Waterman, Friendly and Kaufman, Circuit Judges.
These libels, consolidated for trial before Judge Wyatt in the District Court for the Southern District of New York, see 241 F. Supp. 99 (1965), were brought to recover for damage to cargo on Pier 5, Bush Terminal, Brooklyn. The cause of the damage was a flooding of the pier due to storm surge and wave action created by Hurricane Donna which struck New York harbor in the early afternoon of Monday, September 12, 1960. The pier was 10 foot above mean low water, as against 9 foot required by New York City, but the unusually high level of the water covered the floor to a considerable height.*fn1 Some of the cargo was inbound, having been unloaded on or before Friday, September 9, from the M/V Toreador, the M/V Tatra and the M/V Turandot; other cargo was outbound, having been delivered on or before September 9 for shipment on the Turandot and the Tatra. The respondent shipowners impleaded Atlantic Stevedoring Co., the pier operator, which carried the burden of the defense. After trial Judge Wyatt dismissed the libels on the ground that the loss was attributable to an "Act of God" within the meaning of the Carriage of Goods by Sea Act, 46 U.S.C. § 1304(2) (d).*fn2 In a thorough opinion, he recognized that under the statute the carriers and pier operator were liable for damage to the cargo caused by their negligence and had the burden of showing either freedom from negligence or that the loss could not have been prevented by the exercise of reasonable care. He held, however, that in view of the Weather Bureau estimates as to the probable course of the hurricane they had not been negligent. On appeal no one challenges his analysis of the governing legal principles; the attack by the cargo owners is on the correctness of his conclusion that the burden of negating negligence was satisfied.
We are confronted at the outset by appellees' contention that the only issue before us is whether the district court's conclusion of lack of negligence was clearly erroneous. Appellees recognize the many decisions of this court which hold that, with respect to negligence as distinguished from the evidentiary facts on which it is based, a judge's determination is not entitled to the benefit of the "unless clearly erroneous" rule either in admiralty or in actions governed by F.R.Civ.P. 52; they suggest, however, that most of these cases antedated McAllister v. United States, 348 U.S. 19, 75 S. Ct. 6, 99 L. Ed. 20 (1954), and that in any event they cannot stand in the face of that decision.
The attempt to whittle down the precedents in this court must fail. In Romero v. Garcia & Diaz, Inc., 286 F.2d 347, 355-356, cert. denied, 365 U.S. 869, 81 S. Ct. 905, 5 L. Ed. 2d 860 (1961), where we adhered to our long-standing rule, we cited three earlier post-McAllister decisions in which this court, speaking through eminent judges, had done precisely that. Dale v. Rosenfeld, 229 F.2d 855, 858 (1956) (Swan, J.,); New York, N.H. & H.R.R. v. Gray, 240 F.2d 460, 465 (Frank, J.), cert. denied, 353 U.S. 966, 77 S. Ct. 1050, 1 L. Ed. 2d 915 (1957); and Verbeeck v. Black Diamond S.S. Corp., 269 F.2d 68, 70 (1959) (Clark, J.), cert. denied, 361 U.S. 934, 80 S. Ct. 374, 4 L. Ed. 2d 355 (1960). To this list we can now add Kane v. Branch Motor Express Co., 290 F.2d 503, 506-507, (1961), and, dealing with a somewhat different but related subject, Ellerman Lines, Ltd. v. The S.S. President Harding, 288 F.2d 288, 291-292 (1961). We find nothing to the contrary in Castro v. Moore-McCormack Lines, Inc., 325 F.2d 72, 75 (1963); rather that opinion recognized our established rule. However, the recurrent arguments on this point and the earnestness of appellees' presentation make further consideration appropriate.
The standard explanation of our rule was stated in Romero as follows: "determination of negligence involves first the formulation and then the application of a standard of conduct to evidentiary facts found to be established. When all this has been done by a judge, a reviewing court has no means of knowing whether he formulated the standard correctly, since he does not charge himself. Thus there must be free review of his ultimate determination of negligence although not of the facts on which it was based." See also Kane v. Branch Motor Express Co., supra, 290 F.2d at 506-507. Appellees say in effect, although they phrase it more politely, that this is mumbo-jumbo. Since every first-year law student knows what the standard of care is, how can an appellate court seriously wonder whether an experienced trial judge was aware of it? Particularly, how can any doubt be entertained on this score if the judge has correctly repeated the familiar formula? Does it not then follow that what the appellate court is attempting to oversee, at least in most cases, is the interpretation of particular circumstances in the light of a known rule of law? Is not such work the proper business of the trial courts, subject to review only under the "unless clearly erroneous" standards? And, however the argument on principle might stand, is not this what the Supreme Court decreed in McAllister ?
The argument on principle makes many over-simplifications that fail to take into account the elaborations of the general standard of care and the need for consistency in judicial decision. Of course we do not doubt the knowledge of the district judges that the overall standard of conduct to which an actor "must conform to avoid being negligent is that of a reasonable man under like circumstances." Ali, Restatement (Second), Torts § 283 (1965). But this is a long way from being the whole story. That simple phrase in the Restatement is accompanied by two pages of explanation; twenty-five more sections, §§ 285-309, with comment spreading over eighty pages, are devoted to further specifications of the general standard. The law of negligence has thus followed the path anticipated in the celebrated passage of Holmes' lectures of 1881, fulfilling "the tendency * * * to become more and more concrete by judicial decision and by statute" without interfering with "the general doctrine maintained as to the grounds of liability." The Common Law 89-103 (Howe ed. 1963). This very case presents issues whether the judge gave adequate weight to the principle that "As the gravity of the possible harm increases, the apparent likelihood of its occurrence need be correspondingly less," Prosser, Torts 151 (3d ed. 1964), and whether, as appellants claim, he acted on the erroneous view that unless all the cargo could be protected from damage, none needed to be. Save when the trial judge has gone to unusual lengths in his opinion, the only way in which an appellate court can determine whether he correctly apprehended these manifold specifications of the general standard is by reviewing his result, and if that review is to be effective, it must be unimpeded. Beyond this, the need for consistency militates against considering a judge's application of even an admitted legal standard as a finding of "fact" subject to F.R.Civ.P. 52(a) or its admiralty analogue, even though a jury's doing the same would constitute a determination of "fact" protected by the Seventh Amendment. It would be shocking if contrary decisions of two district judges in this circuit on exactly the same facts had to be left standing, although there would be no similar shock if such a divergence should happen as a result of the deliberation of two different juries. Compare United States v. Maybury, 274 F.2d 899 (2 Cir. 1960). Yet uniformity within a circuit or among circuits*fn3 can be achieved only if appellate review of the application of a legal standard is free of the shackles of the "unless clearly erroneous" rule.
This leaves the question whether McAllister compels a conclusion that our decisions are wrong as a matter of authority, however wise we may think them to be on principle. What proved to be the crucial issue in that case was not negligence, as to which this court had indicated a somewhat hesitant willingness to go along with the district judge's conclusion, but causation, see 207 F.2d 952, 954, 955 (2 Cir. 1953); with respect to that question, this court, speaking through Judge A. N. Hand, did not specify what standard of review it was using. The brief opinion of Mr. Justice Minton assumed that we had applied the "unless clearly erroneous" standard, saying "We do not find that the Court of Appeals departed from this standard, although we do disagree with the result reached under the application of the standard." 348 U.S. at 20, 75 S. Ct. 6, 8, 99 L. Ed. 20. Taking this literally, one could argue that the decision laid down no legal doctrine at all but simply resolved a controversy over facts in favor of the view taken by the district judge. See Staring, Appeals in Admiralty Cases, 35 Tulane L.Rev. 1, 47-50 (1960). But it can be answered that since the Supreme Court ought not to have reversed us if in fact our power went beyond the rejection of a clearly erroneous finding, the Court necessarily decided that causation was a question of "fact" on which the decision of a district court was reviewable only for clear error. Accepting this broad reading for present purposes, we note that the issue in McAllister was simply whether the plaintiff had borne his burden of showing the probability that defendant's negligence rather than other circumstances caused the injury, as distinguished from questions of causation involving a mixture of law and fact such as whether the negligence was the "proximate" or "legal" cause. Finding no discussion in the Supreme Court's opinion directed at what could be considered application of a legal standard to established facts, we adhere to our long-held view that a judge's determination on the issue of negligence does not fall within the "unless clearly erroneous" rule.
Hurricane Donna was first identified on September 2, 1960, as a severe hurricane several hundred miles east of Puerto Rico, but no one then regarded her as presenting any immediate threat to the northeastern United States. In the following week she moved slowly, in a generally westerly direction, passing north of Puerto Rico and Cuba and aiming toward the Florida keys and thence into the Gulf of Mexico. However, hurricanes on this general course are often subject to a phenomenon known as recurvature -- a turn to the north or northeast as a result of encountering southerly and westerly winds near the western end of the Bermuda High, a large high pressure area of varying length which extends west from the Azores. The point at which recurvature occurs, whether off the east coast of the United States or in the Gulf of Mexico, depends on the length of the Bermuda High which tends to shorten progressively from August through the fall months; when recurvature occurs, the hurricane not only changes direction but accelerates from a speed of some 12 miles per hour to several times that rate.
According to the Chief of the National Hurricane Center, a specialized research and warning service of the United States Weather Bureau stationed at Miami, Donna's recurvature began at 7:00 A.M. EST on Friday, September 9, "when it changed from a predominantly western trajectory, first to west-northwest and later to northwest." However, the "advisories" issued by the Weather Bureau during that day did not make alarming news out of this; the 5 P.M. advisory, noting the north-northwest course and the reduced speed of 9 mph, predicted "no material change * * * in intensity, speed or direction during the next 12 hours," and the 11 P.M. advisory went only so far as to forecast that "the direction of movement will likely change to slightly more northerly." In fact, until late in the evening of Saturday, September 10, the Hurricane Center expected that over the weekend Donna would pass up the Gult of Mexico some 60 or 70 nautical miles west of Tampa, cross the Florida coast near Cedar Key, and then move northeast off Cape Hatteras by early Monday. Instead, the center of the hurricane hugged the west coast of Florida only as far as the Tampa area, then turned northeast across the state, and moved up the Atlantic Coast at a considerably greater speed than the experts anticipated. No Weather Bureau report extending hurricane warnings to include New York was issued until Sunday, September 11, at 10 P.M. EDT; although this correctly estimated that Donna would reach the latitude of New York on Monday morning, it erroneously predicted her passage a short distance off the eastern end of Long Island. Later, at 11:55 P.M., the New York Weather Bureau issued a local statement warning that tides on Monday morning might be as high as 5 foot to 8 foot above normal from the northern New Jersey shore to Long Island; the following morning's report repeated this forecast but added that in New York harbor tides would be about 3 foot to 4 foot above normal. In fact, traveling somewhat west of the anticipated course, the "eye" of the hurricane brushed the New Jersey shore, and between 2:30 and 3:30 P.M. on September 12 passed over central Long Island. The ensuing storm surge in New York harbor was unusually great because of the unprecedented width of the eye and the proximity of its northeastern quadrant where surge is maximized. With Donna hitting exactly at the time when the astronomical tide reached its peak of 4.5 foot, New York harbor experienced the highest recorded water level in its history, some 10.7 foot to 10.9 foot above mean low water in the area of the Bush Terminal.
The claimants do not criticize the conduct of the pier operator once the Sunday evening advisories had been issued. The terminal superintendent got down early on Monday morning and, finding that the Turandot and the Tatra refused to open their hatches for loading because of heavy rain, put the stevedores who had been ordered for 8 A.M. to work in nailing up doors, affixing tarpaulins, and placing additional pallets under cargo; only some 3% to 5% of the cargo had been raised by extra pallets when the flooding occurred. What the claimants do criticize is the operator's failure to take precautions before then. Their case centers on the point that under the practices of the Waterfront Commission labor for Saturday, Sunday and Monday must be ordered by 3 P.M. on Friday.*fn4 They say that a pier operator who closes down his pier knowing he will be immobilized for 63 hours must take substantially greater precautions that one who will regain freedom of action in fifteen. The prudent course under the circumstances, they urge, would have been to hire men for the weekend, place extra pallets under the inbound cargo unloaded on Friday or, at least, so much of it as could be feasibly protected in that fashion, and load the outbound cargo on Saturday and Sunday. They say that, despite the lack of official warnings, such action was demanded by a combination of the known severity of Donna, a history of some previous flooding, the unpleasant habit of hurricanes to recurve and threaten the northeast, particularly in early September, and the unpredictability of their course and speed once recurvature has set in.*fn5 Even conceding that reliance on Weather Bureau warnings is a reasonable course of conduct for a pier operator who will resume business on the following morning, the claimants stress that the forecasts are confined to a 24-hour period; they say that when a person disables himself from acting for longer than that, he must bear the consequences of what may have seemed only a slight risk at the time of decision. Though we agree that the irreversible character of a decision is highly relevant to the risk, we concur with the district judge that the pier operator did not fail to exercise the care it was bound to take with the knowledge it had or ought to have had before closing down on Friday.
An actor is required to know "the qualities, characteristics, and capacities of things and forces in so far as they are matters of common knowledge at the time and in the community," Restatement (2d), Torts § 290(a). This will suffice unless he is "engaged in an activity, or stands in a relation to others, which imposes upon him an obligation to investigate and find out, so that he becomes liable not so much for being ignorant as for remaining ignorant." Prosser, Torts 163 (3d ed. 1964). The pier operator clearly met the less stringent test; there was no general knowledge in the New York community of the probable approach of Donna until Sunday evening. And even if the claimants are right in saying that the situation on Friday afternoon had sufficient elements of danger that the operator should have made further inquiry before immobilizing itself for 63 hours, there is no evidence that the knowledge which it would thus have obtained and with which it must be charged should have led to a different course of conduct. If the operator had communicated with the meteorologist in charge of the New York Weather Bureau, it would have learned that, as he later testified, he was not then "very excited about it [Donna] affecting the New York City area, especially over the weekend." If it had gone still further and telephoned the Chief of the Hurricane Center at Miami, it would have found he expected that the center of the hurricane would be on West Florida Coast on Saturday evening and, had he been able to anticipate a forecast in fact made the next day, that on Sunday evening it would lie somewhere between Savannah, Georgia, and Charleston, South Carolina. Neither expert would have predicted that by Sunday evening Donna would be over North Carolina, racing northeastward on a course likely to take her dangerously near New York harbor by Monday afternoon. To be sure, if ...