Folksamerica Reinsurance Company, as Successor in Interest to Christiania General Insurance Company of New York v. Clean Water of New York, Inc.

413 F.3d 307, 2005 A.M.C. 1747, 2005 U.S. App. LEXIS 13041, 2005 WL 1532314

• Court: Court of Appeals for the Second Circuit
• Decided: June 30, 2005
• Docket: Docket 03-9124
• Status: Published

Opinion

WESLEY, Circuit Judge.

Milton Rivera was injured cleaning the oil tank of a barge moored in New York Harbor. Soon thereafter, defendant-ap-pellee Clean Water of New York, Inc., which had subcontracted the work to Rivera’s employer, found itself involved in a negligence action in a New York State court. See Folksamerica Reinsurance Co. v. Clean Water of New York, Inc., 281 F.Supp.2d 530, 531 (E.D.N.Y.2003). Clean Water notified its insurer of the action, and the insurer’s- successor in interest, plaintiff-appellant Folksamerica Reinsurance Company, went to federal court seeking a declaration that it had no obligation to defend or indemnify Clean- Water. Id. at 531-32. The insurance contract at issue contained a Shiprepairers Legal Liability (SLL) policy coupled with a modified Comprehensive General Liability (CGL) policy (collectively, the Policy). Id. at 532. The district court (Block, J.) determined that the contract was neither wholly nor primarily maritime in nature and dismissed for lack of subject matter jurisdiction. Id. at 533-34.' Folksamerica appeals, and we vacate and remand because we find that the primary objective of the contract was to establish marine insurance.

Background

The parties, Folksamerica and Clean Water, are a reinsurance company and a ship tank cleaning business, respectively. Folksamerica is the successor in interest to Christiania General Insurance Corporation of New York. Id. at 531. On behalf of Christiania, Trident Marine Managers, Inc., a managing general agent of marine insurance, issued an insurance policy to Eklof Marine Corporation, the corporate parent of Clean Water. Folksamerica, Christiania, and Clean Water are all New York corporations.

Clean Water cleans the tanks of coastal and ocean-going vessels in New York Harbor and prepares vessels for the shipyard and for changes of cargo. Sometime prior to mid-August 1994, Clean Water contracted to clean the oil tanks of Barge S.T. 85. Id. The barge, an ocean-going vessel, was moored in the Kill Van Kull tidal strait, a navigable waterway within New York Harbor. See id. Union Maintenance Corporation supplied Clean Water with workers for the oil-tank cleaning, and on August 12, 1994, Mr. Rivera, a Union Maintenance employee, was injured on the job. See id. Rivera, perhaps overwhelmed by fumes, lost his grip on a ladder and tumbled backwards into an oil tank. Thereafter, Rivera brought suit against Clean Water in the Civil Court of New York, Kings County, principally alleging negligence. Id.

Clean Water sought defense and indemnification under the Policy. 1 Id. The Policy lists over a dozen named insureds, consisting of Eklof Marine and its subsidiaries. 2 All of the insureds are marine companies whose business operations relate entirely to either ship repair, marine oil transport, marine cargo transport, or, in the case of Clean Water, ship tank cleaning. The Policy has two sections defining coverage, “Section I, Comprehensive General Liability” and “Section II, Shiprepairers Legal Liability.” Id. at 532. Shared liability limits govern these sections; the Policy has one premium. Clean Water asserted its insurance claim under the CGL section of the Policy.

On June 19, 2002, Folksamerica filed a declaratory judgment action against Clean Water in the United States District Court for the Eastern District of New York. Folksamerica alleged admiralty jurisdiction and sought voidance of the Policy, rescission, and a declaration that it was not obligated to defend or indemnify Clean Water. Clean Water raised various defenses, a counterclaim seeking reformation of the Policy, and a challenge to the district court’s subject matter jurisdiction.

Both sides moved for summary judgment. In defending the court’s admiralty jurisdiction, Folksamerica contended, “Since the Policy was issued to maritime companies, clearly covers their marine interests[,] and the underlying accident was on board a vessel in navigable waters, it is a policy of ‘marine insurance.’ ” The admiralty question had implications beyond conferring federal jurisdiction. Folksam-erica pressed the court to employ the federal maritime doctrine of uberrimae fide, or utmost good faith, which “provides that the parties to a marine insurance contract are held to the highest degree of good faith, whereby the party seeking insurance is required to disclose all circumstances known to it which materially affect the risk.” Atl. Mut. Ins. Co. v. Balfour Maclaine Int’l Ltd. (In re Balfour MacLaine Int’l Ltd.), 85 F.3d 68, 80 (2d Cir.1996) (hereinafter Balfour) (internal quotation marks and brackets omitted). Folksamer-ica asserted that Clean Water failed to disclose that it performed tank cleaning on others’ vessels and that the Policy was therefore void.

Clean Water argued that Folksamerica could not invoke admiralty jurisdiction. It characterized the CGL section of the Policy as a standard “all risk policy” and contended that the maritime risks covered by the Policy were “merely incidental” at best. The district court agreed. Folksamerica, 281 F.Supp.2d at 531.

The court focused on the CGL section of the Policy and found it did not constitute “marine insurance.” Id. at 532-33. Relying on Benedict tin Admiralty, the court explained that, for a contract to have “ ‘maritime character,’ ” “ ‘there must be present a direct and proximate judicial link between the contract and the operation of the ship, its navigation or its management afloat.’ ” Id. at 532 (quoting BENEDICT on Admiralty § 182, at 11-7 (7th ed.1985)). The court characterized the CGL section as “a standard comprehensive general liability policy of the type used by many businesses to cover a variety of losses which may arise in .day-to-day general operations.” Id. at 533. Contrasting CGL insurance with “the three traditional forms of marine insurance — hull insurance, cargo insurance, and protection and indemnity insurance,” the court concluded that the CGL section “simply lacks the genuinely salty flavor necessary to constitute a maritime contract.” Id. (internal quotation marks omitted). The court declined to consider the SLL section. Id. at 533-34. It reasoned that, while admiralty jurisdiction exists “where the non-maritime elements are merely incidental in an otherwise maritime contract,” id. at 533 (citing Atl. Mut. Ins. Co. v. Balfour Maclaine Int’l Ltd., 968 F.2d 196, 199 (2d Cir.1992) (hereinafter Atlantic Mutual) (internal quotation marks omitted)), “coverage under Section I can hardly be considered incidental to the "Policy as a whole,” id. at 534. The court dismissed the compláint, and Folksamerica timely appealed.

Discussion

This appeal turns on what would seem a simple inquiry: Is the Policy a maritime contract giving rise to admiralty jurisdiction? See Jeffcott v. Aetna Ins. Co., 129 F.2d 582, 584 (2d Cir.1942). Unfortunately, there are few “clean lines between maritime and non-maritime contracts.” Norfolk S. Ry. Co. v. James N. Kirby, Pty Ltd., 543 U.S. —, 125 S.Ct. 385, 393, 160 L.Ed.2d 283 (2004). The boundaries of admiralty jurisdiction over contracts are conceptual rathep than spatial, id. (quoting Kossick v. United Fruit Co., 365 U.S. 731, 735, 81 S.Ct. 886, 6 L.Ed.2d 56 (1961)), and defined by the purpose of the jurisdictional grant — to protect maritime commerce, see Exxon Corp. v. Cent. Gulf Lines, Inc., 500 U.S. 603, 608, 111 S.Ct. 2071, 114 L.Ed.2d 649 (1991) (citing Ins. Co. v. Dunham, 78 U.S. (11 Wall.) 1, 24, 20 L.Ed. 90 (1870), and Sisson v. Ruby, 497 U.S. 358, 367, 110 S.Ct. 2892, 111 L.Ed.2d 292 (1990)). The Supreme Court has adopted a case-by-case approach in defining these boundaries, see Sisson, 497 U.S. at 372, 110 S.Ct. 2892 (Scalia, J., concurring), and has instructed that “[precedent and usage are helpful insofar as they exclude or include certain common types of contract,” Kossick, 365 U.S. at 735, 81 S.Ct. 886.

While “ ‘[t]he precise categorization of the contracts that warrant invocation of the federal courts’ admiralty jurisdiction has proven particularly elusive,’ ” 3 Ingersoll Milling Mach. Co. v. M/V Bodena, 829 F.2d 293, 301 (2d Cir.1987) (quoting CTI-Container Leasing Corp. v. Oceanic Operations Corp., 682 F.2d 377, 379 (2d Cir.1982)), the abiding instruction of the Supreme Court is that we should look to the contract’s “ ‘nature and character’ ” to see “whether it has ‘reference to maritime service or maritime transactions,’ ” Norfolk S. Ry. Co., 125 S.Ct. at 393 (quoting N. Pac. S.S. Co. v. Hall Bros. Marine Ry. & Shipbuilding Co., 249 U.S. 119, 125, 39 S.Ct. 221, 63 L.Ed. 510 (1919)); see Dunham, 78 U.S. (11 Wall.) at 26. Therefore, the contract’s subject matter must be our focal point. See, e.g., Exxon Corp., 500 U.S. at 611, 111 S.Ct. 2071; Dunham, 78 U.S. (11 Wall.) at 29.

A. A Threshold Inquiry

Several of our cases, however, require that, prior to inquiring into the subject matter of the contract, we first make a “threshold inquiry” into the subject matter of the dispute. Balfour, 85 F.3d at 74-75. “Before attempting to categorize contractual rights as maritime or non-maritime, a federal court must first consider whether an issue related to maritime interests has been raised.” Atlantic Mutual, 968 F.2d at 199; see Balfour, 85 F.3d at 74-75; cf. Thypin Steel Co. v. Asoma Corp., 215 F.3d 273, 278-79 (2d Cir.2000); Sirius Ins. Co. (UK) Ltd. v. Collins, 16 F.3d 34, 37 (2d Cir.1994). “Therefore, in examining whether admiralty jurisdiction encompasses a claim, a federal court must initially determine whether the subject matter of the dispute is so attenuated from the business of maritime commerce that it does not implicate the concerns underlying admiralty and maritime jurisdiction.” Atlantic Mutual, 968 F.2d at 200 (emphasis added).

This Court has twice encountered disputes with “connection[s] with maritime commerce ... simply too speculative and attenuated to justify admiralty and maritime jurisdiction.” Id.; see Balfour, 85 F.3d at 74-75. In both Atlantic Mutual and Balfour, this Court considered claims for coffee that was lost from a Mexican warehouse. See Atlantic Mutual, 968 F.2d at 200; Balfour, 85 F.3d at 74-75. The claims arose under an insurance policy “cover[ing] ... cargo during land transport and warehousing, as well as during ocean carriage..” Sirius Ins., 16 F.3d at 37; see Balfour, 85 F.3d at 74-75; Atlantic Mutual, 968 F.2d at 200. In both cases, the insurer filed declaratory judgment actions seeking determinations that the claimed losses had not been sufficiently substantiated. See Balfour, 85 F.3d at 72-74; Atlantic Mutual, 968 F.2d at 198. This Court invoked its “threshold inquiry” and found admiralty jurisdiction lacking. See Balfour, 85 F.3d at 74-76; Atlantic Mutual, 968 F.2d at 200. We concluded that the disputes were too attenuated from maritime commerce to sustain admiralty jurisdiction because the insurance claims involved coffee — a good without any “inherently maritime character,” Sirius Ins., 16 F.3d at 37—that was designated for land transportation only, never became marine cargo, and never entered maritime commerce. See Atlantic Mutual, 968 F.2d at 200; Balfour, 85 F.3d at 75; see also Sirius Ins., 16 F.3d at 37.

Here, the district court did not conduct this threshold analysis. In contrast to the above cases, the parties’ dispute concerns an insurance claim based on a ship-maintenance-related injury sustained by a ship oil-tank cleaner aboard an oceangoing vessel in navigable waters. The business of ship maintenance has long been recognized as maritime, 4 and the insurance claim arising out of a related on-board injury has more than a “speculative and attenuated” connection with maritime commerce, Atlantic Mutual, 968 F.2d at 200. Coffee lost from a Mexican warehouse clearly has little to do with “the business of maritime commerce.” Id. However, it is fair to say Milton Rivera became a casualty of the business of maritime commerce the moment he tumbled from the tank ladder aboard the Barge S.T. 85. 5 Accordingly, this case easily survives the “threshold inquiry.”

Moreover, we note some uncertainty as to the extent to which this Court’s “threshold inquiry” test survives the Supreme Court’s most recent admiralty decision, Norfolk Southern Railway Co. v. James N. Kirby Pty Ltd., which the Supreme Court introduced as “a maritime case about a train wreck.” 125 S.Ct. at 390. That decision involved a contract for the transportation of goods from Australia to Alabama. Id. at 391-92. The dispute concerned a railroad’s liability for machinery damaged during a train derailment. Id. at 390. The machinery had been transported by ship from Sydney, Australia, to Savannah, Georgia, and was en route from Savannah to Huntsville, Alabama, when the train derailed. See id. at 390-91. The Court focused entirely on the underlying contract and stated, “To ascertain whether a contract is a maritime one, we cannot look to whether a ship or other vessel was involved in the dispute, as we would in a putative maritime tort case.” Id. at 393. Focusing on the contract subject matter, the Court found admiralty jurisdiction. See id. at 395. Although Norfolk Southern Railway Co. is readily distinguishable from Atlantic Mutual and Balfour — the machinery in Norfolk Southern Railway Co. had been shipped through maritime commerce' — the absence of any discussion by the Supreme Court of a “threshold inquiry” akin to that found in our precedents is notable. We simply highlight this discrepancy and leave for a more appropriate case the question of how Norfolk Southern Railway Co. might circumscribe our “threshold inquiry” doctrine, if at all.

B. The Subject Matter of the Contract

What then is the subject matter of the contract in question? “Under the old, now outdated rule, [admiralty] jurisdiction was said to be reserved to ‘contracts, claims, and services purely maritime,’ Rea v. The Eclipse, 135 U.S. 599, 608, 10 S.Ct. 873, 34 L.Ed. 269 (1890) (emphasis added). The test has, however, been loosened considerably.” Sirius, 16 F.3d at 36 (parallel citations omitted). This Court has recognized “two exceptions to the general rule that ‘mixed’ contracts fall outside admiralty jurisdiction.” Transatlantic Marine Claims Agency, Inc. v. Ace Shipping Corp., 109 F.3d 105, 109 (2d Cir.1997). The first of those exceptions, which Folk-samerica does not assert here, is “that a federal court can exercise admiralty jurisdiction over a ‘mixed’ contract if ... the claim arises from a breach of maritime obligations that are severable from the non-maritime obligations of the contract.” Hartford Fire Ins. Co. v. Orient Overseas Containers Lines (UK) Ltd., 230 F.3d 549, 555 (2d Cir.2000); see, e.g., Compagnie Francaise de Navigation a Vapeur v. Bonnasse, 19 F.2d 777, 779 (2d Cir.1927). The second exception, which Folksamerica presses, allows courts to exercise admiralty jurisdiction where the non-maritime elements of a contract are “merely incidental” to the maritime ones. Transatlantic Marine Claims Agency, 109 F.3d at 109; see Sirius, 16 F.3d at 37. In examining this so-called “incidental” exception in the context of insurance policies, this Court has often noted whether the non-maritime portions of the examined policy were subject to separate premiums. See, e.g., Balfour, 85 F.3d at 75; Sirius, 16 F.3d at 36-37; cf. Atlantic Mutual, 968 F.2d at 198.

Once again, the recent pronouncement of Norfolk Southern Railway Co. calls for reconsideration of our precedent. Frequently, this Circuit and others have examined the so-called “incidental” exception in the context of intermodal transportation contracts, i.e., shipment contracts with sea and land components. See Hartford Fire Ins. Co., 230 F.3d at 555-56; Transatlantic Marine Claims Agency, 109 F.3d at 109-10; see also Sea-Land Serv., Inc. v. Danzig, 211 F.3d 1373, 1378 (Fed.Cir.2000); Berkshire Fashions, Inc. v. M.V. Hakusan II, 954 F.2d 874, 881 (3d Cir.1992); Kuehne & Nagel (AG & Co.) v. Geosource, Inc., 874 F.2d 283, 290 (5th Cir.1989). The Court in Norfolk Southern Railway Co. critiqued the analysis employed in these cases, noting, “it seems to us imprecise to describe the land carriage required by an intermodal transportation contract as ‘incidental’; realistically, each leg of the journey is essential to accomplishing the contract’s purpose.” 125 S.Ct. at 394-95. The Supreme Court suggested a shift in analysis. The important question should not be whether the non-maritime components of a bill of lading are “incidental” — that analysis might encourage separate contract negotiations and curb the efficiencies present in an intermo-dal contract, see id. at 394 — but, rather, should be whether the shipment contract’s “primary objective is to accomplish the transportation of goods by sea,” id. at 393. In pursuing this latter inquiry, the Court explained that a bill of lading’s “character as a maritime contract is not defeated simply because it also provides for some land carriage,” id. at 395, and concluded that the intermodal contract before it was maritime: “Conceptually, so long as a bill of lading requires substantial carriage of goods by sea, its purpose is to effectuate maritime commerce — and thus it is a maritime contract.... Geography, then, is useful ... only in a limited sense: If a bill’s sea components are insubstantial, then the bill is not a maritime contract.” Id. Thus, contrary to at least the terms of this Court’s jurisprudence, the Supreme Court exercised admiralty jurisdiction over a contract with non-maritime components deemed to be more than “incidental.”

The Norfolk Southern Railway Co. court’s analysis of an intermodal transportation contract suggests a global principle. In applying what we have previously called the “incidental” exception, we should focus “on whether the principal objective of a contract is maritime commerce,” id. at 394, rather than on whether the non-maritime components are properly characterized as more than “incidental” or “merely incidental” to the contract. Our “incidental” exception then might more accurately be termed the primary or principal objective exception. As an insurance contract is maritime to the extent that it is “marine insurance,” see Jeffcott, 129 F.2d at 584, admiralty jurisdiction will exist over an insurance contract where the primary or principal objective of the contract is the establishment of “policies of marine insurance,” Dunham, 78 U.S. (11 Wall.) at 35.

The parties disagree as to whether and to what extent the contract established marine insurance. Folksamerica focuses on the risks for which the Policy was procured, insists those risks are maritime, and contends the Policy is fundamentally marine — “marine insurance, issued to marine entities by a marine insurer, via a marine broker, to cover the insureds’ ship repair and cleaning operations,” Appellant’s Br. at 5. Clean Water counters that even if the SLL portion of this policy is marine, the CGL section is not. We conclude that the insurance contract at issue is primarily or principally concerned with maritime objectives, although there were incidental non-maritime elements. Accordingly, the district court erred in concluding that subject matter jurisdiction was lacking.

1. The CGL Section

Clean Water makes two chief objections to construing the CGL section of the Policy as marine insurance. The first objection is a matter of form. Clean Water argues that, unlike a traditional marine insurance policy, a CGL policy is a shore side insurance form and that a reasonable person reading it would therefore conclude that it covered nothing but shore side risks. Clean Water’s second objection is based on coverage. It claims, “The function of [ajdmiralty [¡jurisdiction is to ensure uniformity of laws as they apply to the carriage of cargo[e]s and passengers upon navigable waters. The fact [that] this CGL Policy covers none of these risks excludes it from consideration as a policy of marine insurance.” Appellee’s Br. at 13. We reject both arguments.

a. Form

First, despite Clean Water’s suggestion, a long line of precedent establishes that marine insurance need not take on a particular form. The Supreme Court first applied admiralty jurisdiction to an insurance contract in Insurance Co. v. Dunham, 78 U.S. (11 Wall.) at 30-31. See also De Lovio v. Boit, 7 F.Cas. 418 (C.C.Mass.1815) (No. 3,776) (Story, J.). The contract there insured a vessel against certain “perils of the seas.” Dunham, 78 U.S. (11 Wall.) at 2-3. In his opinion for the Court, Justice Bradley described marine insurance as a contract dependent on “maritime casualties” and “affected by and mixed up with all the questions that can arise in maritime commerce.” Id. at 30-31. The Court analogized marine insurance to other maritime contracts in which “payment is made to depend on a maritime risk.” Id. at 30. Significantly, and as we have noted elsewhere, the Dunham court defined the term “marine insurance” “in a general sense, and not in the particular sense of certain types of marine insurance related to the sea in a special way.” Jeffcott, 129 F.2d at 585 (describing Dunham).

Over seventy years later, in Jeffcott, this Court revisited Dunham’s description of marine insurance. The policies in Jeffcott — a hull policy and a shipowners liability policy — insured a schooner, the Dauntless. 129 F.2d at 584. A September hurricane took the previously moored Dauntless 500 feet in-shore and left it flooded and severely damaged. Id. When the insurer rejected the owner’s insurance claims, the owner filed a libel. Id. The insurer challenged federal jurisdiction, objecting that the policies were not maritime in nature because they required the vessel to be “laid up and out of commission” at a shipyard. Id. We disagreed, explaining that whether an insurance policy is marine insurance depends on “whether the insurer assumes risks which are marine risks.” Id. The policies in Jeffcott were simply marine insurance with a “decrease[d] ... quantum of risk.” Id. We stressed their coverage, not their form: “A non-marine insurance company might be familiar with damage from fire, theft, storm; it would not be prepared to assume such risks where the insured article was a ship. It would recognize that the risk took on a different character when a ship became the object insured.” Id.

The Supreme Court’s decision in Wilburn Boat Co. v. Fireman’s Fund Insurance Co., 348 U.S. 310, 75 S.Ct. 368, 99 L.Ed. 337 (1955), illustrates the primacy of substance over form in this regard. The Court in Wilburn Boat Co. exercised admiralty jurisdiction over a fire insurance policy dispute. The petitioners had purchased a policy to insure their houseboat “against loss from fire and other perils.” Id. at 311, 75 S.Ct. 368. After a fire destroyed the boat, the insurance company refused to pay for the loss, asserting that the petitioners breached policy terms. Id. The petitioners argued that Texas law applied, see id. at 311-12, 75 S.Ct. 368, and that, at the time, “no breach by [an] insured of the provisions of a fire insurance policy [could be] a defense to a[ ] suit under Texas law unless the breach contributed to the loss,” id. at 312, 75 S.Ct. 368. This Texas rule of law applied only to “fire insurance policies].” Id. at 312 n. 3, 75 S.Ct. 368. While the Court was split over the question of whether the Texas law should be given effect, not a single member of the Court questioned the propriety of the exercise of admiralty jurisdiction over the fire insurance policy. See id. at 313, 75 S.Ct. 368; id. at 321, 75 S.Ct. 368 (Frankfurter, J., concurring); id. at 325, 75 S.Ct. 368 (Reed, J., dissenting).

Like a fire insurance policy, there is nothing intrinsically “shore side” about a CGL policy. Indeed, other federal courts have treated CGL policies as marine insurance. In Larson Construction Co. v. Oregon Automobile Insurance Co., 301 F.Supp. 1112, 1113 (D.Or.1969), rev’d on other grounds, 450 F.2d 1193 (9th Cir.1971), the defendant issued the plaintiffs “a comprehensive liability policy, ... insuring, among other things, their operations upon navigable waters of the United States.” Looking to “the mobile nature of plaintiffs maritime business,” which consisted of loading barges, the court concluded “that defendant’s policy of insurance assures against maritime risks and is thus marine in nature.” Id. (citing Jeffcott, 129 F.2d at 582). In Granite State Minerals, Inc. v. American Insurance Co., 435 F.Supp. 159, 164 (D.Mass.1977), the district court held that a comprehensive general liability policy insuring a wharfinger against liabilities arising out of its operations was, likewise, a maritime contract.

Two other Circuits have treated CGL policies as marine insurance. Gulf Fleet Marine Operations, Inc. v. Wartsila Power, Inc., 797 F.2d 257, 258, 260 (5th Cir.1986) (Higginbotham, J.), involved a dispute between a ship repairer and its insurer regarding coverage under a broad form general' liability endorsement. The Fifth Circuit described the case as an admiralty case and applied the admiralty choice-of-law principles of Wilburn Boat Co. to the ship repairer’s CGL policy. Id. at 258, 260-61. More recently, in Great Lakes Dredge & Dock Co. v. City of Chicago, 260 F.3d 789, 791-92 (7th Cir.2001) (Easterbrook, J.), the Seventh Circuit exercised admiralty jurisdiction over CGL policies covering a marine excavation company, see Great Lakes Dredge & Dock Co. v. Commercial Union Assurance Co., 57 F.Supp.2d 525, 529 (N.D.Ill.1999).

Ultimately, coverage determines whether a policy is “marine insurance,” and coverage is a function of the terms of the insurance contract and the nature of the business insured. Thus, in Royal Insurance Co. of America v. Pier 39 Ltd. Partnership, 738 F.2d 1035, 1036 (9th Cir.1984), the Ninth Circuit found admiralty jurisdiction lacking even though the subject policies “at first glance look[ed] like typical marine insurance policies” with “[o]ne incorporating] the standard American Institute Hull Clauses” and “the other ... written on a standard protection and indemnity form.” In Youell v. Exxon Corp., 48 F.3d 105, 107, 109 (2d Cir.), vacated, 516 U.S. 801, 116 S.Ct. 43, 133 L.Ed.2d 9 (1995), on remand, 74 F.3d 373 (2d Cir.1996) (admiralty jurisdiction not questioned in subsequent opinions), by contrast, we exercised admiralty jurisdiction over a broadly-sweeping multi-part insurance policy termed a “Global Corporate Excess” agreement.

These cases illustrate that the form of — or label given to — an insurance policy will not always identify the nature of the risks the insurer assumes. Rather, as the First Circuit has explained, “an insurance policy’s predominant purpose, as measured by the dimensions of the contingency insured against ánd the risk assumed, determines the nature of the insurance.” Acadia Ins. Co. v. McNeil, 116 F.3d 599, 603 (1st Cir.1997). “Neither the policy’s form nor its ‘marine insurance’ title, therefore, is dispositive of the jurisdictional issue.” Kenneth H. Yolk, Symposium, Parsing the Admiralty Clause: Jurisdiction of Marine Insurance Transactions, 66 Tul. L. Rev. 257, 272 (1991). Simply classifying a CGL form as' “shore side” cannot resolve the question of. whether a.CGL policy is.marine insurance. We must look to the risks the insurer assumed.

, b. Coverage

. Clean Water contends that; because the CGL section excludes all the risks covered by hull and protection and indemnity (P & I) insurance, the only risks covered by the section are shore side risks such as personal injury and property damage to third parties. We disagree.

In general, CGL insurance provides coverage for liabilities arising from bodily injuries ánd property damage. While CGL insurance originally may hot have been “intended to serve as marine' insurance, ... the far flung activities of major corporate insureds have brought a number of maritime exposures within its scope.” Raymond P. Hayden & Sanford E. Balick, Symposium, Marine Insurance: Varieties, Combinations, and Coverages, 66 Tul. L. Rev. 311, 339 (1991). The insurance agent in this case, Trident Marine, issues CGL policies only to marine companies, usually in conjunction with other marine coverages. Issuance of CGL-like insurance policies to maritime businesses seems to be a growing trend. See Hayden & Balick, supra, at 338-52; Volk, supra, at 291; see also William E. O’Neil, Insuring Contractual Indemnity Agreements Under CGL, MGL, and P & I Policies, 21 Tul. Mar. L.J. 359, 370-73 & n.74 (1997); Hal C. Welch & Reginald R. White, III, Y2K and the Maritime Industry, 24 Tul. Mar. L.J. 125, 148 (1999).

Clean Water is correct that its CGL policy excludes certain traditional marine risks. The Policy requires “that any and all vessels owned and / or operated by the Insured” be “separately insured for Protection and Indemnity [P & I] risks ... and for Collision and Tower’s Liability.” “P & I insurance covers claims that arise in direct connection with the operation of [an] enrolled ship ... [that the insureds] own, operate, or charter,” Hayden & Balick, supra, at 327; see also Robert Force, Federal Judicial Center: Admiralty and Maritime Law 187 (2004), while collision insurance generally “covers an insured’s liability to third parties for damages arising out of a collision involving the insured’s vessel in which the insured vessel is at fault,” Navigators Ins. Co. v. Am. Home Assurance Co., 1999 WL 681161, at *3 (S.D.N.Y. Sept.1, 1999) (citing Hayden & Balick, supra, at 316-18). A provision of the Policy excludes coverage of risks included in certain standard-form P & I and collision policies “and all extensions of coverage thereto.”

These exclusions, however, do not dispose of all of the insureds’ maritime risks. P & I and collision policies leave significant gaps in the spectrum of maritime risks. The CGL section fills in some of these gaps. The basic format of the Policy’s CGL section mirrors that of typical CGL policies, “containing] provisions affording coverage for ‘premises/operations,’ ‘completed operations,’ and ‘products’ hazards.” 1 Barry R. Ostrager & Thomas R. NeWMAN, HANDBOOK ON INSURANCE COVERAGE Disputes § 7.02, at 353 (12th ed.2004). In addition to these three aspects of coverage, the Policy also provides pollution and limited contractual liability coverage. Through tailored endorsements, the Policy extends coverage to exposures particular to marine operations. Cf. Seabulk Offshore, Ltd. v. Am. Home Assurance Co., 377 F.3d 408, 413 (4th Cir.2004) (considering similar maritime endorsements).

Typical CGL policies exclude “liability for damages arising out of the ownership, maintenance, use, loading, or unloading of a watercraft.” 9 Couch on Insurance § 127:38 (3d ed.2004). The Policy voids that exclusion with a ‘Watercraft Liability Coverage” endorsement. A “Gulf of Mexico Extension” endorsement extends coverage to “operations of the Named Insured in the Gulf of Mexico.” The Policy contains an “In Rem” endorsement extending coverage to “any loss, otherwise covered ... even though asserted by an action ‘in rem’ instead of [an] action ‘in personam.’ ” There is little doubt that an in rem suit involving a vessel is decidedly salty — it is an admiralty proceeding. See, e.g., Am. Dredging Co. v. Miller, 510 U.S. 443, 446-47, 114 S.Ct. 981, 127 L.Ed.2d 285 (1994) (citing The Moses Taylor, 71 U.S. (4 Wall.) 411, 431, 18 L.Ed. 397 (1866)). Another endorsement involves pollution coverage. “It became obvious to the petroleum and tanker trades some time ago that extraordinary sources of financial protection would be necessary to meet the increasingly prohibitive costs of pollution clean-up and remediation activities.” Hayden & Balick, supra, at 335. The CGL section of the Policy includes a “buy back” provision, permitting pollution coverage in certain circumstances.

In examining the modifications to the CGL policy and the insureds’ business operations in maritime transport, repair, and maintenance, we discern at least three— perhaps four — aspects of coverage that should be deemed “policies of marine insurance.” They are, respectively, (1) the completed operations hazard coverage, (2) the products hazard coverage, (3) the pollution coverage, and, perhaps, (4) a portion of the premises and operation coverage. We agree with Clean Water, however,- that the contractual liability coverage is not maritime in nature.

The first two marine aspects of coverage are the Policy’s completed operations and products hazard coverage, both of which adopt “broad form property damage” exclusions. These areas of’ coverage have much in common. The completed operations provision “has been construed to mean that once work is completed by an insured, damage that thereafter occurs to the work itself is not covered by the policy, ... [but] consequential damages to property outside the work product arising from completed operations are not excluded.” Gulf Fleet Marine Operations, Inc., 797 F.2d at 260. See generally Roger C. Henderson, Insurance Protection for Products Liability and Completed Operations —What Every Lawyer Should Know, 50 Neb. L.Rev. 415 (1971). The products hazard provision has been understood “to exclude from coverage any obligation of the policyholder to repair or replace his own defective work or defective products,” but to cover “damages to the products of other than the named insured.” Gulf Miss. Marine Corp. v. George Engine Co., 697 F.2d 668, 670 (5th Cir.1983). See generally Henderson, supra. The Policy sets the aggregate liability limit for the completed operations and products hazards at $1,000,000.

Coverage for completed operations hazards, like the coverage here, can expose insurers to the costs of various maritime accidents resulting from an insured’s negligence. For instance, in Gulf Mississippi Marine Corp. v. George Engine Co., 697 F.2d at 669-70, 673, the Fifth Circuit considered a claim under, inter alia, the completed operations hazard of an insurance policy, seeking damages alleged to result from the defective “installation of major components” of certain vessels. Likewise, in Gulf Fleet Marine Operations, Inc., 797 F.2d at 258, 260-61, the Fifth Circuit held that a completed operations hazard provision required an insurer to pay for damages to -a vessel’s engine block and crankshaft that resulted from the. insured’s faulty overhaul of the ship. Cf. Gulf Tampa Drydock Co. v. Great Atl. Ins. Co., 757 F.2d 1172, 1175 (11th Cir.1985) (holding that insurance lacking a completed operations hazard provision would not cover a repairer’s liability for damages to a vessel resulting from faulty workmanship). Here, the insureds’ repair and maintenance operations create similar exposures.

A products hazard may require an insurer to cover liabilities similar to those arising from a completed operations provision. In Todd Shipyards Corp. v. Turbine Service, Inc., 674 F.2d 401, 406, 422-23 (5th Cir.1982), for instance, the Fifth Circuit held that a ship turbine company!s products hazard policy covered damages to a vessel that resulted from faulty turbine blades, including “damages for loss of use of the vessel, general expenses from .the master’s accounts, and pilotage, wharfage, tub, repatriation and recrewing expenses,” id. at 423.

The important difference between the products hazard and the completed operations hazard, of course, is that the prod ucts hazard ties damages to accidents caused by a particular “product,” rather than an “operation.” Accordingly, the Fifth Circuit in Gulf Mississippi Marine Corp. could apply products hazard coverage for liabilities arising from “defects in the component parts” designed and manufactured by an insured and installed into a vessel and completed operations hazard coverage to liabilities arising from “improper installation of the component parts and other construction defects.” 697 F.2d at 671.

We view the coverage set out by the completed operations and products hazards provisions as marine in nature. Several considerations inform this conclusion. First, as noted, the Supreme Court has instructed that “[precedent and usage are helpful insofar as they exclude or include certain common types of contract,” Kossick, 365 U.S. at 735, 81 S.Ct. 886, and here the most apposite precedent, Gulf Fleet Marine Operations, 797 F.2d at 258, 260-61, supports admiralty jurisdiction. In Gulf Fleet Marine Operations, as here, a CGL policy covered, inter alia, the completed operations liabilities of a ship repair and maintenance business. The Fifth Circuit, focusing on the completed operations coverage, treated the CGL policy as one of marine insurance and adopted the choice-of-law analysis presented in Wilburn Boat Co. and other marine insurance cases. See id. at 261 (citing Wilburn Boat Co., 348 U.S. 310, 75 S.Ct. 368, 99 L.Ed. 337, and Eagle Leasing Corp. v. Hartford Fire Ins. Co., 540 F.2d 1257, 1261 (5th Cir.1976) (“In the absence of specifically controlling federal authority, the law of the state where a marine insurance contract is issued and delivered governs the construction of its language.”)).

Second, the coverage afforded to a ship repair and maintenance business by the CGL’s completed operations and products hazards provisions falls within all of the various formulations of “marine insurance.” The risk of injuries from, and damage to, a ship after defective or faulty repair or maintenance is “a maritime risk” that includes the risk of “the loss of the ship or goods” in the language of Dunham, 78 U.S. (11 Wall.) at 30. It is a “risk[] peculiar to ship and sea” as set out in Jeffcott, 129 F.2d at 584, and is “related to hazards encountered in maritime transportation” and “vessels or the maritime industry,” Acadia Ins. Co., 116 F.3d at 601-02. The coverage relates directly to “the navigation, business or commerce of the sea,” De Lovio, 7 F. Cas. at 444, and applies both “against marine risks” and to “maritime interests” (i.e., damage to and injuries upon vessels), Royal Ins. Co. of Am., 738 F.2d at 1036-37.

Third, in the context of ship repair and maintenance businesses, the losses covered by the completed operations and products hazards provisions overlap with those covered by P & I insurance. The district court cited Benedict on Admiralty for the proposition that P & I insurance “addresses losses such as those arising from the injury or death of seamen and passengers, unemployment caused by shipwreck, diversion expenses, stowaway and refugee liability, collision with other ships, quarantine expenses, and towage liability.” Folksamerica, 281 F.Supp.2d at 533 (citing Benedict on Admiralty, Marine Insurance: Protection and Indemnity Club Rules [and] Institute Clauses, Vol. 7A (Matthew Bender 2002)). The district court concluded, “These are not the types of risks that Section I addresses.” Id. Quite the opposite is true: These are precisely the types of losses that come within a completed operations or products hazard. Cf. Todd Shipyards Corp., 674 F.2d at 423; see also 12 Couch on Insurance, supra, § 172:32; 1 Ostrager & Newman, supra, § 7.03[b][2][D], at 399-401. For these reasons, we conclude that both the completed operations and products hazards coverage provisions are “policies of marine insurance.”

A third aspect of coverage that we find marine in nature is the Policy’s pollution coverage. Standard CGL policies include a pollution exclusion clause that limits coverage to “sudden and accidental” pollution or discharge. See 1 Ostrager & Newman, supra, § 10.02[a], at 621. The CGL section of the Policy implements a “buy back” arrangement analogous to that of some contemporary P & I policies. See, e.g., 7A Benedict on Admiralty § 2.06-1; Hayden & Balick, supra, at 334. It first implements “absolute pollution exclusions” that broadly exclude pollution liabilities, see, e.g., 1 Ostrager & Newman, supra, § 10.02[e], at 651-59, and then provides a “Limited Pollution Buy-Back (72 Hour) Clause (02/94)” that allows for reinstatement of coverage under certain conditions. Cf. Certain Underwriters at Lloyd’s London v. C.A. Turner Constr. Co., Inc., 941 F.Supp. 623, 626-29 (S.D.Tex.1996); Jenkins v. Bill Lawrence, Inc., 2000 WL 1751613, at *2-6 (E.D.La. Nov.28, 2000). The insured’s business operations in oil and cargo transportation renders pollution coverage potentially significant. Cf. In re Exxon Valdez, 270 F.3d 1215, 1250-53 (9th Cir.2001); In re Nautilus Motor Tanker Co., Ltd., 85 F.3d 105 (3d Cir.1996); United States v. Oswego Barge Corp. (In re Oswego Barge Corp.), 664 F.2d 327, 334 (2d Cir.1981). The Policy does not set a separate pollution liability limit, simply specifying that “all other coverages” are limited to $2,000,000.

Pollution coverage is widely recognized as marine in nature. See Youell, 48 F.3d at 107 (“[The policy] covers marine liabilities, such as fines and penalties for pollution.”); id. at 109; see also M/G Transp. Servs., Inc. v. Water Quality Ins. Syndicate, 234 F.3d 974, 975-76 (6th Cir.2000) (exercising admiralty jurisdiction over a coverage dispute under marine pollution liability insurance); FORCE, supra, at 184-85, 187; Hayden & Balick, supra, at 334-38. The traditional characterization of P & I insurance as marine insurance supports that recognition, as some level of pollution coverage is often included in P & I insurance policies. See id. at 334; see, e.g., 2 Thomas J. Sohoenbaum, Admiralty and Maritime Law § 19-12 (4th ed. 2004) (“Shipowners look to Protection and Indemnity Clubs for insurance against their liabilities including claims for damage or compensation in respect of ... pollution.”). The pollution coverage is thus a third component of the Policy constituting marine insurance.

Fourth, at least a significant portion of the premises and operations coverage should probably be characterized as marine. Clean Water is arguably right that the slip and fall protection covering injuries on the insureds’ shore-side premises are not marine in nature. Cf. Ocean Marine Ins. Co. v. Wickland Corp., 1995 WL 125478, at *3 (N.D.Cal. Mar.8, 1995) (finding a petroleum trading company’s insurance “against liability arising from the personal injury or death of any person” to be non-marine). However, as Milton Rivera’s accident illustrates, the Policy may be called upon to cover liabilities arising from personal injuries sustained during repair and maintenance operations. Many repair and maintenance tasks were once carried out by crew members. Cf. Atl. Transp. Co. of W. Va., 234 U.S. at 61-62, 34 S.Ct. 733. The liabilities arising therefrom are thus analogous to liabilities that would have been included in a P & I insurance policy, see Force, supra at 185; Hayden & Balick, supra, at 326-27; 2 Schoenbaum, supra, § 19-12, and would probably best be characterized as policies of marine insurance. But cf. Simon v. Intercont’l Transp. (ICT) B.V., 882 F.2d 1435, 1443 (9th Cir.1989). Indeed, Clean Water appears to concede that the premises and operations coverage is maritime to the extent that the premises involved are vessels. But, as explained in the concluding section below, we need not resolve this particular question today.

Finally, as noted above, we conclude that the contractual liability coverage provided by the CGL section is not marine in nature. The section covers liabilities contractually assumed only under “incidental contracts” and defines “incidental” as related to the conduct of the insureds’ businesses. Cf. Lafarge Corp. v. Hartford Cas. Ins. Co., 61 F.3d 389, 395-97 (5th Cir.1995). The CGL policy does not create general contractual liability coverage for contracts, even “incidental” ones, see, e.g., Pace Constr. Co. v. U.S. Fid. & Guar. Ins. Co., 934 F.2d 177, 179-80 (8th Cir.1991); see also 1 Ostrager & Newman, supra, § 7.05, at 460-61, but, rather, provides coverage for personal injury and property damage liabilities that arise under “incidental” indemnification and hold-harmless agreements, see id. at 461; see also Gen. Utils., Inc. v. Interboro Mut. Indem. Ins. Co., 117 A.D.2d 647, 498 N.Y.S.2d 166 (2d Dep’t 1986). Thus, the policy might cover the personal injury or property damage liabilities of, for instance, a subcontractor that one of the insureds had agreed to indemnify as a general contractor, cf. Thermo Terratech v. GDC Enviro-Solutions, Inc., 265 F.3d 329, 333 & n. 7, 335, 337 (5th Cir.2001), or a general contractor that one of the insureds had agreed to indemnify as a subcontractor, cf. United Nat’l Ins. Co. v. Dunbar & Sullivan Dredging Co., 953 F.2d 334, 337 (7th Cir.1992).

The contractual liability coverage is not marine. The Supreme Court has noted that “an agreement to pay damages for another’s breach of a maritime charter is not [a maritime contract],” Kossick, 365 U.S. at 735, 81 S.Ct. 886, and we have held that an “agreement to contribute to the settlement of [a] charter claim is not a maritime contract,” Fednav, Ltd. v. Isoramar, S.A., 925 F.2d 599, 601 (2d Cir.1991). But see Haller v. Fox, 51 F. 298 (W.D.Wash.1892). As we can see no principled distinction between the breach of the maritime contract of charter and the breach of maritime contracts for, e.g., ship repair, we will not construe the contractual liability coverage of the Policy as marine. The Supreme Court in Kossick cited Pacific Surety Co. v. Leatham & Smith Towing & Wrecking Co., 151 F. 440 (7th Cir.1907), as support for the assertion that a contract to pay for the breach of a charter contract is not maritime. Kossick, 365 U.S. at 735, 81 S.Ct. 886. The reasoning of Pacific Surety reveals no ground for distinguishing between a charter contract guarantee and a guarantee of any other maritime contract:

True, the obligation of the surety hinges upon the breach of contract (maritime) on the part of the principal, and the charter party is thus brought in by way of evidence of the breach and of liability to respond in damages under the covenants of the bond. The direct subject-matter of the suit is the covenant to pay such damages, which neither involves maritime service nor maritime transactions; and we are of [the] opinion that the mere fact that the event and measure of liability are referable to the charter party does not make the bond a maritime contract, nor make its obligation maritime in the jurisdictional sense.

Pac. Sur. Co., 151 F. at 443. We have looked to the reasoning of Kossick to explain why “merely agreeing as surety ‘to pay damages for another’s breach of a maritime charter is not’ a maritime con tract.” Fednav, 925 F.2d at 601 (quoting Kossick, 365 U.S. at 735, 81 S.Ct. 886). That logic applies equally to transportation and ship repair and maintenance contracts. Indeed, it applies doubly to CGL policies that provide coverage for a breach of an indemnity agreement, the insurance being a surety agreement to a surety agreement.

To summarize then, of the various protections provided in the CGL section of the Policy — completed operations, products, pollution, premises and operations, and contractual liability — the first three are decidedly marine in nature and the fourth covers both shore side and maritime risks. Clean Water is thus mistaken: The CGL section does reach maritime risks and was, in fact, specifically modified to do so.

2. The SLL Portion

The second section of the Policy, Shiprepairers Legal Liability (SLL), provides coverage for vessels lost or damaged while undergoing repairs by the insureds. SLL coverage is limited to vessels not owned by the insureds when those vessels are in the insureds’ care, custody, and control, and undergoing either repairs or, upon prior notification and agreement, other operations not “within the scope of ... shiprepairing.” A “Travelling Workmen Clause” extends coverage to work “on board the Vessel and/or Drilling Rig at sea or in any port for the purpose of effecting repairs and/or other work entrusted to the Assured .... ”

The SLL section of the Policy is decidedly marine. “There are few objects— perhaps none — more essentially related to maritime commerce than vessels,” Sirius, 16 F.3d at 36, and, as we explained in Jeffcott, although “[a] non-marine insurance company might be familiar with damage from fire, theft, storm; it would not be prepared to assume such risks where the insured article was a ship. It would recognize that the risk took on a different character when a ship became the object insured,” 129 F.2d at 584. A number of courts have exercised admiralty jurisdiction over SLL policies — none have refused to do so. The earliest appears to be Heipershausen Bros. v. Continental Insurance Co. of City of New York, 25 F.Supp. 1010, 1010-12 (S.D.N.Y.1938), in which the United States District Court for the Southern District of New York resolved a dispute between a ship repairer and its SLL insurer regarding coverage of the insured’s liability for a fire that damaged a boat undergoing repair. More recently, in Gulf Tampa Drydock Co., the Eleventh Circuit applied admiralty law to a ship repairer’s liability policy quite similar to the SLL policy here. 757 F.2d at 1173-75. Finally, in Hebert v. Exxon Corp., 770 F.Supp. 314, 317-18 (E.D.La.1991), vacated in part on other grounds by Hebert v. Exxon Corp., 1991 WL 220394 (E.D.La. Oct.21, 1991) (Hebert II), the Eastern District of Louisiana treated as “marine insurance,” id. at 315, a SLL policy covering an company involved in certain repair and maintenance operations, see id. at 317-19; see also Hebert II, 1991 WL 220394, at *1-2 (applying federal maritime law to construe the SLL agreement). In light of these precedents and no persuasive argument to the contrary, we conclude that the SLL section of the Policy creates coverage of marine insurance.

8. Conclusion

The district court erred in concluding that it was without admiralty jurisdiction under 28 U.S.C. § 1333(1). The two sections of the Policy together operate seamlessly to provide coverage that is primarily marine in nature. The sections share a premium and are subject to the same deductible and liability limitations. The Policy is custom-built to fill the gaps that traditional marine insurance policies — hull, collision, and P & I — leave in maritime-industry coverage. The SLL section protects against property damage to vessels undergoing repair, and the CGL section adds completed operations, products, pollution, contractual liability, and premises and operations coverage. Combined, the CGL and SLL provisions round out the insureds’ coverage for maritime transport operations and give fairly robust ship repair and maintenance coverage. In the circumstances of this case — a contract with two sections, one of which provides fully marine insurance, and the other specifically modified to cover maritime risks— we conclude that the Policy is marine in nature without the necessity of determining whether the premises and operations coverage provided here is marine or not.

We therefore hold that, because the primary objective of the Policy is to establish marine insurance, the district court had jurisdiction under 28 U.S.C. § 1333(1).

Conclusion

Accordingly, we vacate the district court’s dismissal and remand for further proceedings consistent with this opinion.

1 . Clean Water has since settled its claim with Rivera.

2 . The parties agree for the purposes of summary judgment that, but for mutual mistake, the Policy would have listed Clean Water among the named insureds. See id. at 531 n. 1.

3 . Indeed, Professor Charles Black once stated that "in this field ... [t]he attempt to project some 'principle' is best left alone. There is about as much ‘principle’ as there is in a list of irregular verbs.” Charles L. Black, Jr., Admiralty Jurisdiction: Critique and Suggestions, 50 Colum. L.Rev. 259, 264 (1950) (footnote omitted), quoted in part by Sisson, 497 U.S. at 372 n. 4, 110 S.Ct. 2892 (Scalia, J., concurring).

4 . See, e.g., N. Pac. S.S. Co., 249 U.S. at 128-29, 39 S.Ct. 221 (exercising admiralty jurisdiction over a ship repair contract); New Moon Shipping Co. v. MAN B & W Diesel AG, 121 F.3d 24, 28 (2d Cir.1997) ("A claim arising out of a contract to repair a ship falls squarely within a federal court’s admiralty jurisdiction ....”); Sundance Cruises Corp. v. Am. Bureau of Shipping, 7 F.3d 1077, 1081 (2d Cir.1993) (exercising admiralty jurisdiction over a ship inspection contract); CTI-Container Leasing Corp., 682 F.2d at 379 ("Traditional texts have defined a ‘maritime’ contract as one ... 'for the furnishing of services, supplies, or facilities to vessels ... in maritime commerce or navigation.' ” (quoting 7A James Wm. Moore et al., Moore’s Federal Practice ¶ .230[3], at 2773 (2d ed.1948))); Wilmar Marine Eng’g & Sales Corp. v. M/V Perseo, 278 F.Supp. 345, 347, 352 (E.D.La.1967) (exercising admiralty jurisdiction over an oil tank cleaning contract).

5 . Cf. Atl. Transp. Co. of W. Va. v. Imbrovek, 234 U.S. 52, 61-62, 34 S.Ct. 733, 58 L.Ed. 1208 (1914) ("[T]he wrong which was the subject of the suit was, we think, of a maritime nature .... The libelant was injured on a ship, lying in navigable waters, and while he was engaged in the performance of a maritime service. We entertain no doubt that the service in loading and stowing a ships’s cargo is of this character. Upon its proper performance depend in large measure the safe carrying of the cargo and the safety of the ship itself; and it is a service absolutely necessary to enable the ship to discharge its maritime duty. Formerly the work was done by the ship’s crew; but, owing to the exigencies of increasing commerce and the demand for rapidity and special skill, it has become a specialized service devolving upon a class as clearly identified with maritime affairs as are the mariners.” (Internal quotation marks omitted)).