This now famous case, Romero v. International Terminal Operating Co., D.C.S.D.N.Y.1956, 142 F.Supp. 570, 2 Cir., 1957, 244 F.2d 409, 1959, 358 U.S. 354, 79 S.Ct. 468, 3 L.Ed.2d 368, returns to us after a voyage that has taken it to the Supreme Court and then back to the Southern District. Most of the dramatis ;personae on the defendants’ side have-disappeared — the steamship owner, Compañía Trasatlántica, as a result of the-Supreme Court’s ruling; International Terminal Operating Co., lessee of the-pier and employer of the longshoremen (who boarded the vessel only after the-accident), through discontinuance by the-plaintiff; and Quin Lumber Co., the marine carpenter contractors, after a settlement for $25,000. Plaintiff went to-trial solely against Garcia & Diaz, Inc., hereafter G & D, the husbanding agent for the vessel. Somewhat surprisingly in view of the history of the case, see D.C., 142 F.Supp. at page 574, and Currie, The Silver Oar and All That: A Study of the Romero Case, 27 U.Chi.L.Rev. 1, 3-8 (1959), the trial was without a jury.
After a ten day trial, the District Judge made findings of fact and conclusions of law which were entered as a. judgment. He concluded that G & D' “had substantial control of the activities, and circumstances proximately related to the plaintiff’s injuries,” and that “The defendant was negligent and its negligence was proximately related to the accident * * * ” He awarded plaintiff" $4,022.50 for hospital and related expenses and loss of earnings from the date of" the accident, May 12, 1954, to May 16, 1955, when the plaintiff again became employed, and $65,000 for mutilation, humiliation, pain and suffering, from which the $25,000 paid in settlement by Quin was to be deducted. He declined to award damages for loss of earnings after May 16, 1955, on the ground that Romero’s present earnings as a power-press operator in the United States considerably exceeded what he had earned' as a Spanish-seaman prior to the accident. Defendant appeals from the entire judgment and plaintiff from the refusal to-award damages for loss of earnings after reemployment. We have concluded that, the evidence did not warrant a judgment against G & D and accordingly that the judgment must be reversed and the complaint dismissed.
Since none of the three previous opinions recite the circumstances of the ac*349cident in any detail and Judge Palmieri’s findings and conclusions have not been reported, it becomes necessary to outline the facts.
The S.S. Guadalupe was a combination passenger and cargo vessel of Spanish ownership and registry, owned by Compañía Trasatlántica and in service between Spain and ports on the eastern and southern seaboards of the United States, Cuba, and Mexico. G & D served Compañía Trasatlántica as agent for the United States and Cuban ports. It had booked a shipment of grain for an eastbound trip; while the Guadalupe was in New. York harbor on her westbound voyage to Cuba, Suarez, a responsible employee of G & D, had a representative of the Board of Underwriters come to the ship and discuss the stowage of the grain with the mate, who gave orders as to what shifting boards, feeders, etc., he wished constructed for stowage on the ship’s return to New York. G & D arranged for Quin to do this work. Learning that the Guadalupe would arrive eárly on May 12, 1954, at an International Terminal Co. pier in Hoboken, Suarez instructed Quin to have its carpenters ready that morning and International to have longshoremen available at 1 P.M., by which time it was thought the carpentry would be completed and the vessel ready to load.1
As the Guadalupe sailed up the harbor, the crew removed her booms from their cradles and placed them at a 30 to 35 degree angle from the mast, parallel with the center line of the vessel, and secured by guys. Shortly after 8:30 A.M., a gangplank was put ashore to Pier I at Hoboken. Suarez came aboard, along with some 20 Quin carpenters under a foreman. The Court found that, in the course of removing tarpaulins and hatch covers, the carpenters cluttered the deck space in the vicinity of the No. 2 hatch and disarrayed the cables used for raising and lowering the forward starboard boom on that hatch which “had been disposed beside the hatch in an orderly fáshion by the crew members who had removed the boom from its cradle earlier that morning.” Then Selvaggio, foreman of the Quin carpenters, approached Sampedro, boatswain of the Guadalupe, who was working at the No. 5 hatch, and requested Sampedro to lower the boom at No. 2 hatch so that lumber needed by the carpenters could be lifted from the pier. Campos, the chief mate, and Suarez were standing together on the deck near the No. 5 hatch. Sampedro reported Selvaggio’s request to Campos, who authorized him to comply; Suarez was in a position to hear this but denied that he did. Sampedro went off, taking with him the assistant boatswain, the plaintiff and another seaman. Romero’s assignment was to handle the topping lift cable which secured the boom as it was being lowered; the cable was being fed around a small drum or “gypsyhead.” A marine carpenter voluntarily took a position several feet behind Romero and handed him the moving cable; a portion of this contained a kink, or “codillo,” which Romero testified had been there before the vessel reached New York. Since Romero could not clear the kink in time to save himself from being caught in the revolving drum, he let go of the cable, whereupon a turn or turns slipped off the drum and entangled his legs. The slipping of the cable caused the boom to drop of its own weight. The force of the fall of the boom, with its consequent pull on the cable, resulted in the amputation of Romero’s left foot and a severe fracture of his right leg.
The theory of plaintiff’s case was that G & D was in control of the loading, either alone or jointly with the owner. Negligence appears to have been predicated on three alternative grounds: first, that it was negligent not to have the boom lowered by longshoremen; second, *350that if the boom was not to be lowered by longshoremen, it was negligent not to have it lowered by the carpenters; and, third, that, in any event, it was negligent to permit the boom to be lowered as it was in view of the disarray alleged to have been created by thé carpenters, their claimed presence on the scene, and the intervention of the volunteer. Recognizing the limitations placed upon the scope of review by F.R.Civ.Proc. 52, 28 U.S.C.A., we are unable to sustain the trial judge’s conclusions as to control and of negligence by G & D.
We are met at the outset by a claim of G & D that the issue of control was settled in its favor by the Supreme Court. The relevant passage is at 358 U.S. 384-385, 79 S.Ct. 487; we quote it in full in the margin.2 The contention is that since the only issue remanded was “a claim independent of the employment relationship or operation and control,” the District Court was forbidden to find that G & D had any degree of control. This is reading the words out of context. All that the Supreme Court ruled out was a contention that G & D had such an “employment relationship or operation and control” as to give rise to claims for unseaworthiness or maintenance and cure or under the Jones Act, cf. Cosmopolitan Shipping Co. v. McAllister, 1949, 337 U.S. 783, 69 S.Ct. 1317, 93 L.Ed. 1692. On the other hand, since it is essential to a conclusion of liability for negligence that defendant was or ought to have been able to prevent the harm, the Supreme Court, in remanding for trial of “a claim based upon the negligence of Garcia & Diaz,” could not have meant to preclude the trial court from finding any facts, established on the trial, that would show such liability.
The relationship between Compania Trasatlántica and G & D was outlined in an agreement made in 1948. This designated G & D as the “consignee” of arriving vessels. G & D was to maintain offices in New York, to obtain passengers and cargo for the vessels, and to enter into all necessary agreements to that end. It was also to collect moneys owing to Compania Trasatlántica, to institute proceedings to recover such moneys, and to expedite the arrival and departure of vessels by carrying out all customs, health, police and other regulations. Further, G & D was to receive cargo for the shipowner’s account, to safeguard it, and to deliver it to its consignee. Finally, G & D was “To organize for the account of the Ship-Owner all operation of boarding and landing passengers, loading and discharging of merchandise and baggage and other tasks called for in port to assure regular service on the part of the Ship-Owner’s vessels, making available to them for the account of the Ship-Owner, all equipment in personnel and material required for the purpose and governed by the availabilities at each port.” For these services G & D was to receive a commission of 5% on originating and 2%% on terminating business, with a minimum of $500 for a freighter and $750 for a passenger vessel. G & D was to bear the cost of the maintenance of its office, including rent and employees’ salaries, but not out-of-pocket expenses which it paid simply for the owner’s account.
The degree of control exercised by G & D must be determined not simply from the language of the agreement but from operation under it. Much evidence was taken on this. G & D’s case was that its function was to serve as a coordinator *351between ship and shore; it would decide where Compañía Trasatlántica vessels would dock and who was to supply longshoremen and then would act as intermediary between the officers of the vessel and the stevedoring company. The usual practice was for Suarez to board Compañía Trasatlántica vessels on their arrival, take aboard the mail and booking sheets for cargo, assist debarking passengers requiring hotel accommodations or other special services, talk with the mate, see that loading or unloading operations had gotten started, and then depart. Normally he would spend one or two hours on the vessel in the morning and again in the afternoon, although he might altogether miss days other than those of arrival and departure. When on board, he would frequently act as liaison between the officers and shore personnel, a task for which he was peculiarly fitted by his fluency in Spanish as well as English. This, according to Suarez, was the limit of what he did or was expected to do.
There is much controversy whether or not Suarez bore the title “Pier Superintendent.” The evidence was that he had acquired that title in earlier years when G & D had itself leased piers and, as is not unusual, the title may have stuck after the office ceased. Plainly Suarez did not have powers of superintendence on Pier I, that function being discharged by Captain Gamer, an employee of International Terminal, and the expert testimony as to the responsibilities of a true “pier superintendent” was thus of little relevance. The only evidence that could be seriously urged as contradicting Suarez was certain testimony of Selvaggio, the carpenter foreman; but when this is read as a whole, it also is consistent with the position that Suarez, when on the ship, acted only as intermediary between the carpenters and the officers — Selvaggio admitted he had never had occasion to appeal to Suarez to give orders to the officers.3 True the judge was not obliged fully to credit Suarez, whom he evidently regarded as a not wholly trustworthy witness, as will be seen below; but Suarez’ testimony as to the general nature of his responsibilities was supported by that of Martinez, the treasurer of G & D, of the mate and of the boatswain, and accords with normal business practice — -an agent like G & D is usually employed to perform functions the ship cannot perform without a shore establishment, not to do things she can do quite as well herself. Clearly G & D was not in control of the vessel during the long periods when Suarez was absent; there would hardly be a shift to joint control whenever he came on board.
Taking the evidence most favorably to the plaintiff, it is apparent that the relation of G & D to the vessel differed toto coelo from that of the defendant in Quinn v. Southgate Nelson Corp., 2 Cir., 121 F.2d 190, 191, certiorari denied 1941, 314 U.S. 682, 62 S.Ct. 185, 86 L.Ed. 546, upon which the Court relied in imposing liability. In that case, the defendant had undertaken to “manage, operate and conduct the business of the line” and “to man, equip, victual, supply and operate the vessels, subject to such restrictions and in such manner as the owner may prescribe.” There the agent was the operator of the vessel, picking the officers and crew subject only to a veto by the Maritime Commission over choice of the former and to other restrictions of a most general sort. Here the officers and crew were selected by the owner, and the ship’s officers continued to function while the Guadalupe was in port.
We accept fully that “The liability of an agent for his own negligence has long been embedded in the law,” Brady v. Roosevelt S.S. Co., 1943, 317 U.S. 575, 580, 63 S.Ct. 425, 428, 87 L.Ed. 471. But it by no means follows that an agent is liable whenever its principal would be. G & D did not have an affirmative duty to see to it that plaintiff was provided with a safe place and with safe *352methods to work. Admittedly, Suarez did not himself harm Romero, and no other employee of G & D was about. Under such circumstances an agent is liable in tort only if he “directs or permits conduct of another under such circumstances that he should realize that there is an unreasonable risk of physical harm to others or to their belongings.” American Law Institute, Restatement of Agency 2d, § 351.
It is thus idle to discuss “control” on an all or nothing basis as the parties have tended to do; what must be determined is whether the judge was warranted in finding that G & D directed or permitted particular conduct entailing an unreasonable risk of harm to Romero. We therefore turn to the various theories of negligence advanced on plaintiff’s behalf:
(1) Failure to have had stevedores aboard to lower the boom.
Judge Palmieri found, as part of his finding 16:
“Suarez knew that if longshoremen had been aboard on the morning of May 12, 1954, they alone would have been responsible for lowering and spotting the No. 2 forward starboard boom in order to raise the lumber and load it aboard ship for use by the carpenters. Suarez also knew it was the prevailing custom and practice in New York harbor in 1954 for the lowering or (2126) raising of ships’ booms, once they were rigged and out of their cradles, to be done by stevedores; and in exceptional circumstances, by the crew in the absence of stevedores. He also knew or should have known that it was unusual to have marine carpenters on board without stevedores.”
Although it is not at all clear that this entered into the judge’s conclusion of negligence, the issue is of peculiar importance since, as explained in footnote 1, he would have been warranted in finding that the decision not to have stevedores come aboard until 1 P.M. was made by G & D, although it was doubtless concurred in by the owner. And if the finding here under discussion supported a conclusion of negligence, plaintiff could defend the judgment on that ground even though it was not the one taken by the judge. Langnes v. Green, 1931, 282 U.S. 531, 535-539, 51 S.Ct. 243, 75 L.Ed. 520.
We may assume it to be generally true that, as stated in the first quoted sentence, if longshoremen had been aboard, they would have handled the lowering and spotting of the boom; but this does not make it negligent not to have had them. If the first clause of the second sentence means only that booms of ships in New York harbor were lowered or raised by stevedores more often than by others, it likewise is immaterial; for it includes the typical ease where the stevedores have come on to unload or load cargo, as well as the case here presented where construction work must first be done. We do not know whether the judge considered the latter to be examples of the “exceptional circumstances” referred to- in the second clause; if so, the finding would seem to negate rather than support a conclusion of negligence from failure to hire stevedores in such a case. Again the last sentence leaves us in doubt whether “unusual” means numerical in frequency, which would be inconclusive, or bad practice; if the latter, we would be constrained to hold it without adequate evidentiary support.
Turning from the finding to the evidence, we think there was none to prove that safety required the hiring of longshoremen in order to lower a single boom to raise lumber for the carpenters. There was testimony that, in Hoboken and some other parts of the Port of New York, there was a practice of having a gang of seven longshoremen come aboard to lower and spot booms somewhat in advance of arrival of the larger gangs. The inference sought to be drawn is that this would not be done if it were deemed safe for the crew to lower the booms. However, there was no evidence that this practice was pursued, at least on foreign ships, when only a single boom had to *353be lowered to raise lumber need for the construction of fittings; and there was much to suggest that the practice with respect to American ships rested on union demands for job security rather than on safety considerations. Nothing in the entire record established that it was dangerous for the crew of a vessel, and particularly for the crew of the Guadalupe, to handle their own booms. Indeed, the judge found that this was not hazardous “under normal conditions,” as we shall see. The only abnormal condition foreseeable to G & D when it was decided not to have longshoremen come aboard on the Guadalupe’s arrival was that carpenters would be on board when a boom would be lowered to raise the lumber. This alone would not create such a hazard as to require longshoremen rather than an adequate complement of the crew.
Romero testified he was familiar with lowering booms and had done it many times; he had said in his deposition that the lowering of booms was a daily maneuver aboard the ship, that “even a cabinboy” knew how to do it, and that he had known how for a long time. True, in direct examination at the trial, he claimed he had never spotted a boom over a dock and had no training in such an operation, and his counsel attempts to distinguish between a mere lowering of the boom and its lateral movement as the boatswain pulled sidewise on the guy to center it over the lumber that was to be brought aboard. This is unpersuasive. Romero had no additional duties to perform by reason of the lateral movement; nor is it even suggested that the boatswain was negligent in his pulling of the guy, or, indeed, that anything that would not have been done if the only movement had been vertical contributed to the accident. Romero was injured because he fed a kinked wire onto the gypsyhead; he would have been equally injured if the only task were the lowering of the boom, a task which he admitted he was quite capable of doing and had often done.
(2) Failure to have the boom lowered by the carpenters in the absence of longshoremen.
The judge made no finding that, in the absence of longshoremen, it was negligent not to have the boom lowered by the carpenters rather than the crew; indeed, his finding seems to have been to the contrary, as we shall see. Nevertheless plaintiff is entitled to seek to defend the judgment on that ground if the record supports it, Langnes v. Green, supra, although in this respect without benefit from the “unless clearly erroneous” rule. The record does not do so, quite apart from the fact, developed in (3) below, that, in contrast to the decision not to summon longshoremen, there was no sufficient evidence that G & D was responsible for the decision to have the crew rather than the carpenters do this work.
Testimony that the practice was to have such work done by the carpenters was given by two of plaintiff’s experts, Ash and Orlando; however, Ash’s evidence was that this practice was designed primarily for the safety of the carpenters and Orlando’s seems to relate to the situation where the carpenters were already working in the hatch rather than, as here, where they were still on the deck, and thus also is oriented to the safety of the carpenters rather than of the crew. As against this the boatswain and the mate testified they had never seen this work done by carpenters rather than the crew; and four other witnesses —the foreman of a large marine carpentry firm, the pier manager for International Terminal, an experienced steamship captain, and a witness who combined experience as a master with that as a stevedoring superintendent — agreed that, under the circumstances here, the practice in the Port of New York was to have the boom lowered by the crew rather than by the carpenters. Moreover, Selvaggio, the Quin foreman, and Amodio, his assistant, stated their practice was to do precisely what was done here, namely, to “get the mate or the boatswain and tell him if he could do it for *354us.” In the light of all this, as well as the discussion as to the capability of the crew under point (1), a finding of negligence on anyone’s part in having the boom spotted and lowered by the crew rather than the carpenters would have been clearly erroneous.
(3) Other acts of negligence.
In addition to the extract from finding 16 quoted in (1) above, the judge’s basic findings of negligence were contained in his finding 21; this we quote in full:
“21. Under normal conditions, the boatswain Sampedro, the plaintiff, and the two other seamen who were engaged in the work task of spotting the boom, would have been sufficient in number and skill to have accomplished the assignment without reasonably foreseeable risk of injury. The conditions at the time and place of the accident were not normal, however, and were, in fact, unusual and hazardous in the following respects:
“(a) The presence and participation of the unskilled marine carpenter, as aforesaid;
“(b) The cluttered condition of the deck and the disarray and resulting kinked condition of the cable which was passed into the plaintiff Romero’s hands shortly before the accident;
“(c) The presence of several marine carpenters in and about the area of the accident;
“(d) The inadequacy of the work crew engaged in the said task because of these unusual conditions.”
Defendant attacks all four subsidiary findings either as unsupported by the evidence, or as relating to matters having no causal relation to the accident, or as both. Although it may not be strictly necessary to rule on all these objections since, for the reason hereafter discussed, the judgment could not stand even if all the findings were warranted, it will add to clarity if we do this; we find merit in some of defendant’s criticisms but not in all.
Taking the subsidiary findings in reverse order, we think there was no substantial evidence that a crew of four men was inadequate, save as the particular condition of the cable may have required more. The judge himself found that, “under normal conditions,” four “would have been sufficient in number and skill to have accomplished the assignment without reasonably foreseeable risk of injury * * we accept that finding. If the boatswain, his assistant and two seamen would have sufficed “under normal conditions,” we see no basis for concluding that “The presence of several marine carpenters in and about the area of the accident,” assuming in plaintiff’s favor that the judge was justified in finding this to be the fact, added such an element of danger as to have made the operation hazardous to the crew or even to have required a larger complement if such hazard was to be avoided. Almost all the testimony that it was bad practice for the crew to lower a boom while shore personnel were standing by seems to have been concerned with danger to the latter, who might be hurt by the boom or other gear, rather than with added hazard to the crew, whose maneuvers would be the same regardless of the bystanders. The cluttered condition of the deck, if such there was, does not appear to have had any causal relation to the accident. On the other hand, the disarray of the cable by the carpenters, a fact which we assume in plaintiff’s favor although seriously disputed, might have caused Romero to have had less notice of the approach of the “codillo” than if the cable had been stretched, and hence may have made it desirable to supplement the normal complement with a seaman experienced in the work and capable of warning Romero in Spanish, — and this even though Romero admitted that the “codillo” that caused the accident did not result from the disarray of the cable and could not have been removed. Indeed, the badly kinked condition of the cable might have made this advisable, quite apart from the alleged disarray. Also, although there is no evidence that “the *355unskilled marine carpenter” affirmatively contributed to the accident, a fellow crewman experienced in handling the cable might have warned Romero and slowed the feeding of the cable to the gypsyhead.
The difficulty which plaintiff cannot surmount is that while the evidence would thus suffice to establish negligence, as well as unseaworthiness, if the action were against an owner, it does not cast G & D as liable without further proof. The additional proof needed was that G & D, knowing or having a duty to know of the dangerous condition created by the carpenters’ disarray of a cable with kinks or by the kinks themselves, directed or permitted the lowering of the boom by an inadequate complement of crew. The evidence established none of this. There was no evidence that Suarez knew of the kinking in the cable or the disarray alleged to have been created by the carpenters at No. 2 hatch. Neither was there any basis for charging G & D with such knowledge; as husbanding agent, G &• D had no responsibility for the ship’s gear and its duty was to arrange for the presence of the carpenters, not to supervise what they did. Whether the judge was justified in finding that Suarez heard the boatswain report the carpenter foreman’s request to the mata is disputed; we shall assume in plaintiff’s favor that he was, since Suarez was in a position to hear the report and the judge was not bound to credit his denial even in the absence of contradiction. However, this is not enough. Even if the conversation apprised Suarez that the boom was to be lowered by the crew, it was not negligent for Suarez to permit this, and nothing gave him any indication that a larger than normal gang might be required or that the boatswain would not provide one if it were. Determination of all this was the responsibility of the representatives of the crew’s employer, namely, the ship’s officers, not of Suarez; there was no reason for him to assume the boatswain would not recruit a complement “sufficient in number and skill” under whatever conditions the boatswain found to prevail when he arrived at No. 2 hatch. We need not speculate whether the mate would have honored a suggestion on Suarez’ part; it was not Suarez’ duty to make one.
Many decisions in this Circuit, some in civil actions and others in admiralty, have held that a judge’s determination of negligence, as distinguished from the evidentiary facts leading to it, is a conclusion of law freely reviewable on appeal and not a finding of fact entitled to the benefit of the “unless clearly erroneous” rule. Sidney Blumenthal & Co. v. Atlantic Coast Line R. Co., 2 Cir., 1943, 139 F.2d 288, 290, certiorari denied 1944, 321 U.S. 795, 64 S.Ct. 848, 88 L.Ed. 1084; Barbarino v. Stanhope S.S. Co., 2 Cir., 1945, 151 F.2d 553, 555; Kreste v. United States, 2 Cir., 1946, 158 F.2d 575, 577; Great Atlantic & Pacific Tea Co. v. Lloyd Brasileiro, 2 Cir., 1947, 159 F.2d 661, 665, certiorari denied 331 U.S. 836, 67 S.Ct. 1519, 91 L.Ed. 1849; Guerrini v. United States, 2 Cir., 1948, 167 F.2d 352, 356, certiorari denied 335 U.S. 843, 69 S.Ct. 65, 93 L.Ed. 393; Johnson v. United States, 2 Cir., 168 F.2d 886, 887; Lynch v. Agwilines, 2 Cir., 1950, 184 F.2d 826, 828. Despite possible negative inferences that might have been drawn from McAllister v. United States, 1954, 348 U.S. 19, 75 S.Ct. 6, 99 L.Ed. 20, we have continued to apply this rule both in civil actions, Dale v. Rosenfeld, 2 Cir., 1956, 229 F.2d 855, 858, and in admiralty, New York, New Haven & Hartford R. Co. v. Gray, 2 Cir., 240 F.2d 460, 465, certiorari denied 1957, 353 U.S. 966, 77 S.Ct. 1050, 1 L.Ed.2d 915; Verbeeck v. Black Diamond S.S. Corp., 2 Cir., 1959, 269 F.2d 68, 70. The basis of these decisions is that 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 *356review of his ultimate determination of negligence although not of the facts on which it was based.
The rule of free review applied in these cases would plainly require reversal here. However, even if we are no longer permitted to adhere to what we regard as the sound view taken in this long line of decisions, as our colleagues in the Ninth and Sixth Circuits tell us, Pacific Tow Boat Company v. States Marine Corp., 9 Cir., 1960, 276 F.2d 745, 752; Imperial Oil, Ltd. v. Drlik, 6 Cir., 1956, 234 F.2d 4, 10, certiorari denied, 1956, 352 U.S. 941, 77 S.Ct. 261, 1 L.Ed. 2d 236, an issue which we may leave for another day, we are constrained to reverse here on the basis, stated in United States v. United States Gypsum Co., 1948, 333 U.S. 364, 395, 68 S.Ct. 525, 542, 92 L.Ed. 746, that “a finding is ‘clearly erroneous’ when although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.”
The judgment is reversed and the complaint dismissed.