Friendly, Kaufman and Anderson, Circuit Judges.
This action to recover for the loss of a cargo of soybean oil was originally tried before the late Judge Archie O. Dawson, who entered judgment in favor of the plaintiffs but limited liability.*fn1 On appeal, this court reversed and remanded because the findings of fact did not make clear apparent inconsistencies between the issue of defendant's due diligence in making the vessel seaworthy and the issue of limitation of liability*fn2 nor were those two issues discussed in the opinion at all.
At the second trial the shipowner defendant denied unseaworthiness, claimed due diligence in its effort to assure seaworthiness of the vessel when it broke ground to put to sea and claimed that any liability should be limited. In addition, as it did in the first trial the defendant relied principally upon the defense of fire, not caused by the design or neglect of the owner, as provided in the fire statute, 46 U.S.C. § 182, and in the fire exception of the Carriage of Goods by Sea Act (COGSA), 46 U.S.C. § 1304(2) (b) (1958). Judge Croake concluded that the fire had nothing to do with the loss of the vessel and cargo, that the ship was unseaworthy, that the owner had not shown due diligence in making the ship seaworthy, and he denied limitation of liability.*fn3 We affirm.
The trial court made detailed findings of fact and it is not necessary to do more than summarize them for the purpose of this opinion. The tank vessel Perama was built in 1936 for the American Oil Co. After remaining idle for almost two years in a very run-down condition, she was sold in 1956 to the defendant-appellant, who put her in the yard of the Alabama Drydock and Shipbuilding Company for reconditioning, renewals and repairs, which cost $890,000. After this work had been completed and between January 18, 1957 and March 9, 1957, she carried a cargo of gas oil from Houston, Texas to Thameshaven, England, and returned in ballast to New Orleans, Louisiana. There she went into the Todd Shipyard for repairs, most of which were to remedy some small cracks which had appeared in the transverse bulkheads. The Perama then went to Baton Rouge, Louisiana, where she loaded the cargo in suit for carriage to Genoa, Italy. She left Baton Rouge on March 15, 1957 for the Gulf of Mexico. En route she was required to anchor for two hours to repair her steering gear. When out in the Gulf at 1400 on March 17, the Perama hove to after shutting down her engines to effect repairs to the lubricating oil pump in her engine room. At about 0130 on March 18th, before the repairs had been fully completed, fire and smoke were observed in the area of the tanker's after pump room and after cofferdam, accompanied by an explosion-like sound. The Captain noted in the log that the fire was put out by 0240, although some smoke still remained. At 0315 a second explosion-like noise occurred, with considerable vibration in the stern of the ship. There were no evidences of fire or smoke coinciding with or following this sound, but a diagonal crack, three to four feet long, appeared in the hull in the engine room. The fissure ran down to the floor of the engine room and sea water poured through it and up through the floor plates. A vertical crack of about the same length also appeared in the forward bulkhead of the engine room. This bulkhead was also the after side of the port fuel tank, and through the crack a mixture of water and fuel oil also discharged into the engine room. There was also evidence that sea water was entering through a crack in the pump room. Before leaving the tanker, the Perama's Captain observed another crack in the after deck on the starboard hand through which cargo oil was leaking. The court inferred from the trim of the ship until shortly before she sank that cracks had opened up and let in sea water in the forward part of the vessel. In connection with the observed cracks no standard damage control procedures were adopted except for an effort with the pumps to keep ahead of the intake of water, which proved to be ineffective. As sea water flooded the engine room, the Captain ordered the ship abandoned. Distress calls had been sent out by radio, and a vessel which responded, sought to tow the Perama to more shoal water for grounding but at 1631 on March 18th the tanker sank in water too deep for salvage of vessel or cargo.
It is not disputed that the cargo was delivered to the carrier in good order and that subsequently it was entirely lost when the ship sank in fair weather and calm seas. Under these circumstances it is presumed that the loss was occasioned by the unseaworthiness of the Perama. Commercial Molasses Corp. v. New York Tank Barge Corp., 314 U.S. 104, 62 S. Ct. 156, 86 L. Ed. 89 (1941); South, Inc. v. Moran Towing and Transportation Co., 360 F.2d 1002 (2 Cir. 1966); Lekas & Drivas, Inc. v. Goulandris, 306 F.2d 426 (2 Cir. 1962). The defendant-appellant could therefore escape liability only by sustaining the burden of proving one of its defenses*fn4 which were that the facts brought the case within the fire exception under the COGSA or the provisions of the fire statute, or that it had exercised due diligence to make the Perama seaworthy before she broke ground and put to sea, or that it was entitled to limitation of liability.
The fire defense was rejected by both of the trial judges who had heard the case. Both found that the loss of the cargo was not caused by any fire or an explosion resulting from fire. Judge Croake found himself in substantial agreement with Judge Dawson's findings concerning the fire and its effect, though we are concerned on this appeal only with the findings and conclusions of Judge Croake.
The appellant's attacks on the findings are for the most part differing interpretations of the evidence of what those present on the vessel heard and observed and did from the time the engines of the Perama were shut down until she sank. This kind of attack resolved itself into a battle of experts who gave sharply conflicting exegeses of the events in the last twenty-seven hours in the life of the vessel. The defendant-appellant argues that a fire of unexplained origin sprang up in the after cofferdam and pump room and heated the fuel oil in adjoining tanks, causing it to vaporize and explode so violently that it opened breaches in the ship's side and let in sea water which sank her. The trial court concluded, however, that the fire did not produce such an explosion. It noted what is a commonly known phenomenon testified to by some of the expert witnesses, that the cracking of steel hull plates or the rending of a steel bulkhead may make a noise like a violent explosion. There was evidence from which the court could have found and did find that the fire was confined to the after cofferdam and pump room; that it did not generate heat of sufficient intensity to vaporize the fuel oil; that the damage to the ship's adjacent internals was not characteristic of that caused by an explosion; that no explosion took place in the fuel tanks; that the flames were subdued in approximately an hour, though the smoke continued a little longer; that the second explosive noise occurred about a half-hour after the flames of the first fire had been put out; and that no fire or smoke coincided with or followed the second noise, but cracks appeared in the hull of the ship and a bulkhead in the engine room. From these and other facts found from the evidence and stated in its opinion the trial court concluded that the fire did not cause the openings in the ship's hull which let in the sea water. We cannot say that it committed clear error in so doing. The conclusion is fully supported by findings based on competent evidence. The court was not bound to accept the interpretations of the appellant's experts as against those of the appellees. Their function was to aid the court, and it was within the province of the court to accept and use only the explanations and opinions which in its judgment were most logical and consistent with the facts it found. The appellant's claimed fire defense did not avail to meet the presumption in favor of the appellees.
The trial court found that the presumption of unseaworthiness not only remained unrebutted but that the greater persuasiveness of the plaintiffs' evidence carried "the claim of unseaworthiness beyond the area of legal presumption and into the domain of evidentiary proof." A great deal of time was consumed in the trial and large parts of the briefs submitted on this appeal were devoted to analyses of the chemistry and structure of steel plates used in various parts of the vessel, the susceptibility of the steel plates made during World War II to brittle fractures due to deficiencies of manganese, and to the improved plates used in making the repairs to the Perama, although no analysis was offered of the plates used in her when she was built in 1936. There were also conflicting views and opinions of the experts produced by both parties. Those for the plaintiffs sought to prove that the extensive replacement of wasted plates in a riveted ship by welding new plates in substitution for the old, by failing to replace a large number of wasted plates, by cropping many of the old plates to remove rivet holes at their ends, by welding approximately 13,000 new rivet heads on old stems all tended to create crack triggers which made the hull open up in numerous places and caused the vessel to sink. Those for the defendant sought to show that the new welded plates were of a post-war composition which was not subject to cracking; that the original 1936 steel of her hull was also not susceptible to cracking; that at the time she put to sea the Perama had no brittle steel in her; that in the stern where cracks were observed in the engine room and pump room there were no wasted plates, no replacement by welded plates, no welding of new rivet heads and no changes or repairs which could operate as crack triggers; and that the plaintiffs' claim that there were in the hull insufficient crack arrestors is meaningless because the lines of rivets in a riveted ship are in themselves crack arrestors.
The trial court in its opinion discussed the evidence and found facts from which it concluded that the Perama was lost because of her unseaworthy condition. There was sufficient evidence to support the findings and the conclusion logically followed. There is certainly no basis for a holding of clear error. The appellant argues that cracks in the hull could come only from brittle steel, that the evidence demonstrated that there was no brittle steel whatever in the Perama and that therefore any cracks which appeared could have been caused only by fire and explosion. But the appellant failed to prove that fire and explosion had anything to do with cracking the steel plates and the proposition that, absent some outside force as a procuring cause, no steel plates of pre-World War II riveted ships could crack is, of course, entirely untenable in view of the many cases in which cracks appeared in steel plates in such vessels. The appellant never offered proof of the chemical composition or other intrinsic characteristics of the hull plates of the Perama, but it asserts that the trial court committed reversible error in stating that the tanker's "prewar plating was presumably of pre-war sensitivity and susceptibility to brittle fracture. * * *" This amounts to no more than a criticism of the use of words. For want of a better one the court mentioned a brittle fracture as distinct from ductile crack, i.e., one resulting from bending or drawing out the steel. There was little or no evidence that the cracks observed were ductile cracks and the court was not wrong in finding, in effect, that they were non-ductile. What such cracks should be called -- whether brittle cracks or fractures or ordinary fissures -- should not be determinative of this case. Even if the court were wrong in calling them brittle cracks, it would not change the result. The undisputed facts are that after the first voyage cracks appeared in some of the transverse bulkheads and had to be repaired at the Todd Shipyard; on the day of the sinking cracks appeared in the hull and a bulkhead in the engine room and in the hull in the pump room. A crack had also been observed in a deck plate in the after part of the ship.
The essence of the trial court's decision is that the extensive replacement of plates by welding in the hull of a riveted ship, including the replacement of all of the "H" or the "wind and water" strakes, adjacent to cargo spaces, with welded plates, set up stresses within the hull between the fixed and rigid areas of welded plates and the less rigid riveted areas which resulted in cracking in various parts of the hull and the internal structures. Just when and where such stresses in the hull will produce cracks -- whether near to or distant from the place of repair -- cannot usually be determined, but the substitution of large areas in a riveted hull with welded plates is likely to cause cracks. The court, in effect, found that this is what happened in the present case. The evidence of what was observed to have happened and expert opinion support this conclusion. We find no error in it.
The appellant claims that in any event it cannot be held liable for the loss because it exercised due diligence to make the Perama seaworthy.*fn5 In seeking to fulfill its burden of proof of this defense, it produced evidence to show that it had fully complied with the recommendations of the American Bureau of Shipping to make the vessel seaworthy and that it contracted to have the work done by a reputable shipyard, Alabama Shipyard & Dry Dock Company. The appellees, on the other hand, argue that the sole function of the American Bureau of Shipping was to advise as to the work needed to make the vessel comply with the Bureau's own classification rules and to inspect to see that the work required was performed. The trial court held that the responsibility for exercising due diligence in making the ship seaworthy could not be delegated to others, that the appellant had present, and in charge of the repairs, fully authorized representatives who acted as and for the corporation,*fn6 that it offered no satisfactory or persuasive explanation for the deficiencies in the repair work which made the ship crack-prone, and that the appellant had therefore failed in its burden of proving due diligence. General Motors Corp. v. The Olancho, 115 F. Supp. 107 (S.D.N.Y.1953), affirmed on opinion below 220 F.2d 278 (2 Cir. 1955); Gilmore and Black, The Law of Admiralty (1957) § 3-27, p. 129.
There was evidence that the American Bureau of Shipping in its standards approves and accepts unlimited replacements of the plates in a riveted hull by welded plates. There was also evidence that the United States Coast Guard will approve only minor and very limited substitutions of this kind; and, had the Perama been an American flag vessel, the very extensive substitutions of welded for riveted plates made in her repairs would not have been approved by the United States Coast Guard because of the likelihood of the creation of stresses in the hull with resultant cracking of plates. That this method of repair used so extensively in the Perama was dangerous was something the appellant knew or should have known. Great Atlantic & Pacific Tea Co. v. Brasileiro, 159 F.2d 661, 665 (2 Cir. 1947).
The appellant, in describing the repair work done by the Alabama yard, said that internally the cargo tank bulkheads and framing were almost entirely renewed. After the first voyage, however, cracks appeared in the transverse bulkheads of several of her cargo tanks and had to be repaired at the Todd Shipyard at New Orleans. Although these were small, they should at least have put the appellant on notice that some stress or strain was causing cracks in newly repaired bulkheads which the appellant considered to have been of the highest quality of material and workmanship. But there is nothing to indicate that any special effort was made to investigate or determine the cause of these cracks in the newly installed steel plates. We see no reason to reverse the trial ...