PATRICK E. HIGGINBOTHAM, Circuit Judge:

This is a Carriage of Goods by Sea Act claim for rust damage to steel coils which their owner alleges was caused by seawater when shipped from Latvia to the United States on the M/V LAKE MARION. The district court awarded damages against the vessel, its owner, Lake Marion, Inc., its manager, Bay Ocean Management, Inc., collectively the "vessel interests," and the time charterer, Western Bulk Carriers K/S Oslo. Finding no error, we affirm.

I.

The plaintiff, Steel Coils, Inc., is an importer of steel products with its principal office in Deerfield, Illinois. It ordered flat-rolled steel from a steel mill in Russia. Itochu, which then owned ninety percent of the stock of Steel Coils, purchased the steel and entered into a voyage charter with Western Bulk for the M/V LAKE MARION to import the steel to the United States. Western Bulk had time chartered the vessel from Lake Marion, Inc. As Lake Marion, Inc.'s manager, Bay Ocean employed the master and crew of the vessel.

The LAKE MARION took on the steel coils at the Latvian port of Riga between February 26 and March 2, 1997. The steel had traveled to port by rail from the Severstal steel mill 400 miles north of Moscow. At Riga, the hot rolled coils were stored outside, while the cold rolled and galvanized coils were encased in protective steel wrappers and stored in a warehouse at the port.

After departing Riga, the vessel stopped at another Latvian port, Ventspils, where it took on more steel coils. The ship departed Ventspils on March 7, 1997 and arrived at Camden, New Jersey, on March 28, 1997. After Camden, the ship stopped at New Orleans and Houston. Steel Coils alleged that the coils unloaded at New Orleans and Houston were damaged by saltwater, which required Steel Coils to have the cargo cleaned and recoated.

II.

Steel Coils filed suit under COGSA against the M/V LAKE MARION in rem and against Lake Marion, Inc., Bay Ocean Management, and Western Bulk Carriers in personam, requesting $550,000 in damages, with a separate claim of negligence against Bay Ocean. The vessel interests and Western Bulk filed cross-claims against each other for indemnification, and Western Bulk filed a third party complaint for indemnification against Itochu.

After a bench trial, the district court held the defendants jointly and severally liable to Steel Coils for $262,000, and Bay Ocean liable for an additional $243,358.94. The court further found that Western Bulk was entitled to indemnity from Lake Marion, Inc. for any amount it pays to Steel Coils. It dismissed with prejudice Lake Marion's cross-claim against Western Bulk and Western Bulk's third party complaint against Itochu. From this judgment the vessel interests appeal, and Steel Coils and Western Bulk cross-appeal. Steel Coils's and Western Bulk's cross-appeals become relevant only if we find the vessel interests' points of error meritorious.

III.

Defendants M/V LAKE MARION, Lake Marion, Inc., and Bay Ocean contend that the district court improperly shifted the burden to them to prove that the steel cargo was not in good condition prior to loading or was in undamaged condition at discharge, that it erred in finding that they failed to exercise due diligence to ensure that the vessel was seaworthy at the commencement of the voyage, and that it was wrong in disregarding their defenses to COGSA liability of peril of the sea and latent defect. They also assert the district court should not have held Bay Ocean liable to Steel Coils in tort separate from the COGSA claim, depriving Bay Ocean of the COGSA $500-per-package limitation on damages. We find these arguments unavailing.

IV.

In admiralty cases tried by the district court without a jury, we review the district court's legal conclusions de novo, and its factual findings under the clearly erroneous standard. "The clearly erroneous standard of review does not apply to decisions made by district court judges when they apply legal principles to essentially undisputed facts."

COGSA provides a complex burden-shifting procedure. Initially, the plaintiff must establish a prima facie case by demonstrating that the cargo was loaded in an undamaged condition and discharged in a damaged condition. "For the purpose of determining the condition of the goods at the time of receipt by the carrier, the bill of lading serves as prima facie evidence that the goods were loaded in the condition therein described." If the plaintiff presents a prima facie case, the burden shifts to the defendants to prove that they exercised due diligence to prevent the damage or that the damage was caused by one of the exceptions set forth in section 1304(2) of COGSA, including "[p]erils, dangers, and accidents of the sea or other navigable waters" and "[l]atent defects not discoverable by due diligence." If the defendants show that the loss was caused by one of these exceptions, the burden returns to the shipper to establish that the defendants' negligence contributed to the damage. Finally, "if the shipper is able to establish that the [defendants'] negligence was a contributory cause of the damage, the burden switches back to the [defendants] to segregate the portion of the damage due to the excepted cause from that portion resulting from the carrier's own negligence."

A.

The vessel interests first assert that the district court reversed the burden of proof, requiring them to demonstrate that the goods were loaded in a damaged condition or were unloaded in an undamaged condition instead of requiring Steel Coils to prove that the coils were loaded undamaged and discharged damaged. These defendants mischaracterize the district court's decision. The district court properly explained that under COGSA a plaintiff must establish a prima facie case by "proving that the cargo for which the bills of lading were issued was loaded in an undamaged condition, and discharged in a damaged condition." Applying this law the trial court determined that Steel Coils demonstrated its prima facie case by proving that "the cargo was delivered to the LAKE MARION in good order and condition" and "was unloaded at the ports of New Orleans and Houston in a damaged condition."

The district court cited specific evidence proffered by Steel Coils to support these conclusions. In determining that the coils were loaded in good condition, it examined "mates receipts, bills of lading containing comments on the condition of the cargo, and a cargo survey taken at the load port in Riga that contained commentary about and photographs of the cargo." It explained that although some of these documents contained notations regarding "atmospheric rust on the hot rolled coils and damage to the wrapping of the cold rolled and galvanized coils," the evidence showed that "these conditions did not damage the coils" and were not the result of exposure to seawater prior to embarkation.

In looking at the evidence of the coils' unloading, the court found that "[a]ll of the surveyors at the discharge ports testified that the cargo was damaged when it was discharged . . . and their survey reports support their testimony." Moreover, relying on these surveyors' reports, as well as those of chemists who tested the rust on the coils, the court concluded that the rust damage was a result of seawater contamination. The trial court affirmatively rejected the testimony of the defendants' expert, Sanchez, who testified that the cargo was not contaminated by seawater.

The trial court correctly placed the burden upon Steel Coils to prove a prima facie case and examined the evidence to determine whether in fact this had been done. Although the vessel interests couch their complained-of error as improper burden shifting, their argument attacks the district court's factual findings that the cargo was undamaged before loading and damaged at unloading. They assert that the district court wrongly determined the coils were undamaged at loading because the bills of lading noted that the hot rolled coils were rust stained and wet and noted that the condition of the cold rolled and galvanized steel coils, which were encased in steel wrappers, was unknown. They also argue that the district court did not rely on competent evidence in determining that the cargo was damaged at discharge.

The district court found from the evidence presented that the notations on the bills of lading indicating rust staining and moisture on the hot rolled coils did not affect their good condition prior to loading. Evidence in the record supports this conclusion. Captain Sparks, Steel Coils's expert on carriage of steel cargo by sea, explained that the notations regarding rust on the bills of lading were "non-restrictive clause[s] indicating that the goods [were] undamaged but affected by a form of atmospheric rust normal on all mild steel surfaces which are untreated against oxidation." This atmospheric rust, which he also termed "fresh water rust," is caused by the storage of hot rolled coils in open air prior to loading onto the vessel. Sparks stated that this type of rust does not result in harmful deterioration of the material, and only if the hot rolled coils come into contact with "an electrolyte more aggressive than fresh water[,] e.g. saltwater," does permanent damage to the plating occur. Moreover, he explained that "[i]t is not unusual for the goods to be shipped in an apparent rusty condition and wet," and despite the notations on the bills of lading, all of the bills of lading were signed clean, and therefore "all cargo when shipped was in apparent good order and condition."

Despite this evidence, the vessel interests argue that the rust found on the hot rolled coils prior to loading was probably caused by exposure to saltwater, because the coils were stored outdoors for an indefinite amount of time in a "marine environment." They maintain that saltwater could have been carried by the wind from the sea to the place at which the coils were stored. However, as the district court noted, the hot rolled coils loaded at Ventspils, which were also noted as being rust stained on their bills of lading and, in accordance with custom, had likely also been transported and stored outdoors, tested negative for exposure to saltwater when subjected to silver nitrate tests prior to loading. The vessel interests argue that these silver nitrate tests are irrelevant because they were conducted at Ventspils, not Riga. Nevertheless, that the hot rolled coils at Ventspils had the same notations on their bills of lading and tested negative for saltwater exposure prior to loading clearly bears on the plausibility of the defendants' theory that the pre-loading rust on the Riga hot rolled coils was seawater rust caused by the marine environment.

In Thyssen, Inc. v. S/S Eurounity, the Second Circuit confronted similar facts. The shipper sued the carriers for rust damage to hot rolled steel coils. However, the bills of lading contained notations such as "rust stained," "partly rust stained," and "wet before shipment." The defendants urged that these notations prohibited a finding that the coils were in good condition upon loading. The Second Circuit disagreed, finding "ample evidence that the steel was in good condition." For instance, experts the district court found credible testified that the clauses on the bills of lading indicated steel of good condition. One expert testified that "[t]he Port . . . had used these standardized notations for approximately thirty years to refer to nondamaging, atmospheric rust that does not affect the value of steel. [The expert further] testified that steel is considered to be in `prime' condition when the bills of lading include these standardized notations." On the basis of this evidence, the Second Circuit affirmed the district court's determination that the shipper had made out a prima facie case.

We conclude that the district court in this case did not clearly err in finding that the hot rolled coils were in good condition prior to loading. That the rust noted on the coils was atmospheric and nondamaging in nature, and that the moisture on the coils also did not affect their good condition is supported by the evidence.

As for the cold rolled and galvanized coils, the vessel interests argue that the bill of lading notation that the condition of these coils was unknown fatally undermined Steel Coils's attempt to prove a prima facie case of good condition. For this argument they rely on Caemint Food, Inc. v. Lloyd Brasileiro Companhia de Navegacao, in which the Second Circuit reasoned:

Although a clean bill of lading normally constitutes prima facie evidence that cargo was in good condition at the time of shipment . . . it does not have this probative force where . . . the shipper seeks to recover for damage to goods shipped in packages that would have prevented the carrier from observing the damaged condition had it existed when the goods were loaded.

Caemint held that a plaintiff could not recover for corned beef it claimed was ruined during the voyage because it could not present evidence as to the condition of the corned beef, which was inside metal containers, before shipment.

We have similarly stated that "[w]here because of the perishable or intrinsic nature of the commodity, the internal condition is not adequately revealed by external appearances, cargo may have a considerable burden of going further to prove actual condition." That is not the case here. Captain Sparks testified that although the wrappers of the cold rolled and galvanized coils loaded at Riga were wet due to condensation, there was no evidence of "drip-down" or "run-down" of moisture to the coils and no mention in the bills of lading of "white rust or white oxidation marks," which are normal preshipment clauses indicating possible rust damage to the coils. He concluded that "[t]he amount of moisture on those coils must have been negligible" and "[t]here was no damage to those coils."

The evidence at trial showed that, had the cold rolled or galvanized coils been damaged by rust, their outer wrappers would have revealed it. Because the wrappers had no indication of rust, and the moisture on the outside of the wrappers was not dripping down into the coils, it was not clearly erroneous for the district court to conclude that the cold rolled and galvanized coils were in an undamaged state prior to loading.

The vessel interests further assert clear error in the finding that the steel coils were damaged upon unloading. The contention is that seawater could not have entered through the hatches because the top-stowed cargo unloaded at Camden had no seawater damage, and that perhaps the steel coils rusted on the way from the ship to their ultimate inland destinations.

The vessel interests' arguments are belied by a wealth of evidence relied upon by the district court that at unloading the cargo was damaged by seawater rust. For instance, the McLarens Toplis survey conducted in New Orleans noted "rust stains to coils to varying extents," and "[r]andom tests on the rust stained areas with a solution of silver nitrate proved positive" with respect to chlorides, "indicating water ingress." The McLarens Toplis survey in Houston similarly stated that "[t]he cargo was examined and found to be extremely rusty," and that "[e]xtensive silver nitrate tests were conducted with strong positive results. It is our opinion [that] the . . . cargo came into contact with sea water, most likely through the poorly maintained hatch covers. . . ."

The vessel interests have not cited any evidence in the record that disputes these conclusions. Their argument that the district court "simply accepted plaintiff's survey reports and testimony en masse as setting forth the proper measure of damages, and that the damage was proven at discharge" implicitly acknowledges that substantial evidence in the record supports the district court's conclusions as to the damage evident at unloading.

B.

Facing a prima facie case, a defendant may escape liability if it shows that it exercised "due diligence . . . to make the ship seaworthy, and to secure that the ship is properly manned, equipped, and supplied, and to make the holds . . . and all other parts of the ship in which goods are carried fit and safe for their reception, carriage, and preservation." The vessel interests urge that even if Steel Coils carried its initial burden, they exercised due diligence in making the vessel seaworthy and thus should have escaped liability.

In making its determination that the defendants did not exercise due diligence, the district court correctly noted that seaworthiness is defined as "reasonable fitness to perform or do the work at hand," and explained that, under COGSA, the carrier's duty to exercise due diligence in making the vessel seaworthy is nondelegable. It concluded that the ship was not reasonably fit to perform the work at hand — shipping steel coils — because the hatches were not maintained in good condition and had not been tested for watertightness before embarkation, which had resulted in an ingress of seawater during the voyage, and because the holds, which had previously carried a cargo of rock salt, had not been washed out with fresh water before the steel was loaded.

The vessel interests dispute these findings, arguing that the surveys done prior to embarkation show that the hatches were in watertight condition. They further deny any obligation to test the watertightness of the hatch covers prior to the voyage, because by the voyage charter that obligation belonged to Itochu. Finally, they deny any contractual obligation to wash the holds with fresh water instead of Baltic Sea water.

Although preload surveys conducted at Riga found the hatches sufficient, one of the surveys also noted deficiencies in the hatch covers. The SKS International Cargo Service survey found that "[t]he hatchcover tightening rubbers are deformed (pressed inside) more or less everywhere. Also[,] the ends of these rubbers are pressed out or glued out on the extremes of hatch-covers." The report further detailed that "[t]he condition of hatch trackways and coamings [was] found as satisfactory but rusty and with the traces of corrosion. The compression bar is partly bent (deformed) . . . due to wear and tear."

Captain Sparks explained that the description of the hatches found in the SKS survey "indicates that the required maintenance was not performed." Moreover, he opined that the preload surveyors' conclusion that despite these deficiencies the hatches were watertight is not reliable because the "watertight integrity of the hatches was not determined [by a hose test or ultrasound test], as is customary with steel cargoes in the industry prior to commencement of the sea voyage."

That the hatches were insufficiently maintained is further supported by the observations of the Seaspan Marine Consultants surveyor who inspected the hatches after the vessel docked in Camden. He noted that the rubber gaskets of the hatch panels had deep grooves in them, were worn out in several places, were heavily rusted and bent or waved in certain areas, and that parts of the gaskets were missing or cut in some places.

In line with these observations, the McLarens Toplis surveyor in Houston reported after inspection of the vessel:

The vessel appeared to be very poorly maintained. Our inspection of the hatch covers noted them to be in extremely poor condition. We noted large amounts of rust flaking off at the slightest of touches. We noted the compression bars, channel bars, and general areas around the hatch coaming to be severely dented, gouged, and holed to varying degrees. We noted the hatch packings to be gouged, missing, and the general area around the hatch packings to be wasted severely. We also noted the channel areas to be bent. Severe drip downs were noted in all hatches.

It is our opinion [that] the cargo hatches were poorly maintained and appeared to have severe sea water ingress in all cargo hatches that were of concern at this port.

Captain Sparks concluded from these descriptions that "when the vessel commenced the voyage from the Baltic Sea to Camden the hatches were defective owing to lack of maintenance," and as a result they "were not weather tight." Although the vessel interests take issue with the competency of this evidence and argue that the preload survey reports should have been given more weight than Captain Sparks' "hindsight" opinions, there was sufficient evidence to support the district court's conclusion that the hatches were inadequately maintained and it was not clearly erroneous.

The vessel interests also argue that they were not responsible for conducting a watertightness test on the hatch covers prior to embarkation, because pursuant to the voyage charter Itochu was supposed to "make an inspection of holds and test watertightness of hatches." This argument ignores the COGSA carrier's nondelegable duty to ensure that the vessel is reasonably fit to carry steel cargo.

In Jamaica Nutrition Holdings, Ltd. v. United Shipping Co., Ltd., we rejected a similar argument. There the trial court had found the defendant carrier liable for failing to adequately clean out the pipes of the vessel before loading its cargo of soybean oil. The ship's previous cargo had been molasses. Prior to loading the oil, a surveyor had visually inspected the ship's pipes and tanks and determined that they were suitable for carrying the oil. After the ship reached its destination, another surveyor examined the oil and found it contaminated with molasses. Based on this evidence, the district court concluded that the defendant's failure to clean adequately the vessel's tanks, pipes, and pumps rendered the vessel unseaworthy.

On appeal, the defendant argued it should escape liability because the voyage charter party provided: "Vessel to clean tanks, lines and pumps to Charterer's surveyor's satisfaction." It contended that because the charterer's surveyor had inspected the vessel's tanks and found them suitable, the carrier's obligation to the shipper was fulfilled. However, we concluded otherwise, reasoning:

COGSA, whether applicable by its own force or by virtue of the clause paramount, imposed a nondelegable duty on [the carrier] to exercise due diligence to make the vessel fit for carriage of the cargo shipment. This duty was not abrogated by its covenant also to clean the vessel to the charterer's satisfaction. By permitting molasses residue to remain in the system, [the carrier] violated its duty.

Because the duty to exercise due diligence to ensure the seaworthiness of a vessel is nondelegable, the district court here did not reversibly err in concluding that the vessel interests failed to exercise due diligence in part because they did not test the watertightness of the hatches.

Finally, the vessel interests maintain that they had no contractual duty to rinse out the holds with fresh water. Again we remind that the vessel interests need not be contractually bound to perform a task for its omission to be a lack of due diligence. They do not dispute that before transporting the steel coils the vessel had transported coal, and before that, rock salt. They also do not deny that before loading the coils onto the ship the holds were washed out with Baltic Sea water, which has a higher salt content than fresh water. When asked whether the holds should have been washed out with fresh water before loading the steel cargo, Captain Sparks explained, "as chlorides are, even in a reduced form, so devastating to steel surfaces, . . . we always advocate that after washing out the holds, the final wash-down should be done with fresh water."

The district court's finding is supported not only by Captain Sparks's opinions but also by silver nitrate tests conducted both at Riga and Ventspils after the holds were washed out that revealed the presence of chlorides in the holds. Even after these tests proved positive for salt, the crew did not wash the holds out with fresh water. The results of the silver nitrate tests show not only that the crew should have washed the holds out with fresh water to begin with but also that the crew did not take the necessary precautions to protect the cargo even after they knew salt was in the holds. We therefore can find no error in the district court's determination that the crew's failure to wash out the holds with fresh water was a lack of due diligence.

C.

The vessel interests contend that even if the district court's finding on due diligence can be sustained, the coils became damaged due to causes for which COGSA liability is excepted.

1.

Under section 1304(2) of COGSA, "[n]either the carrier nor the ship shall be responsible for loss or damage arising or resulting from . . . [p]erils, dangers, and accidents of the sea or other navigable waters." At trial the defendants argued that a storm encountered by the M/V LAKE MARION on its transatlantic voyage constituted a "peril of the sea" and caused the saltwater to enter the holds. The ship's captain testified that he encountered very rough weather during the journey, with strong winds that occasionally reached Beaufort Scale Force 10 and, at their peak, reached force 11 to 12 for approximately two hours on March 26.

The trial judge rejected the peril of the sea defense for two reasons. First, such weather conditions were foreseeable in the North Atlantic during the late winter months. Second, no damage to the vessel resulted from the voyage, and the only conditions noted in the surveys at the discharge ports indicated preexisting damage as a result of prolonged neglect.

The vessel interests contend that it is irrelevant that the conditions encountered by the M/V LAKE MARION were foreseeable, because with their force 12 winds they were severe enough as to have been unpreventable even if foreseeable, like hurricane weather. They claim these conditions resulted in significantly more movement of the hatch covers than is normal.

The vessel interests rely upon J. Gerber Co. v. S.S. Sabine Howaldt, a case from the Second Circuit which concluded that similarly rough conditions were a peril of the sea. The shipper had chartered the vessel to carry steel products across the Atlantic. Upon unloading, the steel showed extensive signs of rust damage from exposure to seawater. The defendants claimed the peril of the sea defense. The district court found the weather conditions through which the ship sailed were not a peril of the sea, and awarded damages to the shipper. The Second Circuit reversed, finding that the conditions were sufficiently perilous.

The vessel in J. Gerber faced days of force 9 and 10 winds, and over ten hours in which the winds reached force 11 and 12. It had to heave to and remain in that state for twelve hours. These fierce winds twisted the ship's hull and caused her to roll violently and shudder and vibrate "as she was pounded and wrenched by the heavy seas." After the storm abated, the crew found damage to the ship in several areas: a piece of equipment was torn off the deck, leaving a hole; a porthole in the galley was smashed; a gangway was destroyed; two winch covers were ripped off; and the vessel had several dents.

In determining whether the storm constituted a peril of the sea, the court found of great importance "the wind velocity in terms of the Beaufort Scale." Although it explained that "[n]o exact Beaufort Scale wind force can be referred to as the dividing line which will determine those cases in which a peril of the sea is present and those, below that mark, in which it is not," it found "few cases in which the winds are force 9 or below . . . in which there has been found to have been a peril of the sea, whereas there are many where the force has been 11 or above."

It also cautioned that wind velocity is "a rough measure at best and not sufficient standing by itself." There are other indicia. Assuming a seaworthy ship, they include the "nature and extent of the damage to the ship itself, [or] whether or not the ship was buffeted by cross-seas which wrenched and wracked the hull and set up unusual stresses in it." Looking to these other considerations, the court concluded that the damage to the cargo was caused by a violent peril of the sea that, "through wrenching and twisting the vessel, set up torsions within the hull which forced up the hatch covers and admitted sea water to the holds." The vessel interests argue that we should reach the same conclusion as the J. Gerber court, because the wind force encountered by the ship in that case was in line with that faced by the LAKE MARION.

Although the conditions experienced by both vessels contain some parallels, they were not identical. In J. Gerber the ship faced force 11 to 12 winds for approximately ten hours; the LAKE MARION only encountered such winds for two hours. Captain Sparks testified that because the ship's exposure to those winds was short, the wave height likely did not build up to that normally encountered with force 11 to 12 winds. He explained, "[s]ea condition builds up slowly as the wind increases" and therefore "the wind force is way ahead of it." As a result "it'll take three to four hours to build up to what it should be." Although with force 12 winds the accompanying sea height is typically fourteen meters, Sparks opined that the highest it reached with the LAKE MARION was eleven and a half to twelve meters before the wind force started to reduce.

The vessel interests also overlook the fact that the strength of the wind is but one among several factors that should be considered in determining the applicability of the peril of the sea defense. As emphasized by J. Gerber, the extent of damage to the vessel is also important. The vessel in that case sustained a great deal of damage, while the captain of the LAKE MARION reported none. Captain Sparks confirmed that the ship was not damaged by the weather, as the only problems noted in the discharge surveys referred to preexisting damage as a result of prolonged neglect.

We, like the district court here, find telling that the ship sustained no reported damage as a result of the stormy conditions. As one court has noted,

[t]he absence of damage to the vessel itself is always an important consideration. Indeed, the courts have often required that the structural damage be substantial. . . . [I]f there was no structural damage or if the damage was limited to one fitting, such as a hatchway or a ventilator cowl, which led to damage to the cargo, courts will find with reluctance that the damage resulted from a peril of the sea.

We sustain the district court's refusal to find the rough weather encountered by the M/V LAKE MARION to have been a peril of the sea given the ship's lack of injury. We cannot conclude on this record that the noted storm, even with its force 12 winds, constituted a peril "of an extraordinary nature or aris[ing] from irresistible force or overwhelming power" which could not "be guarded against by the ordinary exertions of human skill and prudence."

2.

The vessel interests also urge another exception to COGSA liability. COGSA exempts any damage caused by "[l]atent defects not discoverable by due diligence." Defendants argued at trial that a crack found in Hold No. 1 while the vessel was docked in New Orleans, which ruined 123 coils in that hold, was a latent defect that could not have been discovered through due diligence. The trial judge rejected the contention that the fracture was a latent defect.

"A true latent defect is a flaw in the metal and is not caused by the use of the metallic object" or by "gradual deterioration." Such a defect "is one that could not be discovered by any known and customary test." The ship owner has the burden to demonstrate that the defect was not discoverable.

The vessel interests posit that since a latent defect is one not discoverable in the ordinary course of surveys or inspections, and the M/V LAKE MARION's holds were inspected during the loading process, the crack was by definition a latent defect. However, Marine surveyor Captain Rasaretnam inspected the crack and determined that it was old, and had existed in some form since crews installed a doubling plate at the fracture site. The district court concluded that the crack was an extension of an old crack, and at least part of it had been present since the doubling plate had been put in place. Moreover, Captain Sparks hypothesized that the crack was caused by gradual deterioration, not by a defect in the metal. We cannot conclude that the district court clearly erred in finding that the fracture was old and in rejecting the latent defect defense.

V.

In addition to its COGSA claims, Steel Coils asserted a general maritime negligence claim against Lake Marion, Inc.'s managing agent, Bay Ocean. The claim is that Bay Ocean, as vessel manager, hired the crew and was responsible for maintaining the vessel's condition, and that it was negligent in maintaining and testing the hatch covers, failing to repair the crack in Hold No. 1, and in washing the holds with seawater. Bay Ocean contended at trial that Steel Coils could not assert a negligence claim against it outside of COGSA.

The district court disagreed, holding Bay Ocean liable in tort for its negligence separate from the COGSA claim, and finding Bay Ocean liable for the entire amount requested by Steel Coils because Bay Ocean was not entitled to claim the $500-per-package limitation on liability found in COGSA. These are conclusions of law and we conduct a de novo review.

"One of COGSA's most important provisions limits a [vessel or] carrier's liability to five hundred dollars . . . per package unless a higher value is declared by the shipper." The term "carrier" includes "the owner or the charterer who enters into a contract of carriage with a shipper." We have held that as long as an entity is a party to the contract of carriage, it is a carrier. In Sabah Shipyard Sdn. Bhd. v. M/V HARBEL TAPPER, we stated, "[t]o determine whether a party is a COGSA carrier, we have followed COGSA's plain language, focusing on whether the party entered into a contract of carriage with a shipper. . . . [A] party is considered a carrier under COGSA if that party `executed a contract of carriage.'"

It is undisputed that Bay Ocean is not explicitly named in the applicable contract of carriage, the voyage charter between Western Bulk and Itochu. Nevertheless, Bay Ocean maintains that it was a party to the time charter between Lake Marion, Inc. and Western Bulk, and therefore Western Bulk acted as Bay Ocean's agent, as well as Lake Marion's, in entering into the voyage charter with Itochu. The district court rightly concluded that Bay Ocean was not a carrier.

The time charter states that it is between "Lake Marion, Inc. — Managers: Bay Ocean Management, Inc." and Western Bulk. However, the charter party only recites the duties and rights of the time charterer, Western Bulk, and the owner, Lake Marion, Inc. It does not bind Bay Ocean in any way. Bay Ocean's presence in the contract is simply as a signing agent of the owner, as evidenced by the signature line, which provides that Bay Ocean signed the time charter for Lake Marion Inc. "As Agents Only." Because the time charter party was solely between Western Bulk and Lake Marion, Western Bulk did not enter into the voyage charter party with Itochu on Bay Ocean's behalf, but rather only on Lake Marion's, and Bay Ocean's only status in this case is as an agent of the carrier, Lake Marion, Inc.

In Robert C. Herd Co. v. Krawill Machinery Corp., the Supreme Court clarified that agents of a carrier do not qualify for the $500-per-package limitation. The precise question presented was whether the limitation "also appl[ied] to and likewise limit[ed] the liability of a negligent stevedore." The stevedore's employees had, while loading the plaintiff's cargo onto the vessel, dropped one of the cases, which contained a press weighing nineteen tons, into the harbor. After the plaintiff filed suit against the stevedore, the stevedore asserted that its liability was limited to $500 by COGSA. In determining whether the package limitation applied, the Supreme Court first looked to the language and legislative history of COGSA to determine whether Congress intended to limit the liability of "negligent agents of a carrier." It observed:

The Act is clearly phrased. It defines the term "carrier" to include "the owner or the charterer who enters into a contract of carriage with the shipper." It imposes particularized duties and obligations upon, and grants stated immunities to, the "carrier." Respecting limitation of the amount of liability for loss of or damage to goods, it says that "neither the carrier nor the ship" shall be liable for more than $500 per package. It makes no reference whatever to stevedores or agents.

Moreover, "[t]he legislative history of the Act shows that it was lifted almost bodily from the Hague Rules of 1921," which "do not advert to stevedores or agents of a carrier," and "[t]he debates and Committee Reports in the Senate and the House upon the bill that became the Carriage of Goods by Sea Act likewise do not mention stevedores or agents." The Court concluded that nothing in the language or the legislative history of the Act either "expressly or impliedly indicates any intention of Congress to regulate stevedores or other agents of a carrier, or to limit the amount of their liability for damages caused by their negligence."

Despite this language in Herd, Bay Ocean argues that even if it is nothing more than an agent of the carrier it may avoid liability altogether on Steel Coils's separate negligence claim because COGSA is the exclusive remedy for suits for damage to cargo. However, in a similar case, Citrus Marketing Board of Israel v. J. Lauritzen A/S, the Second Circuit held that a plaintiff may sue a ship's manager in tort for damage to cargo and that COGSA does not govern such an action. The Citrus Marketing court rejected the manager's argument and the district court's holding that COGSA controlled the claim, explaining that COGSA only applied to disputes between shippers and carriers. Relying on Herd, the court concluded that COGSA did not preclude a separate action against the manager. It explained, however, that a Himalaya Clause, which extends a carrier's rights under COGSA to agents of the carrier, might apply to save the manager from liability and remanded that issue for the district court to consider at trial. In finding that COGSA did not prohibit Steel Coils's negligence claim against Bay Ocean, the trial judge relied upon Citrus Marketing.

Bay Ocean charges that the district court's ruling that Bay Ocean cannot take advantage of the $500-per-package limitation "ignores the reality of maritime commerce," because it is common for one-vessel corporations such as Lake Marion, Inc., who have no employees, to act solely through their managing agents. It also argues that this result will "allow shippers to circumvent not only the package limitation, but all of COGSA, when contracting with a vessel with a separate managing agent." However, Bay Ocean chose to separate itself from Lake Marion by binding only Lake Marion to the time charter. In doing so Bay Ocean chose that only Lake Marion would become a carrier for purposes of COGSA.

We agree with the Second Circuit that a noncarrier can be held liable in tort outside of COGSA. Steel Coils's negligence action against Bay Ocean was not subject to the COGSA package limitation.

VI.

For these reasons, we AFFIRM the district court's judgment.