposition to the Silverstein Parties' motion, Coyle attested that he responded to Boyd's offer by stating that if Travelers were to participate in the primary layer, it would insist on using its own form, but that Travelers would agree to the use of the WilProp form if Travelers participated only in the excess layers. At deposition, Boyd testified that he told Coyle that if Travelers would permit changes to its form to meet the Silverstein Properties' needs, Silverstein Properties would accept the Travelers form. On July 9, Coyle e-mailed to Boyd an authorization for coverage, to which he attached Travelers' specimen excess policy. Two days later, on July 11, Coyle e-mailed Boyd again setting forth a revised authorization for coverage and attaching Travelers' specimen primary policy. On July 17, Boyd e-mailed Coyle, stating that Willis was "pleased to bind [Travelers'] participation." Coyle's colleague Robert Malm confirmed Travelers' coverage by e-mail on July 18. On July 20, Boyd e-mailed to Travelers a binder similar to the one he had sent to Hartford and Royal, stating, "Attached is Binder to be used with Marketplace. We are aware of use of Travelers Form requirement and are reviewing same to make sure we and you agree where we can amend if need be. In the meantime, went out with specs as binder." Although Travelers never returned a signed copy of the binder, the parties agree that Travelers had bound coverage as of July 18.

After July 18, Boyd and Fenn Harvey, also of Willis, reviewed the WilProp and Travelers forms to identify the areas of difference between the two forms. Boyd then submitted a list of more than 76 differences to Travelers as a starting point for negotiating the terms of the final policy form. The list did not identify the presence or absence of a definition of "occurrence" as a point of difference between the two. During the next several weeks, the parties met to negotiate policy terms and exchanged drafts of the policy. By September 11th, however, the parties had not agreed to a final policy form. Following the destruction of the WTC, on either September 12 or 13, Boyd asked Coyle for a copy of the policy incorporating the changes that had been agreed to as of September 10, 2001. Boyd stated at deposition that he and Coyle agreed to freeze the draft as of that date. On September 14, 2001, Coyle issued the Travelers Policy for final review and approval by Silverstein Properties.

On June 3, 2002, the district court denied the Silverstein Parties' motion for summary judgment against Travelers. See SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props. LLC, 2002 WL 1163577 (S.D.N.Y. June 3, 2002) ("Travelers Dec."). In construing Travelers' binder, which incorporated by reference the Travelers form, the district court found

that the dispositive issue on this motion is whether the term "occurrence" has such a clear and unambiguous meaning that the trier of fact should be barred from considering the available extrinsic evidence concerning the meaning that the parties gave to that term when they were negotiating the insurance coverage for the World Trade Center.

Id. at *3. The district court then cited to Curry Road Ltd. v. K Mart Corp., 893 F.2d 509 (2d Cir.1990), for its enunciation of the standard for determining whether a contract term is ambiguous:

A term is ambiguous when it is capable of more than one meaning when viewed objectively by a reasonably intelligent person who has examined the context of the entire integrated agreement and who is cognizant of the customs, practices, usages and terminology as generally understood in the particular trade or business.

Travelers Dec., 2002 WL 1163577, at *4 (quoting Curry Road, 893 F.2d at 511) (internal quotation marks omitted). Applying the Curry Road test, the district court found that the meaning of "occurrence" to be incorporated into the Travelers binder was ambiguous given the history of litigation over the term, the fact that the term has been variously defined in different insurance policies, and the varying treatment of the term in different cases. Id. at *5-*6. The district court rejected the Silverstein Parties' argument that there is a single generally accepted meaning of "occurrence" under New York law. Id. The district court also found that the fact that Willis had circulated a policy that defined "occurrence" lent further support for a finding that the term as used in the Travelers binder was ambiguous. Id. at *5.

Accordingly, the court held that summary judgment had to be denied and that what the parties intended to be the meaning of "occurrence" would have to be determined by reference to extrinsic evidence, including

the specific definition of the term occurrence circulated by the insurance agent for the [insureds], testimony and documents relating to the negotiations prior to September 11th and the overall structure of the insurance program from the World Trade Center, and testimony and documentary evidence concerning statements made after September 11th by those who had been involved in negotiating the insurance contracts, in which they expressed their views on the question of whether there had been one or two occurrences.

Id. at *6.

On October 22, 2002, the district court amended the Opinion and Order to include a certification pursuant to 28 U.S.C. § 1292(b), thereby permitting the Silverstein Parties to seek leave to immediately appeal the denial of the summary judgment motion against Travelers together with the Rule 54(b) grants of summary judgment in favor of Hartford, St. Paul, and Royal.

As an initial matter, the parties dispute which contract governs the present dispute. The Silverstein Parties argue that the final Travelers policy issued on September 14th governs Travelers' obligations for the events of September 11th. They premise this claim on the fact that the Travelers policy that issued on September 14 specified an effective period of July 19, 2001 to July 18, 2002, and on the legal contention that under New York law, the terms of the insurance policy that is ultimately issued govern coverage during the binder period.

However, the case that the Silverstein Parties assert establishes this rule, Employers Commercial Union, 45 N.Y.2d 608, 412 N.Y.S.2d 121, 384 N.E.2d 668, is inapposite. There, the court sought to determine whether there was insurance coverage at the time a loss occurred, which was after a binder had been given but before the final policy was issued. In response to plaintiff's argument that "its ... coverage was not in place" at the time of the loss, the court stated, "[W]e reject any notion that the [insurer's] policy was not in force" at the time of loss. Id. at 669-70. The Silverstein Parties seize on this language as support for the proposition that the later-issued policy was the instrument governing the insurer's obligations. The language is taken out of context, however. As the rest of the opinion makes plain, the court's use of the word "policy" here is a reference to "insurance coverage" generally, and not to the later-issued formal policy. See id.; see also id. at 671 (stating in consecutive sentences, "[the Insurer's] insurance was in effect" and "[the Insurer's] policy was in effect"). The court's conclusion that insurance coverage was in place at the time of the loss rested in part on the fact that the final policy, although issued after the loss had occurred, specified coverage during the period in question. Id. at 669-70. But the relevance of this fact to the court's reasoning was not that the final policy terms therefore governed coverage, but simply that they demonstrated that the insurer had done "nothing to repudiate its coverage" after the loss occurred. Indeed, the court rejected the insurer's claim that the binder had not created an enforceable contract, observing that

[d]aily, important affairs and rights in our society are made to depend upon [binders]. It is a common and necessary practice in the world of insurance, where speed often is of the essence, for the agent to use this quick and informal device to record the giving of protection pending the execution and delivery of a more conventionally detailed policy of insurance. Courts, recognizing that the cryptic nature of binders is born of necessity and that many policy clauses are either stereotypes or mandated by public regulations, are not loath to infer that conditions and limitations usual to the contemplated coverage were intended to be part of the parties' contract during the binder period.

Id. at 670.

We do not read Employers Commercial Union to stand for the proposition that a final policy that is issued after a loss is the relevant contract governing the loss. In fact, as we have explained earlier in this opinion, New York law is quite to the contrary. See, e.g., Springer, 710 N.Y.S.2d 298, 731 N.E.2d at 1108 (noting that binder and final policy are "two distinct agreements"); Rosenblatt, 594 N.Y.S.2d at 459 (holding in case where loss occurred prior to the issuance of final policy, that "the dispositive issue ... concerns the risks covered under the binder agreement ..., and this is unaffected by any changes regarding insurance policies issued subsequent to the loss"); Del Bello, 585 N.Y.S.2d at 918-19 (holding that because binder identified plaintiff as an insured, it was covered for fire damage even though plaintiff was not so listed in policy that issued post-loss). Thus, as we held in connection with the Rule 54(b) appeal, it is the Travelers binder, not the September 14 Travelers policy that applies to determine Travelers' obligations.

D. Is the Binder Ambiguous?

The Silverstein Parties argue that the meaning of "occurrence" as used in the Travelers insurance coverage is not ambiguous and, therefore, that resort to extrinsic evidence to construe it is both unnecessary and improper. Because nothing in the documents that constitute the Travelers binder defined "occurrence," we must decide whether the undefined term "occurrence" when used in a first-party property damage contract is ambiguous.

Applying New York law, we have held that [t]he cardinal principle for the construction and interpretation of insurance contracts — as with all contracts — is that the intentions of the parties should control. Unless otherwise indicated, words should be given the meanings ordinarily ascribed to them and absurd results should be avoided. As we have stated before, the meaning of particular language found in insurance policies should be examined "in light of the business purposes sought to be achieved by the parties and the plain meaning of the words chosen by them to effect those purposes."

Newmont Mines Ltd. v. Hanover Ins. Co., 784 F.2d 127, 135 (2d Cir.1986) (internal citations omitted).

Whether a contract is ambiguous is a question of law for a court to determine as a threshold matter. As noted by the district court, an ambiguity exists where a contract term "could suggest `more than one meaning when viewed objectively by a reasonably intelligent person who has examined the context of the entire integrated agreement and who is cognizant of the customs, practices, usages and terminology as generally understood in the particular trade or business.'" Morgan Stanley Group Inc. v. New England Ins. Co., 225 F.3d 270, 275 (2d Cir.2000) (internal quotation marks omitted). Moreover, "a contract may be ambiguous when applied to one set of facts but not another. Therefore, ambiguity is detected claim by claim." Id. at 278.

Once a court finds that a contract is ambiguous, it may look to extrinsic evidence to determine the parties' intended meaning. See id. at 275-76. If factfinding is necessary to determine the parties' intent, however, the matter must be submitted to the finder of fact. See id. at 279.

The first argument made by the Silverstein Parties invokes the doctrine that "whether an ambiguity exists must be ascertained from the face of an agreement without regard to extrinsic evidence," Reiss v. Fin. Performance Corp., 97 N.Y.2d 195, 738 N.Y.S.2d 658, 764 N.E.2d 958, 961 (2001) (internal quotation marks); see also Md. Cas. Co. v. W.R. Grace & Co., 23 F.3d 617, 625 (2d Cir.1993) ("Interpretation of unambiguous contract language does not bring extrinsic evidence into play."). This argument fails because it is based on the faulty premise that the September 14 Travelers policy rather than the Travelers binder governs the parties' obligations. While New York law is clear that extrinsic evidence may not be used to contradict clearly unambiguous language contained in an insurance binder, see Am. Sur. Co. v. Patriotic Assurance Co., 242 N.Y. 54, 150 N.E. 599, 601 (1926) (holding that it was error to admit extrinsic evidence to contradict unambiguous description of location in insurance binder), it is just as well settled in New York that extrinsic evidence is admissible to determine the parties' intentions with respect to the incomplete and unintegrated terms of a binder. See, e.g., Underwood v. Greenwich Ins. Co., 161 N.Y. 413, 55 N.E. 936, 938-39 (1900) (holding that because binder was not "in and of itself, ... such a complete and perfect instrument that it embodie[d] all the mutual stipulations of the parties ... [it] was open to explanation by parol proof as to the intention of the parties, and the established custom of the business"); see also Thomas v. Scutt, 127 N.Y. 133, 27 N.E. 961, 962-63 (1891) (noting that exception to general rule that parol evidence is inadmissible to contradict a written contract is where "the written instrument, [though] existing and valid,... [is] incomplete, either obviously, or at least possibly, and ... parol evidence [is admitted], not to contradict or vary, but to complete, the entire agreement, of which the writing is only a part"). Indeed, the Silverstein Parties have relied on this exception to the parol evidence rule in the context of their disputes with other insurers. See, e.g., SR Int'l Bus. Ins. Co. v. World Trade Ctr. Props. LLC, 2003 WL 289600, at *1 (S.D.N.Y. Feb 11, 2003) (agreeing with the Silverstein Parties that question of whether defendant-insurer Zurich's binder "provid[ed] coverage on a `per occurrence' basis [could] not be resolved without resort to extrinsic evidence"); Appellants' Rule 54(b) Br. at 2-3 (arguing that due to the fact that "binders are a species of temporary, unintegrated contracts," a court must resort to extrinsic evidence of industry usage and custom to discern the expectations of the parties, particularly where, as here, the binders are "not remotely unambiguous"). In fact, in their brief appealing the district court's grants of summary judgment in the Rule 54(b) appeal, the Silverstein Parties take the clear position that

the parol evidence rule does not come into play at all when dealing with un integrated contracts, such as appellees' binders. Rather, extrinsic evidence is "admissible to supply the terms that the parties intended to incorporate into their agreement." Saxon Capital v. Wilvin Assocs., 195 A.D.2d 429, 600 N.Y.S.2d 708, 709 (1st Dep't 1993); see also Bourne v. Walt Disney, 68 F.3d 621, 627-28 (2d Cir.1995). And, in such circumstances, "summary judgment does not lie" unless the extrinsic evidence itself is so one-sided as to negate the existence of a triable issue of fact. Lowell v. Twin Disc, 527 F.2d 767, 770 (2d Cir.1975); see also Saxon Capital, 600 N.Y.S.2d at 709.

Appellants' Rule 54(b) Br. at 77. We agree with this statement of the law, which fully applies to the Travelers binder.

The Silverstein Parties' next contention, that the undefined term "occurrence" is not ambiguous because it is typical for insurance policies not to define "occurrence" and, further, that the WilProp definition is "atypically broad," is undercut by the policy forms of the two other WTC insurers who provided their own forms for coverage, each of which defined occurrence. IRI issued a binder expressly incorporating its own policy form, which defines "loss arising out of one Occurrence" as "the sum total of all loss or damage insured against arising out of or caused by one event." Allianz, in the only final policy to issue before September 11, 2001, defined occurrence in language similar to the WilProp definition: "any one loss, disaster or casualty, or series of losses, disasters or casualties arising out of one event."

In addition, in order to demonstrate the ambiguity of the undefined term "occurrence," Travelers has proffered evidence of industry custom and usage concerning the meaning of occurrence that differs from the definition asserted by the Silverstein Parties. For example, a Willis forms specialist testified that she did not believe that the WilProp form definition of an occurrence as, inter alia, losses attributable to "one series of similar causes" deviated from the commonly understood meaning of "occurrence." Similarly, Daniel McCrudden, an underwriter at Travelers, testified that "it's recognized that multiple causes of loss can be involved in a single occurrence, and it's recognized that all loss arising out of an overriding cause or group of causes is considered a single occurrence. It's never been a question."

Although the Silverstein Parties argue that it was improper for the district court to consider such evidence of custom and usage in deciding whether the policy is ambiguous, we have specifically instructed courts to consider the "customs, practices, usages and terminology as generally understood in the particular trade or business" in identifying ambiguity within a contract. Int'l Multifoods Corp. v. Commercial Union Ins. Co., 309 F.3d 76, 83 (2d Cir.2002) (quoting Morgan Stanley Group, 225 F.3d at 275). And New York courts have long held that such evidence is admissible for purposes of construing an insurance binder. See, e.g., Underwood, 55 N.E. at 937 (holding that binder "was open to explanation by parol proof as to... the established custom of the business").

Finally, the Silverstein Parties assert that the mere fact that the word "occurrence" was not defined in the binder is not enough to render it ambiguous. They contend that in the absence of a definition in the binder, a court seeking to construe the meaning of "occurrence" must first turn to well established New York precedent. If there is a clear and uniform meaning of the term under the law, they argue, then a court must reject a claim of ambiguity and apply that definition. This argument fails because its underlying premise — that there is a uniform meaning of "occurrence" under New York law — is erroneous.

The Silverstein Parties maintain that under New York law, there is but one meaning of "occurrence," which is the direct, physical cause of a loss and not more remote causes. This definition is so accepted and well settled, they contend, that it must be implied into the Travelers binder as a matter of law. Applying this definition to the facts of this case, it follows that because the destruction of the WTC was the result of two physical impacts from two separate planes, there were two occurrences as a matter of law.

To support their argument, the Silverstein Parties rely on a string of authorities beginning with Arthur A. Johnson Corp. v. Indem. Ins. Co., 7 N.Y.2d 222, 196 N.Y.S.2d 678, 164 N.E.2d 704 (1959). In Arthur A. Johnson, the court considered whether there was one or two accidents within the meaning of a third-party liability insurance policy where two separate walls constructed by the same insured contractors in two adjacent buildings collapsed 50 minutes apart during the course of an unusually heavy rainfall and caused flooding within the buildings. The insurer argued that there was only one occurrence because all the damage was ultimately caused by the heavy rainfall. In rejecting this argument, the court started from the premise that in determining the number of accidents, it must consider "the `reasonable expectation and purpose of the ordinary business man when making an ordinary business contract.'" Id. at 706 (quoting Bird v. St. Paul Fire & Marine Ins. Co., 224 N.Y. 47, 120 N.E. 86, 87 (1918)). The court then held that "the term [accident] is to be used in its common sense of an event of an unfortunate character that takes place without one's foresight or expectation ... [t]hat is, an unexpected, unfortunate occurrence." Id. at 707 (internal quotation marks and emphasis omitted). Reviewing the facts before it, the court concluded that there had been two separate accidents on the rationale that the walls that collapsed belonged to separate buildings, there was no indication that the flooding in the first building would have caused the flooding in the second building in the absence of a second defective wall, and the two walls collapsed almost an hour apart. Id. at 708.

Because Arthur A. Johnson and nearly all of the other cases relied on by the Silverstein Parties to provide the definition of "occurrence" are third-party liability insurance cases, however, they involve different interests, both public and private, than first-party property insurance cases such as the instant case. See generally Great N. Ins. Co. v. Mount Vernon Fire Ins. Co., 92 N.Y.2d 682, 685 N.Y.S.2d 411, 708 N.E.2d 167, 170 (1999) ("[W]holly different interests are protected by first-party coverage and third-party coverage."); see also Port Auth. v. Affiliated FM Ins. Co., 311 F.3d 226, 233 (3d Cir. 2002) (finding third-party definitions of contract terms unhelpful in first-party context because of "[t]he fundamental differences between liability policies and first-party contracts"). For example, for third-party liability policies, there is no reason to look any further back in the chain of causation than to the insured's acts of negligence, because it is the insured's negligence that triggers liability. See, e.g., In re Prudential Lines Inc., 158 F.3d 65, 80-81 (2d Cir.1998) (holding, in third-party liability context, that "courts should look to the event for which the insured is held liable," regardless of whether it is the physical impact closest in time) (internal quotation marks and emphasis omitted); Stonewall Ins. Co. v. Asbestos Claims Mgmt. Corp., 73 F.3d 1178, 1213-14 (2d Cir.1995) (same), modified on other grounds, 85 F.3d 49 (2d Cir.1996). Thus, the approach taken by courts reviewing the number of occurrences in the context of third-party liability — such as the court's focus in Arthur A. Johnson on the separate wall collapses rather than the rain — makes sense for such policies because the insured is held liable only for its own negligence and not for the act of nature that may have been the initiating cause. See Arthur A. Johnson, 196 N.Y.S.2d 678, 164 N.E.2d at 708 ("Here the proximate cause cannot be said to be the heavy rainfall but separate negligent acts of preparing and constructing separate walls which, for all we know, may have been built at separate times by separate groups of workmen.").

In addition, construction of the term "occurrence" in a liability insurance context is influenced by the public policy concern of ensuring adequate compensation for injured third-parties who are not parties to the insurance contract, and, perforce, played no role in negotiating its terms. See Affiliated FM Ins. Co., 311 F.3d at 233. It is no surprise, therefore, that a salient characteristic of the third-party liability cases relied on by the Silverstein Parties is that each one involved multiple liability claims filed against the insured by multiple parties. See, e.g., In re Prudential Lines, 158 F.3d at 68 (liability on multiple asbestos claims); Stonewall Ins., 73 F.3d at 1187 (liability on thousands of asbestos claims); Travelers Cas. & Sur. Co. v. Certain Underwriters at Lloyd's of London, 96 N.Y.2d 583, 734 N.Y.S.2d 531, 760 N.E.2d 319, 322 (2001) (liability claims involving decades of commercial activities at numerous industrial and waste disposal sites); Hartford Accident & Indem. Co. v. Wesolowski, 33 N.Y.2d 169, 350 N.Y.S.2d 895, 305 N.E.2d 907, 908 (N.Y.1973) (car liability insurance where insured struck two different cars); Arthur A. Johnson, 196 N.Y.S.2d 678, 164 N.E.2d at 704 (liability for destruction of retaining walls of two buildings owned by different owners). And the "test" that the Silverstein Parties argues is universally applicable in all insurance contexts was described by the Arthur A. Johnson court as applying "in a given set of circumstances when the damage is to several persons." Arthur A. Johnson, 196 N.Y.S.2d 678, 164 N.E.2d at 706 (emphasis added). In such cases, of course, a finding of a separate occurrence as to each claimant ensures compensation for the injured third parties. See Affiliated FM Ins. Co., 311 F.3d at 233.

In a first-party property case, by contrast, the insured's negligence is not at issue; rather, the policy insures against external perils such as fires, floods, and intentional acts that cause damage to the insured's property, and against which a property interest holder can take adequate measures to protect his investment in advance of any loss. See Newmont Mines, 784 F.2d at 136 ("The goal of such a [first-party] policy, simply stated, is to provide financial protection against damage to property."). As a result of these differences, a court's construction of the undefined term "occurrence," or the synonymous term "accident," as intended by the parties for use in the third-party context is not necessarily applicable in the context of first-party property insurance.

We also find it noteworthy that while the Silverstein Parties assert that Arthur A. Johnson "provides the applicable legal test for determining number of occurrences," they do not try to apply that case's definition to the facts of this case. Instead, they assert that "the governing test under `well-established precedent' under New York law is to look to the immediate, efficient, physical cause of the loss, and not to some indirect or remote cause of causes." The Silverstein Parties' source for this rule appears to be some New York cases that have applied a similar rule to determine causation in the context of whether an exclusionary clause applies to a loss. See, e.g., Album Realty Corp. v. Am. Home Assurance Co., 80 N.Y.2d 1008, 592 N.Y.S.2d 657, 607 N.E.2d 804, 805 (1992) (applying "proximate, efficient and dominant cause" test to hold that covered risk of water damage, which was caused by frozen pipes, rather than excluded risk of freezing, was cause of loss); but see Tonkin v. Cal. Ins. Co., 294 N.Y. 326, 62 N.E.2d 215, 216-17 (1945) (applying proximate cause test to hold that covered risk of fire, rather than excluded risk of collision was cause of loss, where blinding smoke from fire caused driver to collide with another car). But no New York case of which we are aware has set forth such a test for purposes of determining the number of occurrences that comprise a loss.

The test actually enunciated by Johnson and its progeny in the third-party liability context is considerably more nuanced:

We have expressed the rule of these cases as follows: "In determining the number of occurrences for deductible purposes, New York inquires whether multiple claims result from `an event of an unfortunate character that takes place without one's foresight or expectation.'... [A]lthough a single `occurrence' may give rise to multiple claims, courts should look to the event for which the insured is held liable, not some point further back in the causal chain."

In re Prudential Lines, 158 F.3d at 80 (quoting Stonewall Ins., 73 F.3d at 1213). Even if we were to accept the Silverstein Parties' contention that this statement sets forth the uniform and well settled definition of "occurrence" under New York law, its application to the facts before us would not establish as a matter of law whether the events of September 11th were one or two occurrences. For example, what, precisely, is the "event" here for which insurance coverage is being sought? In the context of third-party liability insurance — the type of insurance at issue in Arthur A. Johnson — the "event" is the insured's negligence. According to the Silverstein Parties, under the insurance with Travelers, "the `event' that triggers coverage ... is `direct physical loss or damage' to property." This wording traces the language contained in the Travelers form upon which the Travelers binder was issued, which provides that

[t]he Company will pay for direct physical loss or damage to Covered Property at premises ..., caused by or resulting from a Covered Cause of Loss. Covered Cause of Loss means risks of direct physical loss unless the loss is excluded in Section D., Exclusions; limited in Section E., Limitations; or excluded or limited in the Supplemental Coverage Declarations or by endorsements.

A jury construing this language, however, could reasonably conclude that the "event" that triggers coverage is the Covered Cause of Loss, rather than the damage itself. It also could find that the cause of the destruction of the WTC was either the individual impacts caused by each plane or a single coordinated terrorist attack. The latter cause, Travelers argues in its opposition to summary judgment, was a Covered Cause of Loss expressly specified in the Underwriting Submission and, thus, intended by the parties to be covered by the policy. In fact, a jury could find that the words "direct physical loss or damage" does not refer to the "event" that triggers coverage at all, but rather sets forth the scope of the damage resulting from the "event" that the insurer will pay for, namely, direct physical damage as distinct from remote or incidental damage.

In any event, we are not called upon here to decide whether there was one occurrence or two in this case, only whether the district court properly concluded that because there is no well settled definition of the term "occurrence" under New York law, the Travelers binder was sufficiently ambiguous to preclude summary judgment and to permit the factfinder to consider extrinsic evidence of the parties' intent.

We think the case most directly on point is our decision in Newmont Mines, the only first-party property insurance case cited by the Silverstein Parties that addresses the meaning of "occurrence." 784 F.2d at 135-37. In that case, we were "not persuaded ... [by defendants] that the term `occurrence' has obtained any ... specialized or singular meaning in the context of property insurance," and we interpreted both Arthur A. Johnson and Wesolowski as "rejecting any single definition of occurrence." Id. at 136. Our conclusion, at bottom, was that "the meaning of `occurrence' must be interpreted in the context of the specific policy and facts of th[e] case." Id. at 136 n. 9.

In Newmont Mines, a heavy accumulation of snow caused two separate parts of a roof to collapse several days apart, requiring two independent repairs. Id. at 129-31. We held that there was sufficient evidence to support the jury's verdict that the two partial losses constituted two occurrences under the policies. Id. at 137. We also upheld the instruction given to the jury on the meaning of occurrence:

it is for you to decide whether or not the losses which are alleged to have occurred or the loss that's alleged to have occurred in this case was the result of a single, continuous event or incident, or whether or not it was the result of two separate incidents. If you find that the collapse of the two sections of the roof was a single, continuous event or incident, then the collapse constituted a single occurrence — and there would be only one loss. If, on the other hand, you find that the collapse of the two sections of the roof constituted separate events or incidents that were not causally related, then of course you would have two separate losses.

Id. at 134. We held that the instruction was proper on the rationale that given the goals of first-party property insurance, "the parties ... must have intended to provide coverage for property damage each time it occurred unexpectedly and without design, unless the damage occurring at one point in time was merely part of a single, continuous event that already had caused other damage." Id. at 136.

Notwithstanding the express statements to the contrary in our decision, see id. at 135-36 & n. 8. (rejecting any one definition of "occurrence" and stating that "the meaning of `occurrence' must be interpreted in the context of the specific policy and facts of th[e] case"), the Silverstein Parties contend that Newmont Mines sets forth "the rule of law" with respect to the meaning of "occurrence" in the context of first-party property insurance, and that this rule favors them. However, even if we were to agree with the Silverstein Parties that the approved jury instruction and our separate definition of "occurrence" are applicable to this case (a question we need not and do not reach), the one thing Newmont Mines makes certain is that the question of how many occurrences the events of September 11th constituted is a question properly left to the fact-finder. To be sure, a jury could find two occurrences in this case, as it did in Newmont Mines, or it could find that the terrorist attack, although manifested in two separate airplane crashes, was a single, continuous, planned event causing a continuum of damage that resulted in the total destruction of the WTC, and, thus, was a single occurrence. Instead of supporting the Silverstein Parties' argument that New York law mandates a finding of two occurrences under the Travelers binder as a matter of law, Newmont Mines confirms our belief that in a first-party property insurance case, the meaning of the undefined term "occurrence" is an open question as to which reasonable finders of fact could reach different conclusions.

Accordingly, we conclude that given the significant distinction between first-party and third-party insurance policies, the fact-specific nature of the inquiry, and the fact that it is the parties' intent that controls, the district court properly concluded that the meaning of "occurrence" in the Travelers binder is sufficiently ambiguous under New York law to preclude summary judgment and to warrant consideration by the fact finder of extrinsic evidence to determine the parties' intentions. We therefore affirm the denial of summary judgment against Travelers.

For the foregoing reasons, we affirm the judgments of the district court.