ANDERSON, Circuit Judge:
The S.S. Sabine Howaldt, a small cargo vessel of a gross tonnage of 2288.43 tons, was time chartered to Contramar S/A for a voyage from Antwerp, Belgium to Wilmington, Delaware and Alexandria, Virginia. Contramar contracted to carry a quantity of steel products consigned to each of the two plaintiffs. The cargo was in good condition when loaded aboard the Sabine Howaldt at Antwerp. On arrival at the ports of destination in the United States, however, the steel showed extensive salt waiter damage from rust and pitting. In the course of her voyage across the North Atlantic the Sabine Howaldt encountered extremely heavy weather. The trial court found that as a result of shipping seas over the deck and hatches, sea water in substantial quantities penetrated the #4 hold between the coam-ing and hatch cover and to a lesser degree entered the other holds in like manner and to some extent through the ventilators.
The defendant both in the district court and on appeal has argued as its principal defense that any damage to cargo was caused solely by a peril or danger of the sea, an exception to liability provided by the Carriage of Goods by Sea Act, 46 U.S.C. § 1304(2) (c). The carrier also claims to have shown that the ship was not unseaworthy on her departure from Antwerp for “want of due diligence” on its part, or for any other reason. The district court, however, concluded that the vessel was un-seaworthy when it set out on its voyage, and that this condition was due to the negligence of the defendant. It further concluded that the winds and seas which the vessel encountered did not constitute a peril of the sea; and, even if they [584]*584did, the negligence of the defendant nullified that defense. We reverse.
It is not disputed that when the coils óf wire and other cargo were loaded aboard the Sabine Howaldt at Antwerp they were in good condition and free from sea water damage. When delivered at the ports of destination they were pitted and rusted from contact with sea water which had entered the holds— principally in #4 and some in #1.
As the plaintiffs-appellees assert, this created a prima facie case of liability against the defendant-appellant. The burden of proof then shifted to the defendant-appellant to show that the loss came within the exception of a peril or danger of the sea, § 1304 (2) (c). Commercial Molasses Corp. v. New York Tank Barge Corp., 314 U.S. 104, 109, 62 S.Ct. 156, 86 L.Ed. 89 (1941); Gilmore and Black, The Law of Admiralty § 3-43 at 162 (1957).
The Sabine Howaldt had an overall length of 306' 4"; a beam of 40' 7"; a deadweight tonnage of 3507 long tons; and a net tonnage of 1499.51 long tons. Her bridge was amidships and her en-gineroom aft. Forward of the bridge there was a hatch in the weather deck measuring 59.8' fore and aft and 17.75' athwartships. Aft of the bridge and forward of the engineroom there was a hatch in the weather deck measuring 68' fore and aft and 17.75' athwartships. Each of these hatches served two holds separated by bulkheads. Holds #1 and #2 were served by the forward hatch, and holds #3 and #4 were served by the afterhatch. Both of these hatches were covered by MacGregor hatch covers.1 The ship’s rail on either hand of the forward deck area was a solid bulwark, and on both sides of the after deck there were open pipe railings. The after deck, however, was 3' 3" higher than the forward deck.
The Sabine Howaldt was seven years old in 1965 and was rated in the highest class in Germanischer Lloyd, which had surveyed her in April, 1965. She was reported well-found in all respects, and her International Load Line Certificate authorized her use on North Atlantic routes during the winter. As the loaded vessel made her departure from Antwerp in fresh water her draft was 19' 6" forward and 20'8" aft. She was not down to her winter marks.
The ship’s log records that during the late evening of the second day of the voyage, December 16, the wind freshened considerably and by midnight of December 17th had reached a force of 9 on the Beaufort Scale or 41-47 knots. On December 18th at 0900 the wind force was 10 or 48-55 knots, and continued at that strength until 1700 when it began to abate. During this period there were high seas and the ship pitched and rolled. Waves broke over the whole vessel and it was “very heavily stressed in its seams.” On December 20th the wind force again increased and reached a force of 9/10 in hurricane-like squalls but abated by 2400 and on December 21st did not exceed force 7/8 or about 34 knots. At 2400 on December 22nd the wind again increased to force 9/10 or about 48 knots. By 0200 on December 23rd it went up to force 11 with gusts at force 12 in “hurricane-like rain squalls.” By 0500, due to the criss-cross running swells and high breaking seas, the ship was badly strained in her seams and sea water was breaking over forecastle deck, [585]*585hatches and upper works. It was necessary for the vessel to heave to and she so remained for 12 hours. At 0900 the force 11 wind with gusts in squalls of force 12 steadied at a constant hurricane force of 11/12 or about 63 knots, which continued to a time between 1200 and 1300.2 The hull of the Sabine Howaldt [586]*586was twisted and strained in the turbulent cross seas; she rolled from 25°-30°; waves constantly broke over her; and she shuddered and vibrated as she was pounded and wrenched by the heavy seas. At about 1700 the wind and sea diminished to force 9/8, and by 2000 to 7/8.
Subsequently it was discovered that during this period of hurricane, the pedestal holding the master switch which controlled the turning of the capstan at the stern was torn loose leaving a hole in the deck; a port hole in the galley was smashed; the cat-walk or gangway from the amidships housing aft over the hatches and the well-deck to the poop was destroyed when it was torn loose and landed against a ventilator, which it dented. The winch covers from two winches were ripped off and disappeared. Dents appeared in several parts of the ship’s superstructure and fixtures. While the damage was substantial, it is likely that it would have been more extensive had the ship not been only about seven years old and therefore relatively new.
The trial court found: “39. The character of the seas and weather recorded in the log book do not constitute a ‘peril of the sea.’ ” This factual conclusion appears to have been predicated, in substantial part, on two other findings which are: “37. The sea and weather conditions entered in the ship’s log are less severe than those described by the master in his deposition;” and “35. The log book entries show a wind force of 10 on the Beaufort Scale during the various watches on five days of an eighteen day voyage. On one day only, during the morning watch, viz., December 23, the log entry states that the ship passed ‘through hurricane-like rain squalls.’ The ship’s chief officer testified that the winds ‘were gusts and not hurricanes.’ ” The district court tended to dismiss the testimony of the ship’s captain as exaggerated. But both his testimony and that of the chief officer were of importance as the stories of the only eyewitnesses to the weather conditions December 17-23 and their effect on the ship, and both accounts should have been carefully considered. The captain was German and his deposition discloses many instances of language difficulty, both when he spoke through an interpreter and when he testified directly. It appears, however, that he sought to answer questions by both the plaintiffs and defendant fully and accurately and was not evasive or uncooperative. The record discloses nothing to justify the conclusion that his description of the hurricane and its effect on the ship was exaggerated; on the contrary, his testimony was consistent with the entries in the log. As the evidence given by the captain and chief officer were presented in depositions, this court, on review, is in as good a position as the trial court to examine, interpret and draw inferences from their testimony. Severi v. Seneca Coal & Iron Corp., 381 F.2d 482, 488 (2 Cir. 1967); Dempster Bros. Inc. v. Buffalo Metal Container Corp., 352 F.2d 420, 423 (2 Cir. 1965), cert. denied, 384 U.S. 940, 86 S.Ct. 1458, 16 L.Ed.2d 539 (1966); Iravani Mottaghi v. Barkey Importing Co., 244 F.2d 238, 248 (2 Cir.) cert. denied, 354 U.S. 939, 77 S.Ct. 1402, 1 L.Ed.2d 1538 (1957).
One likely reason why the trial court assumed the captain was overstating the facts relating to the sea and weather was the court’s misunderstanding and misinterpretation of the log which resulted in its finding that the Sabine Howaldt did not encounter wind and sea conditions of sufficient severity to create a peril of the sea. Apparently the district court did not read from the log a continuous, sustained Beaufort Scale wind force in excess of 10 except for occasional gusts in hurricane-like rain squalls which, it found, only occurred during the morning watch (0400-0800) on December 23rd. The log, however, does not record any wind, sustained or in gusts, as low as force 10 between 0200 and 1500 on that day. The entries for those hours show no force below 11, which the trial court’s finding never mentions; and from 0500 to between 1200 and 1300 the log recorded [587]*587force 11(12) or 11/12 with the notation that the ship was “hove to in hurricane.” This is a sustained wind force of 62-63 knots or 73 land miles per hour. Both the captain and chief officer testified the ship encountered winds of hurricane force. As above quoted from the finding, however, the trial court found that the chief officer testified that the winds “were gusts and not hurricanes.” The evidence does not support such a finding. The quoted statement was made by the chief officer, not about the weather on December 23, 1965, in which the Sabine Howaldt was involved, but about the chief officer’s past experiences in hurricane weather.3
The above quoted findings, numbers 35, 37 and 39, were not supported by the evidence and were erroneously made. Other findings which depend upon these particular findings for their assumptions that the ship did not encounter any sustained winds of force 11 or more are in this regard also erroneous.
The nature and severity of the storm as described in the log and in the testimony of the captain and chief officer, are strongly corroborated by the account of the storm in the Mariner’s Weather Log, Volume 10, Number 3, May 1966, a publication of the United States Department of Commerce, a certified copy of which is in evidence.4 From this it appears that from a point about 500 nautical miles east of Cape Hatteras, the disturbance moved northeast at a rate of a little over 41.6 knots so that at 0000 on December 23 its center was 45.5° N. and 44.6° W., which was the position of the S.S. American Clipper which reported the winds at 80 knots. At 0100, when the center of the storm was about 200 nautical miles northwest of the United States Coast Guard Cutter Yakutat, that station ship reported the winds at 78 knots. At 0500 when the Sabine Howaldt hove to she was about 350 nautical miles from the center of the disturbance. She was forced off of her course and her normal daily run of over 200 nautical miles was [588]*588reduced to 96. Meanwhile the course of the storm had veered from northeast to east so that the distance between the disturbance and the Sabine Howaldt lessened over the following seven hours. At 1200, therefore, the vessel was at 45° 40' N. and 33° 11' W. and the storm, even assuming its course was slightly north of east with its target Ireland and the Midlands of England, would have been centered at approximately 48° 35' N. and 34° 14' W., or on a conservative estimate between 150 and 200 nautical miles from the Sabine Howaldt. In this connection it must be borne in mind that the disturbance had a radius of 600-800 nautical miles in the southern quadrant in which the Sabine Howaldt was located, and it is entirely consistent with the report in the Mariner’s Log that she recorded winds of hurricane force. There is, moreover, no evidence to the contrary.
[587]*587“A developing storm centered about 500 mi. east of Cape Hatteras early on the 22d moved rapidly northeastward and deepened. By the 23d central pressure was down to 975 mb. as the LOAV passed about 300 mi. east of Newfoundland, then curved toward the east, passing over the British Isles on the 24th. From the 22d through the 24th gale, whole gale, and occasional hurricane force winds extended 600-800 mi. south and west of the center. At 0000 on the 23d the American Clipper, near the storm center, reported 80-kt. winds at 45.5° N„ 46.6° AY., while the USCGC Yakutat, on Ocean Station “D,” recorded 78 kt. at 0100. Six other ships, the American Resolute, Attleboro Victory, Augsburg, Exiria, Keystone State, and Teniers all reported 55- to 60-kt. winds on the 23d or 24th in association with this relatively small but tightly organized storm.” (page 89).
[588]*588This court, therefore, concludes that the findings of the trial court relating to the nature and severity of the disturbances of the wind and sea encountered by the Sabine Howaldt and to its factual conclusions that they were not “perils of the sea” were clearly erroneous as they unavoidably leave us with the “definite and firm conviction that a mistake has been committed.” United States v. United States Gypsum Co., 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746 (1948); Iravani Mottaghi v. Barkey Importing Co., supra, 244 F.2d at 248.
The district court also concluded as a matter of law that “the character of the seas, winds and weather encountered by the S.S. Sabine Howaldt during the voyage in suit did not constitute a peril of the sea.” With this determination we disagree.
As plaintiffs’ cargo when loaded aboard at Antwerp was, barring some atmospheric damage, in good condition and was unloaded in damaged condition at the ports of destination, the plaintiffs established a prima facie case of liability against the carrier defendant. The defendant asserted exoneration under exception 2(c) of § 1304 of Title 46 U.S.C. on the ground that any loss sustained was caused by perils and dangers of the sea, the generally accepted definition of which is “those perils which are peculiar to the sea, and which are of an extraordinary nature or arise from irresistible force or overwhelming power, and which cannot be guarded against by the ordinary exertions of human skill and prudence.” The Giulia, 218 F. 744, 746 (2 Cir. 1914). See, Allen N. Spooner & Son, Inc. v. Connecticut Fire Insurance Co., 314 F.2d 753, 757 n. 3 (2 Cir.) cert. denied 375 U.S. 819, 84 S.Ct. 56, 11 L.Ed.2d 54 (1963); R. T. Jones Lumber Co. v. Roen Steamship Co., 270 F.2d 456, 458 (2 Cir. 1959). A similar definition phrased a little differently is, “a fortuitous action of the elements at sea of such force as to overcome the strength of a well found ship or the usual precautions of good seamanship.” Gilmore & Black, The Law of Admiralty § 3-32 at 140 (1957).
While in general the rule is that the carrier, to be exonerated, must prove that it comes within one of the exceptions 2(a-p) of § 1304, once it has done so, the burden rests upon the shipper to show that there were, at least, concurrent causes of loss in the fault and neglect of the carrier, unless it is one of the kinds of negligence specifically excluded, e. g. 2(a). If the shipper does so, then the carrier has the practically insuperable burden of proving the portion of the loss caused by the particular exception invoked and the portion caused by its negligence. If the carrier fails to do so, it loses all exoneration and must bear full liability for the loss. See Lekas & Drivas, Inc. v. Goulandris, 306 F.2d 426, 432 (2 Cir. 1962). Nevertheless, these usual rules governing what is to be proved by whom are altered in cases in which the carrier seeks exoneration under 2(c), “the perils of the seas” exception, because of the definition of that exception. As stated in Gilmore & Black, supra, p. 140, “ * * * in a sense, the absence of negligence as a concurring cause may be said to enter into the very definition of a sea peril, so that, in order to establish an exception under this [589]*589clause, the ship would have to establish freedom from negligence.”
The district court not only found that the character and nature of the winds and seas were not sufficiently severe to constitute a peril of the sea in fact, but it also found that the Sabine Howaldt was unseaworthy due to the neglect of the defendant carrier. It concluded that the defendant was negligent in permitting the ship to proceed on the voyage with defective hatch covers without tarpaulins over them and also because its ventilators were insufficiently protected. In its findings the district court also found the vessel unseaworthy in proceeding on the North Atlantic voyage with such hatch covers when the ship had a freeboard of only one to two feet.
With regard to the MacGregor hatch covers and the tightness of their fastening to the hatch coamings, the defendant produced evidence of the chief officer who had, with another officer, inspected the hatches after they had been battened down just prior to departure. Prior to this the surveyor for the charterer had inspected the hatches and the covers when the hatches were still open. He, therefore, did not check the juncture of the knife-edge of the coaming with the rubber gaskets of the cover, but he found no dents or bending or other damage to the edges of the coamings nor was there any staining of the insides of the coam-ings to indicate there had been leakage. He found the ship seaworthy. The chief officer made an entry in the log that the MacGregor hatch covers were “seaworthy, closed and wedged.” On December 17, just after 0800 when according to the log the wind force was 7 to 8 and when the ship pitched and rolled heavily in high seas and swells, and water had been shipped over deck and hatches, the chief officer and another inspected the hatches and cargo holds and found them in order and no sign of leakage.
On arrival at Wilmington, Delaware, on January 3, 1966 the chief officer examined the hatches and found no damage to the hatches, the hatch covers or their rubber gaskets — all were in good condition. A surveyor for the shipowner checked the hatches the following day and found no damage to the hatches, their coamings or their MacGregor hatch covers and rubber gaskets. He was accompanied by the surveyor for the plaintiffs, who did not call him as a witness. There was no evidence in the case that at any relevant time anything was wrong with the hatches or their MacGregor covers, or that they were at any time damaged or defective. No alterations, changes or repairs were made to the hatches or covers after arrival and before the ship’s next voyage from Santo Domingo to Norway in which she encountered heavy seas and shipped green water over decks and hatches but no leaks occurred. All the evidence showed they were in good condition both before and after the voyage. The trial court made no finding of anything which the defendant did to the hatches or their covers which it should not have done or, with the exception of the use of tarpaulins, discussed infra, omitted to do anything with respect to them which it should have done; nor is there any finding as to any specific condition of the hatches or their covers or the manner in which they were used and operated which rendered the ship unseaworthy.
It is undisputed that in the course of the voyage the hatch covers leaked and admitted sea water into #4 hold and a lesser amount into #1 hold. The trial court concluded from the fact that sea water entered the holds through the hatch covers, that they must have been defective when the ship left Antwerp and that the carrier was negligent in putting to sea with them in such a condition. It said, in effect, that because damage to cargo occurred through the entry of sea water between the hatch covers and coaming, therefore the ship must have been negligent. But this only goes back to the prima facie case with which the plaintiffs started. It is clear that the district court so held because it had already decided that the weather encountered by the Sabine Howaldt did not carry sufficient force to substantiate a [590]*590claim of perils of the sea regardless of other attendant circumstances.
The trial court did, however, make a finding which, in the abstract, implied some support for the opinion of several of the defendant’s experts, which was that the hull of the ship was wrenched and twisted momentarily, particularly by violent cross seas, during its rolling and pitching, setting up unusual stresses in the vessel and resulting in a torsion in the hull which, for a matter of seconds, deformed the rectangular opening of the hatch and forced the hatch cover and its gaskets up from the knife edge of tht coaming, permitting the entry of the sea water which was pouring over the decks and hatches.
In its finding #31 the district court said:
“MacGregor hatch covers, if properly maintained, will not, under almost all forseeable circumstances, raise and lower above the coaming. They should prevent the entrance of sea water under almost all foreseeable sea and wind conditions. However, no suitable means have yet been found to completely seal a hatch so as to provide absolute insurance against the entry of sea water under every possible combination of circumstances. Conceivably, an extremely severe storm of prolonged duration could so affect a vessel having the architectural characteristics of the SS Sabine Howaldt as to place so severe a strain upon properly maintained MacGregor hatch covers that the jerking, shaking and vibration of the vessel, coupled with continuous pounding by and submersion in green seas, could overcome such hatch covers to the point of leakage. So severe a state of affairs did not, however, occur during the voyage in question.”
The final sentence reflects the trial court’s conclusion that the ship was not involved in a severe storm and that the wind never exceeded force 10 except for momentary gusts of greater but not specified force during the morning watch, 0400 to 0800. As above stated, this was plain error because the uncon-tradicted evidence showed that the ship encountered extremely severe disturbances of wind and sea, particularly on December 23, which required the ship to heave to for twelve hours.
The district court also found that the ship was negligent in not covering its hatches with tarpaulins. There was no evidence that there was a customary or usual standard in the exercise of good seamanship that called for the use of canvas tarpaulins over MacGregor hatch covers. It was quite apparent that the customary practice of most steamship lines was not to use tarpaulins over such hatch covers. It was undisputed that the MacGregor hatch covers had not leaked before the 17th of December nor had they on subsequent voyages though no changes or repairs had been made to them. The expert on whose testimony the plaintiffs rely for their claim that it was negligent not to have put tarpaulins over the hatches of Sabine Howaldt on leaving Antwerp, testified that a ship was all right to sail without tarpaulins over the hatch covers where the ship had had no problems with the MacGregor covers. The Sabine Howaldt had had no such problems.
The expert’s opinion was not related to the particular wind and wave conditions to which the vessel was being subjected when the covers were forced up to admit sea water. He had no knowledge of the Sabine Howaldt and her hatches. His expert conclusion was based upon hearsay reports of other ships in the Pacific Ocean in undisclosed circumstances. He gave the opinion that a ship fitted with MacGregor hatch covers but without supplemental tarpaulin covers was not unseaworthy. The other expert testifying for the plaintiffs on the subject of tarpaulins over MacGregor hatch covers said that anyone who put a tarpaulin cover over a properly maintained MacGregor hatch cover would be a fool.
As there was no evidence in the case that the MacGregor hatch covers on the Sabine Howaldt were not properly maintained and as there was substantial, [591]*591uncontradicted evidence that they were, it was plain error to hold there was negligence in regard to a failure to cover the hatch covers with tarpaulins. Tarpaulins which were fastened as covers over winches and over ventilators on the Sabine Howaldt were blown to shreds.
The trial court also found the defendant negligent because water had entered the holds through the lower ventilators. There is, however, no evidence in the case that any water entered the holds through the ventilators. As the plaintiffs based their assertion that such evidence was brought out by their counsel in his cross-examination of the ship’s captain, it was necessary for this court to review again the entire deposition of the captain and to interpret it and draw such inferences from it as were warranted. At no time while the captain was recounting the difficulties of the voyage and the entry of sea water into the holds did he state or imply that sea water entered the vessel’s holds otherwise than through the hatches. The water damaged cargo was directly under the hatches and not near the ventilators. The captain pointed out that the surveyors at the ports of destination found no staining of the insides of the ventilators and found nothing to indicate that water had entered the hatches through the ventilators. On cross-examination the captain testified that the only two possible ways water could have entered #4 hold were through the hatches or down the ventilators.5 The captain did not testify that water had actually entered through the ventilators. Asked again if some water possibly could enter through the ventilators, the captain answered, “Possibly, yes.” Plaintiffs’ counsel then phrased a question which assumed that the captain had said that the water had entered through the ventilators.6 It is not sur[592]*592prising that with his language difficulty the captain missed the subtle change and answered as if it were not only possible that water entered the ventilators but that it had actually done so. As this court is in as good a position as the trial court to interpret the deposition and draw the inferences therein implied, it concludes that under the circumstances it must overrule the trial court’s construction and interpretation of the deposition. On reviewing the captain’s testimony as a whole we are satisfied that he never testified to anything more than that it was possible that sea water entered the ventilators. Had he testified that water had actually so entered the holds or that it was likely that it had, there would have been something to support the trial court’s finding. A mere possibility, however, is not sufficient and the finding must be disregarded as not supported by the evidence.
Even if this were not so, it could not be used as a basis for finding the ship unseaworthy for lack of due diligence. The ventilators were constructed to prevent water going down them in ordinary weather and in anticipation of especially rough weather, the captain, at the beginning of the voyage, had two tarpaulins securely made fast to the ventilator heads. This satisfied the usually accepted practices of good seamanship, and there is no evidence to indicate a lack of due diligence.
The only remaining charge of negligence against the ship has to do with her putting to sea with what the district court found to be a freeboard of one to two feet. Relating to the matter of free-board are two findings of fact which are utterly contradictory. In one paragraph the trial court found that the vessel was not down to her winter marks; and in another it found the vessel’s free-board was one to two feet. Both of these statements cannot be true. The winter marks were 2.24 meters or 7 feet 4 inches down from the upper edge of the quarter-deck stringer plate on the [593]*593ship’s side. The forward deck was lower than the quarter-deck by 3 feet 3 inches, according to the stipulation of the parties. As the vessel was loaded, the water line was below her winter marks and the freeboard could not have been as little as one to two feet as the court found.
The trial court in finding this extremely low freeboard relied upon the captain’s testimony to that effect. It is patently erroneous because it is contrary to the plans of the ship, its load line certificate and the mathematics of subtracting the mean draft of the vessel on departure, found by the trial court to be 20' 1", from the distance from the bottom of the keel to the top of the quarter-deck stringer, which is 26’ 10" or a result of 6’ 9". A fair inference is that the captain, as one of his language difficulties, was unable to convert meters to feet and inches correctly. The chief officer, who was apparently more conversant with the English language, said that the captain was mistaken about the one to two feet freeboard because it was impossible. The exhibits, however, disclose what the truth of the matter was.
The vessel was designed to run, and was accustomed to run, in rough seas with her decks awash. On this voyage she had so operated prior to 0800 on December 17th when the chief officer and another examined the hatches and holds and found there had been no intake of sea water. By the time the severe weather struck the ship she was in sea water, the buoyancy factor of which increased the freeboard by about five inches. Also the vessel had used up a sizeable portion of bunker and other consumable stores which further lightened her. We are satisfied that the defendant showed there was no negligence of the ship in putting to sea with her free-board as it was.
The plaintiffs’ case consisted of the opinions of experts about MacGregor hatch covers and attacks on the deposition testimony offered by the defendant. The experts extolled the cover’s watertight characteristics under all weather conditions. They cited hundreds of instances in which the Mac-Gregor covers had leaked on other ships and in other circumstances, but they testified that every such instance had come about because of damage to the covers or coamings by stevedores or others and there was evidence of faulty maintenance. From this they formed the opinion that the leakage of sea water into the holds of the Sabine Howaldt, through hatches which they had never inspected, was caused by faulty maintenance. This was incompetent evidence to prove unseaworthiness of the Sabine Howaldt because of a lack of due diligence by the defendant. The plaintiffs produced no evidence of defects in the coamings or the gaskets or other parts of the hatch to contradict the testimony of all witnesses who had examined them at the beginning and end of the voyage. The most important witness available to them on the subject of damaged hatches was their own surveyor, Captain Charles Bryant, who surveyed the Sabine Ho-waldt shortly after her arrival at Wilmington, Delaware at the same time she was checked by Captain Quistgaard, the defendant’s surveyor. The defendant is entitled to. the inference that Captain Bryant found the hatches and their covers in good condition, as did Captain Quistgaard, because the plaintiffs did not call on Captain Bryant to testify and gave no adequate reason for not producing him.
Plaintiffs’ counsel also asked hypothetical questions of some of the experts. These questions were not the same in each instance and were an effort to relate the expert opinion to the Sabine Howaldt, concerning which the experts had no first hand knowledge of their own nor had the plaintiffs offered any evidence of their own from which they could reasonably claim to have proven the assumed facts. There were presented to the expert witnesses selections which were taken from the defendant’s exhibits and depositions [594]*594(which by itself is, of course, permissible) but there were added some of plaintiffs’ own factual claims and interpretations. The result was that, because the assumption of material facts propounded by plaintiffs’ counsel either included statements which, as heretofore discussed, were simply not so or omitted highly relevant facts which could be fairly claimed to have been proved, the questions were misleading and unfair, and the answers were therefore meaningless on the issue of negligence as it concerned the Sabine Howaldt on the voyage in question. II Wigmore, Evidence § 680 (3rd ed. 1940).
A review of the entire record before this court makes it apparent that there was no relevant or competent evidence of negligence on the part of the defendant and that, on the contrary, the defendant proved that the ship was not unseaworthy because of a lack of due diligence. Although several of the trial court’s findings of fact, which it relied upon to support its legal conclusion of negligence, have been held wanting because not supported by the evidence and were, therefore, clear error, its determination of the issue of negligence does not fall within the clearly erroneous rule, 52(a), F.R.C.P. Serra, Inc. v. SS. Francesco C., 379 F.2d 540, 541 (2 Cir. 1967); Mamiye Bros. v. Barber Steamship Lines, Inc., 360 F.2d 774 (2 Cir.), cert. denied, 385 U.S. 835, 87 S.Ct. 80, 17 L.Ed.2d 70 (1966).
There remains the issue whether or not as a matter of law the action of the wind and sea upon the vessel was sufficiently violent to come within the exception 2(c) of § 1304 as a peril or danger of the sea. The district court decided that it was not and cited in support of this conclusion two United States Supreme Court cases and seven prior decisions of this court. All of them are readily distinguishable and none supports the lower court’s decision. The Propeller Niagara v. Cordes, 62 U.S. (21 Howard) 7, 16 L.Ed. 41 (1858), concerned stranding of a steam vessel caused by its arbitrary and unnecessary change of course when she was proceeding without difficulty on her usual course in a fresh gale which increased to a heavy (whole) gale, i. e., a wind of force 9. The issues in The Delaware, 81 U.S. (14 Wall.) 579, 598-599, 20 L.Ed. 779 (1872), had to do with the propriety of carrying a portion of a cargo of pig iron as deck cargo, which was jettisoned in bad weather, and the admissibility of parole evidence to vary terms of the bill of lading. The Rappahannock, 184 F. 291, 294 (2 Cir. 1911), suffered damage to a pipe which leaked. The vessel went through three gales, one on each of Lakes Superior, Huron and Erie. None of the witnesses testified the winds were unusual or extraordinary. The captain and mate estimated them at 60 land miles per hour and described them as “gales” the highest velocity of which is 54 land miles per hour; and is force 9 on the Beaufort Scale; 60 land miles per hour would be force 10. The court held there was not sufficient proof to show whether the pipe, weakened by age and not inspected for over a year gave way under ordinary and expected strains or was a seaworthy pipe damaged by an extraordinary strain. In the ease of The Rosalia, 264 F. 285 (2 Cir. 1920), the court found the vessel liable because the cargo was damaged by sea water on its arrival. Although the ship had had a rough voyage, it did not meet with any great or destructive storm. The leakage through the hatch was ascribed to the fact that the tarpaulins over the hatch came loose but no explanation of this was offered. See comment on broad language in the opinion in this case by L. Hand, J., in Philippine Sugar Centrals Agency v. Kokusai Kisen Kabushiki Kaisha (The Naples Maru), 106 F.2d 32 (2 Cir. 1939), infra. In The Rosalie Hull, 4 F.2d 985, 987 (2 Cir. 1925), the seas, cross-seas and winds encountered by the schooner resemble those in the present case, though the winds in Rosalie Hull were 60-70 land miles per hour or force 10/11 rather than 11/12, but the vessel was a [595]*595wooden schooner. The court found the vessel seaworthy on leaving port, and she was exonerated because the damage was caused by a peril of the sea.
In the case of The Edith, 10 F.2d 684, 686 (2 Cir. 1926), there was a sharp discrepancy between the captain’s estimate that the wind was “at 75 [land miles per hour], or about 10 on the Beaufort scale” (which he should have called force 11 or a storm), and the ship’s log which showed the wind accelerating from a “moderate breeze” to a “strong northeast breeze.” The maximum wind velocity was therefore force 6 or 31 land miles per hour. The tarpaulin hatch cover became loose and blew back so that the hatch was partly opened for about 15 minutes when repairs were completed. The court disbelieved the witness and relied upon the log entries from which it concluded there was no peril of the sea. In the Edmond Weil, Inc. v. American West African Line, 147 F.2d 363 (2 Cir. 1945), the ship had taken on deck cargo of empty ammonia cylinders and mahogany roots. In a rough sea with strong or .whole gale winds, force 9-10, the insecurely lashed tubes came adrift and careened about the after deck. Some of them struck and broke off conduits or gooseneck vents which protruded from the deck. The openings thus made let in sea water and damage to cargo resulted. The court held that the cylinders should have been made fast against all but the most unexpected and catastrophic storms, which was not done; and it in effect found the storm not that severe, at least the wind was only at force 9 when the cylinders broke loose. The highest velocity of the wind was force 9-10, which is 55 to 63 land miles per hour, as compared with winds at 56 to 63 knots or 64-73 land miles per hour met by the Sabine Howaldt. The court held that whether or not the storm reached catastrophic proportions the day after the first cylinders broke loose was of no relevance as the remaining cylinders would have followed even if the wind did not exceed force 9 or 54 land miles per hour. Under these circumstances the carrier could not claim exoneration under the exception of perils of the sea.
Judge L. Hand, who wrote the opinion in the Edmond Weil case, had, a few years earlier, authored the opinion, not cited by the district court, in Philippine Sugar Centrals Agency v. Kokusai Risen Kabushiki Raisha (The Naples Maru), supra. In that case the steam vessel, found to be seaworthy, ran into heavy seas and winds between 50 and 60 land miles per hour or 44 and 52 knots respectively. This translates into Beaufort Scale forces 9-10, a strong to whole gale. The vessel was blown off her course three times and made runs of eighty, sixty and forty-four nautical miles, respectively, in the three days of heavy weather instead of her normal 230 miles. A lifeboat was crushed and a good deal of the steel superstructure was twisted, broken or carried away. The tarpaulins over two of the hatches were torn or got loose and the iron hatch coamings were broken so that sea water entered the holds. The court held that the storm encountered by the Naples Maru was a “peril of the sea” and the carrier was entitled to exoneration. The court said :
“True, it [the storm] was no more than was to be expected in those waters at that time; but in some waters at some seasons, even hurricanes are not infrequent. Although this was not a hurricane, it was bad enough to damage the gear and superstructure of a seaworthy ship. The phrase, ‘perils of the sea', has at times been treated as though its meaning were esoteric: Judge Hough’s vivid language in The Rosalia, 2 Cir., 264 F. 285, 288, has perhaps given currency to the notion. That meant nothing more, however, than that the weather encountered must be too much for a well-found vessel to withstand. Duche v. [Thomas and John] Brocklebank, 2 Cir., 40 F.2d 418. The standard of seaworthiness, like so many other legal standards, must always be uncertain, for the law cannot fix in advance those precautions in hull and gear which will be [596]*596necessary to meet the manifold dangers of the sea. That Judge Hough meant no more than this in The Rosa-lia, supra, is shown by his reference to the definition ,in The Warren Adams, 2 Cir., 74 F. 413, 415, as the equivalent of what he said. That definition was as follows: ‘That term may be defined as denoting “all marine casualties resulting from the violent action of the elements, as distinguished from their natural, silent influence.” ’ It would be too much to hope that The Rosalia, supra, will not continue to be cited for more than this, but it would be gratifying if it were not.” 106 F.2d at 34-35.
The standard of seaworthiness must remain uncertain because of the imponderables of the forces exerted upon a ship by the winds and seas. Ship design and construction over many centuries of experience have evolved to meet the dangers inherent in violent winds and tempestuous seas. But for the purpose of deciding whether or not they constitute perils of the sea for a particular vessel for the purpose of the statutory exception there is the question of how violent and how tempestuous. These are matters of degree and not amenable to precise definition. Moreover, there are variations in kinds both in the winds and seas: whether the winds are steady for a number of hours from one direction or are in frequent gusts lasting less than a minute on each occasion or are cyclonic and shifting. The seas may proceed in a fairly regular procession of coamers or consist of turbulent cross seas which can be very destructive. In a constricted sea-way the drift and set of currents and their relationship to the velocity and direction of the wind are matters to be considered. As there is a direct relationship between the wind velocity and the build up and size and shape of waves, a very important measure considered on perils of the sea issue is the wind velocity in terms of the Beaufort Scale. Most ocean going vessels today are equipped with anemometers or other wind gauging device which will function up to about 80 knots. No 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. There are, however, few cases in which the winds are force 9 or below (i. e., 54 land miles per hour or 47 knots) 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. This is still, however, a rough measure at best and not sufficient standing by itself. Other indicia are, assuming a seaworthy ship, the nature and extent of the damage to the ship itself, whether or not the ship was buffeted by cross-seas which wrenched and wracked the hull and set up unusual stresses in it and like factors. While the seaworthiness of a ship presupposes that she is designed, built and equipped to stand up under reasonably expectable conditions this means no more than the usual bad weather which is normal for a particular sea area at a particular time. It does not, however, include an unusual combination of the destructive forces of wind and sea which a skilled and experienced ship’s master would not expect and which the ship encountered as a stroke of bad luck. Hurricane force winds and turbulent cross-seas generating unpredictable strains and pressures on a ship’s hull are an example.7
[597]*597We are satisfied that the Sabine Howaldt was a seaworthy vessel when she left Antwerp on Deeeinber 15, 1965; she had been certified in the highest class of Germanischer Lloyd and the latest inspection by that classification society was April 27, 1965. Throughout the voyage she was operated in a good and seamanlike manner. There was no negligence on the part of the carrier. The damage to the cargo was caused by violence of the wind and sea and particularly by the resulting cross-seas which, 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 judgment of the district court is reversed and judgment is ordered to be entered in favor of the defendant with its costs.