dissenting:
The majority opinion approves the district court’s refusal to instruct the jury on superseding cause despite evidence that the stevedore company had knowledge of the dangerous condition and declined to adopt the safest procedures. The stevedore deliberately elected instead to disregard the defective stacking of the carpet backing rolls and, in its own interest, to “use” the condition to speed unloading. Because I believe that in such a ease a general instruction on proximate cause is wholly insufficient and an instruction on superseding cause is required, I respectfully dissent.
I.
I.T.O. Corp. (I.T.O.), a stevedore company, had undertaken as an independent contractor the job of unloading from the EXPORT AGENT, a ship owned by defendant American Export Lines (Export), rolls of carpet backing. At the time of the accident, DiRago, Stephen Pakech and gang foreman William Collins were part of an I.T.O. crew working on the ’tween deck of No. 3 hold. Approximately 104 rolls were stowed on the deck, half on each side of the vessel, arranged athwartship so that the lengths of the rolls were perpendicular to the centerline of the vessel. The rolls had been stacked in Bangladesh in four or five tiers, with rolls of each tier above the first resting in recesses in the tier below, as a load of pipe would be stowed. The accident occurred on the starboard side after many of the rolls had already been removed, creating a “V” cut in the stowage that went to the floor. DiRago was standing in that cut when the remaining rolls of stacked carpet backing shifted and rolled into the cut, striking and injuring DiRago.
Although there was conflicting evidence about which procedures were used in the unloading and about the propriety of some of those procedures, certain facts are undisputed, and some of the disputes are not material to the issue on appeal. Export allegedly was negligent in failing to provide a safe working place for the longshoremen (1) by not requiring that the load of rolls be chocked when stowed in Bangladesh, and (2) by failing to warn the longshoremen of this unsafe condition. Export admits that no chocks, small wooden blocks, had been placed in the rolls. DiRago’s theory in the district court was that customarily chocks are placed against each roll in the bottom tier, and that had chocks been placed there, the load would not have shifted causing him injury. Although at trial, Export offered conflicting testimony about this “custom,” on appeal Export does not challenge the finding of the jury that the rolls were improperly stowed and that Export was negligent for allowing such stowage. Rather, Export’s contention is that the negligence of a third party, the stevedore company, I.T.O., is of such a nature as to relieve Export of liability for its negligence. Expressed in other words, Export argues that the alleged negligence of I.T.O. is a superseding cause of DiRago’s injuries, as provided in Restatement (Second) of Torts §§ 440-442, 452 (1965). To evaluate this argument it is necessary to understand the facts leading to DiRago’s injury.
DiRago, Pakech and Collins began their work on the ’tween deck of No. 3 hold at approximately 10:30 a. m. on April 14, 1977. To remove the rolls of carpet backing they used a “chisel,” a forklift equipped with a prong that could be inserted into each roll. By using the chisel, each roll could be lifted from the stack and transported to an area below the hatch. There the rolls were placed on a sling which, when a number of rolls were gathered together, was lifted with a hoist through the hatch and out of the hold.
When the stevedore crew began work, they did not know that the rolls were un*870chocked. At no time did Export warn them of this condition or advise them to follow special procedures. Despite the lack of chocking, had the rolls been removed tier-by-tier, the accident would never have occurred. When unloading began, tier-by-tier removal was not possible. First, the approach with the forklift to the forward part of the stowage on each side was blocked by two stacks of plywood. Collins, the gang foreman, asked an officer of the vessel if the plywood could be placed temporarily on the dock to facilitate the unloading, but the officer refused, fearing the theft of the plywood if it were removed. Collins could have “appealed” this denial, but he declined to do so. Second, access was further limited by two-foot diameter stanchions, one at each of the two forward corners of the hatch. Third, there was disputed evidence that removal of the top tier was limited by the coaming extending into the hold from around the hatch above. The coaming however, did not limit removal of lower tiers and did not make necessary a cut to the floor. Further, apparently the stanchions alone would not have prevented tier-by-tier removal, although they might have made it more difficult. Thus tier-by-tier removal would have been feasible except for the plywood, the removal of which might have been accomplished had Collins pursued the issue.
The carpet backing on the port side was removed first without incident. The crew used the “V” cut method. By so doing, when they first reached the floor, long before they finished the port side, they discovered that the rolls were not chocked. Despite this discovery, they continued to unload by using the cut in the stack. In fact, DiRago testified tier-by-tier removal was rejected because it would have taken “too much time.” The forklift operator actually took advantage of the lack of chocks, undermining the stack and allowing several rolls at a time to fall to the floor where they were more easily picked up. There was evidence that this method violated Occupational Safety and Health Administration (OSHA) regulations.1 See 29 C.F.R. §§ 1918.83(a) & (b) (1980). Even though the rolls were not chocked and the crew purposely let them collapse to the floor, all the rolls stowed on the port side had been removed without mishap shortly before the noon lunch hour.
Work began on the starboard or opposite side before noon. At that time, with the port side clear, the several bundles of plywood could have been easily moved out of the way without taking them from the hold, thus removing the only real obstacle to tier-by-tier unloading. This was not done.2 Instead, the stevedore crew continued to unload by using a cut to the floor. The lunch hour came and work stopped. Shortly after work resumed following lunch, when approximately 15 feet of floor had been cleared in the middle of the stack, the accident occurred. DiRago testified that he had just finished guiding, by hand, a sling of rolls being hoisted from the hold. He stepped into the clearing in the stack to get clear of the load going up. At this time, according to DiRago, the stack of rolls remaining in the forward part of the hold collapsed. The falling rolls struck and injured him. Despite this accident, when work resumed after the injured DiRago was removed, the I.T.O. crew continued to unload by the same method of rolling the carpet backing into the cut.
*871At the trial, the defendant requested that the jury be instructed on superseding cause because, Export argued, I.T.O.’s negligence in its choice of a method to unload cargo was negligence of such a nature that the jury could find it to be the superseding cause of the injury, thus relieving Export from liability. The district court declined, stating:
I will not charge as to superseding cause. It seems to me that if there is negligence, which is a proximate cause, that is all that is necessary, and that if, in fact, there is a superseding cause, it’s a superseding cause, because the negligence itself is not a proximate cause.
In other words, if there is negligence which is a proximate cause, and is a substantial factor, then I don’t see how there could be, at least in this case, a superseding cause.
Subsequently, the jury found (1) that Export was negligent and that the negligence of Export was a proximate cause of DiRago’s injury, and (2) that DiRago was negligent, that DiRago’s negligence was a proximate cause of his injury, and that his negligence contributed five percent to his injury.
II.
The district court’s reasons for refusing the instruction are not persuasive. That court apparently could not discern the distinction between proximate cause and superseding cause. In refusing the instruction, the court stated that “if, in fact, there is a superseding cause, it’s a superseding cause, because the negligence itself is not a proximate cause.” The district court correctly defined proximate cause as “a substantial factor and a direct cause of the injury.” Export’s negligence in allowing the stowing of the carpet backing in Bangladesh without chocks admittedly was a proximate cause of the injury. But that does not preclude a finding of superseding cause. As Professor Prosser stated, “the problem is not one of causation at all, since it does not arise until causation is established. It is rather one of the policy as to imposing legal responsibility.” W. Prosser, Law of Torts § 44, at 270 (4th ed. 1971). This distinction is obvious from the language of the Restatement definition of superseding cause: “A superseding cause is an act of a third person or other force which by its intervention prevents the actor from being liable for harm to another which his antecedent negligence is a substantial factor in bringing about.” Restatement (Second) of Torts § 440 (1965) (emphasis added). The Restatement’s Comment on that section explains that “[a] superseding cause relieves the actor from liability, irrespective of whether his antecedent negligence was or was not a substantial factor in bringing about the harm.” Id., Comment b (emphasis added). The Comment further notes that if in looking back and tracing the sequence of events by which the harm was produced it is found that a superseding cause has operated, there is no need to determine “whether the actor’s antecedent conduct was or was not a substantial factor in bringing about the harm.” Id. Plainly the existence of proximate- cause does not preclude a jury’s finding that the stevedore’s negligence was a superseding cause. True, factors that indicate that something is a superseding cause may also indicate that proximate cause is absent. But I do not believe, as the district court apparently did, that in all situations the law of negligence has been reduced to the two words “proximate cause.” Therefore, I conclude that the district court’s analysis was erroneous.
The question must then be faced as to whether in this case, regardless of whether Export’s negligence was a proximate cause of DiRago’s injury, a jury could reasonably have found that I.T.O.’s negligence was a superseding cause of the accident. If a jury could so find, an instruction on superseding cause was required and the district court’s failure to give that charge was error. The majority, in holding that no such charge was required, necessarily found as a matter of law that the facts of the case could not support a jury finding that I.T.O.’s negligence was a superseding cause. I turn now to the majority’s analysis.
*872III.
I agree with the majority that the plaintiff satisfied the three-part threshold test for recovery of damages under the Longshoremen’s and Harbor Workers’ Compensation Act (LHWCA), 33 U.S.C. § 905(b) (1976), as established by this court in Griffith v. Wheeling-Pittsburgh Steel Corp., 610 F.2d 116 (3d Cir. 1979), appeal docketed, 48 U.S.L.W. 3374 (U.S. Dec. 4, 1979) (No. 79-813). In Griffith this court stated that section 905(b) “imposes on vessel owners the same duty to exercise ‘reasonable care un-, der the circumstances of each case’ that would be applicable to a land based business.” Id. at 125, quoting Kermarec v. Compagnie Generate Transatlantique, 358 U.S. 625, 632, 79 S.Ct. 406, 410, 3 L.Ed.2d 550 (1959).3 In Brown v. Ivarans Rederi A/S, 545 F.2d 854, 863 (3d Cir. 1976), cert. denied, 430 U.S. 969, 97 S.Ct. 1652, 52 L.Ed.2d 361 (1977), this court cited Restatement (Second) of Torts § 452 (1965), which defines when the negligence of third parties may be a superseding cause, as an example of one of the “principles of the law of negligence” applicable in LHWCA actions under 33 U.S.C. § 905(b). My sole, but important disagreement with the majority is that I believe section 452(2) of the Restatement should have been applied to the facts of this case.
The Restatement (Second) of Torts (1965) provides:
§ 452. Third Person’s Failure to Prevent Harm
(1) Except as stated in Subsection (2), the failure of a third person to act to prevent harm to another threatened by the actor’s negligent conduct is not a superseding cause of such harm.
(2) Where, because of lapse of time or otherwise, the duty to prevent harm to another threatened by the actor’s negligent conduct is found to have shifted from the actor to a third person, the failure of the third person to prevent such harm is a superseding cause.
Export’s position is that on its facts this case falls within subsection (2). The majority is correct that subsection (2) is the exception to the general rule. I also agree with the majority that LHWCA cases are not, as a class, to be treated as exceptional. But neither of these conclusions decides the only real question in this case: whether these particular facts are sufficient to require the instruction of the jury on superseding cause.
Although it may be “impossible to state any comprehensive rule” as to when responsibility for the prevention of harm has shifted to a third person, Restatement (Second) of Torts § 452, Comment f (1965), there are certain factors to be considered in making that decision:
Among them are the degree of danger and the magnitude of the risk of harm, the character and position of the third person who is to take the responsibility, his knowledge of the danger and the likelihood that he will or will not exercise proper care, his relation to the plaintiff or to the defendant, the lapse of time, and perhaps other considerations.
*873Id. The majority is correct that the general warranties given by I.T.O. to Export are insufficient to support a jury finding of superseding cause.4 However, I believe that the lapse of time between the discovery by the stevedore of the dangerous condition and the accident, the continuing failure of I.T.O. to use a safe method in the face of a known danger, and the inability of Export to correct the dangerous condition constitute facts that together warrant an instruction on superseding cause.5 In my view, the facts of this case, when evaluated according to the factors enumerated in the Restatement comment, indicate firmly the need for such an instruction.
The first factor, “the degree of danger and the magnitude of the risk of harm,” id., according to the majority, weighs against permitting a jury finding of superseding cause. See typescript at 12 n.5. I believe their “rough calculus,” id., is in error. Although the degree of danger from these unchocked rolls might be abnormal were they stacked in a public area — along a sidewalk or in a department store for example, here they were in a hold of a ship. The only persons present were the ship’s crew or employees of the independent stevedore contractor. All these persons were familiar with the handling of cargo. These stevedore employees are trained, seasoned cargo handlers, with all possibly necessary loading equipment available to them. These days cargo is often shifted in large containers, much heavier than carpet backing. At any rate, cargo in 1,500 pound units, as this was, is not abnormal in shipping. Further, DiRago introduced evidence for the expressed purpose of showing that longshoremen were capable of dealing with different types of cargo in the condition it was found. See Evans v. Transportation Maritime Mexicanna SS “Campeche”, 639 F.2d 848 at 859 (2d Cir. 1981) (“The stevedore is hired precisely for his expertise in a specialized and often dangerous trade.”). Thus, I conclude that although injuries from the heavy objects that longshoremen move are likely to be severe, the degree of danger posed by the carpet backing rolls for such workers is not abnormal and the magnitude of risk slight in light of their admitted skill. Therefore, this factor may actually support a superseding cause instruction.
The second factor, “the character and position of the third person who is to take the responsibility,” Restatement (Second) of Torts § 452, Comment f (1965), strongly supports the issuance of an instruction on superseding cause. The third person is I.T.O., hired by Export for the purpose of safely and efficiently unloading the cargo. I.T.O. is a professional stevedore company and as such is as familiar, and probably more familiar, with unloading procedures and risks than the vessel owner. See Evans v. Transportation Maritime Mexicana SS “Campeche”, supra, at 852 (“Since the stevedore is an independent contractor in control of the cargo operation and is hired specifically for his expertise, the stevedore is ordinarily in the best position to prevent accidents which might arise in the course of loading or unloading a ship.”). Further, as an employer, it should have a great interest in protecting its workers. Finally, once the cargo had been stowed in Bangladesh, I.T.O. alone was in a position to remedy the dangerous situation when the vessel arrived in Philadelphia. At that point, there was no way for Export to insert chocks. I.T.O., however, had full control of the situation and it could have used a safe method, either by tier-by-tier removal or by insertion of dunnage as chocks during the cut removal process. Unfortunately for DiRago, those safe methods were not as fast as the method of rolling them down, taking advantage of the lack of chocks. The nature and *874character of I.T.O.’s stevedoring operations and its knowledge of the known danger placed it in such a superior position that the issuance of an instruction on superseding cause was required.
The third factor, the third person’s “knowledge of the danger and the likelihood that he will or will not exercise proper care,” Restatement (Second) of Torts § 452, Comment f (1965), again favors the issuance of an instruction on superseding cause. Several hours before the accident, I.T.O. learned the stowage was unchocked, yet the I.T.O. crew did not alter their method of unloading. The lack of chocks was discovered early in the uneventful unloading of the port side and it was apparent again as unloading progressed on the starboard side, yet the stevedore took no precautions. I.T.O. had knowledge of the danger. I also believe it is proper to conclude that there was a likelihood that I.T.O. would exercise proper care. First, I.T.O. had the expertise and all of the necessary equipment of a stevedore and was hired with the expectation that proper care would be exercised. Second, I.T.O.’s own employees bore the risk from the danger and it owed them a special degree of care. The stevedore has primary responsibility for the safety of the longshoremen in its employ. Evans v. Transportation Maritime Mexicana SS “Campeche”, supra, at 852, 859-860. Finally, injury to the employees would be economically costly to I.T.O. The unloading would be delayed costing the stevedore money and the stevedore might incur losses for workmen’s compensation, if the injured longshoreman were unable subsequently to recoup his losses and those of the stevedore under the LHWCA.6 Thus, I believe there is an expectation that I.T.O. would exercise proper care, and again the factor supports a jury finding of superseding cause.
The fourth factor, the third person’s “relation to the plaintiff or the defendant,” Restatement (Second) of Torts § 452, Comment f (1965), has been noted already, and again strongly supports a superseding cause instruction. Export hired I.T.O. with the expectation that it would conduct its business expertly and safely. As the employer of the plaintiff DiRago, it had a duty to provide him with safe “places of employment”, 33 U.S.C. § 941(a) (1976), and with safe working procedures. See 29 C.F.R. §§ 1918.83(a) & (b) (1980).7 This factor also *875indicates that an instruction on superseding cause was required.
The final factor, “lapse of time,” Restatement (Second) of Torts § 452, Comment f (1965), is the weakest in terms of supporting a conclusion that a superseding cause instruction was required. While there was a significant lapse of time between Export’s negligence in Bangladesh and the accident in Philadelphia, there were only a few hours between the discovery by the stevedore of the dangerous condition and the accident. However, for three reasons, I believe sufficient time elapsed to warrant a superseding cause instruction. First, the stevedore had enough time to change its method; that would have taken no time. Second, a stevedore never has much time' between the arrival of a ship and the beginning of its operation. Shipping companies want to unload quickly and leave. At most a stevedore would have a few days notice, and I can find no meaningful distinction between a few hours and a few days in this setting. If the majority is correct that a great lapse of time between discovery by a stevedore and an accident is a prerequisite to superseding cause, then I can imagine no longshoreman’s case where it would be appropriate. Third, I.T.O. cannot complain about lack of time; even after the accident it continued to use the same method of unloading. It is unlikely that six months of accident-free unloading would have led I.T.O. to change its method. Whether or not sufficient lapse of time occurred to require by itself a superseding cause instruction, its combination with the other factors all indicate that such an instruction was necessary.8 Because I believe the failure to give such an instruction was reversible error, I would vacate the judgment and remand the case to the district court for a new trial.
IV.
The majority disparages Export’s argument that a superseding cause instruction was needed as an attempt “to avoid the impact” of Griffith, supra, and Edmonds v. Compagnie Generate Transatlantique, 443 U.S. 256, 99 S.Ct. 2753, 61 L.Ed.2d 521 (1979). But neither of those cases addressed the situation found here, nor will their holdings be eroded by the giving of instructions on superseding cause in cases such as this.
Griffith involved a barge with damaged hatch covers. The dangerous condition caused by the owner’s negligent failure to maintain the barge could have been corrected at anytime. Failure to correct that condition was ongoing, concurrent negligence. In this case, once the carpet backing had been stowed, there was no way to correct the condition. Rather the only remedy was careful unloading. Given that the stevedore and its employees knew of that condition, and that only they could take the steps to prevent harm from arising from that condition, an instruction on superseding cause was required.
The Supreme Court in Edmonds, supra, was not faced with a superseding cause situation. Edmonds was injured when, on the instruction of a member of the ship’s crew, he went behind a large cargo container to remove a jack. A co-worker of Edmonds then drove a truck into the container, causing it to roll backwards and pin Edmonds against the bulkhead. 443 U.S. at 274, 99 S.Ct. at 2764 (Blackmun, J., dissenting). The negligence of the shipowner, occurring immediately before the accident, *876was the crew member’s ordering of Edmonds into the dangerous position. The co-worker’s negligence had none of the attributes of a superseding cause, attributes that I believe are plain in this case. Thus, Edmonds does not suggest that the doctrine of superseding cause has no place in LHWCA eases. This is not surprising because in Edmonds the sole issue was apportionment of damages, not a question of when legal responsibility will or will not be imposed. See W. Prosser, Law of Torts § 44, at 270 (4th ed. 1971). Because the Court in Edmonds never considered the issue of superseding cause, nor was that issue presented on the facts, Edmonds cannot stand for the proposition that superseding cause instructions are precluded in LHWCA cases.
The majority cites Edmonds, 443 U.S. at 265 n.15, 99 S.Ct. at 2759 n.15, for the proposition that a vessel owner’s “principal defense,” typescript at 9, in a damage suit is absence of proximate cause. Note 15 states: “In many cases, of course, the shipowner whose act or omission contributed only a very small percentage of the total negligence will avoid liability on the ground of lack of causation.” There is nothing there to indicate that lack of causation is the “principal defense.” In saying that the question of proximate cause subsumes the issue of superseding cause, the majority makes the same error as that made by the district court. The two doctrines, at least in my understanding of the law of torts, have not yet been merged. If a vessel owner’s negligence is not a proximate cause, then a third party’s negligence is not a superseding cause for there is nothing to supersede. Rather, only when it is shown that a defendant’s negligence is a proximate cause, does the question of superseding cause arise.
V.
Because I believe that the district court relied on incorrect reasoning in denying the requested instruction, that no cases or statutes suggest that the superseding cause doctrine is inappropriate in LHWCA cases, and that the facts of this case could sufficiently support a jury finding that I.T.O.’s negligence was a superseding cause of DiRago’s injury, thus absolving Export of any liability for its negligence, I conclude that an instruction on superseding cause is required. Accordingly, I would reverse the judgment and remand^ the case for a new trial.
. Collins, the gang foreman, disputed DiRago’s testimony that the cut method was used in this way and to save time. Collins maintained that such a method would be unsafe. That was proved by this injury. Because such a method may violate OSHA regulations, it is not surprising that the stevedore supervisor would not want to admit the use of such a procedure. As is proper, I will assume the truth of the testimony of appellee DiRago, testimony corroborated by co-worker Pakech, and thus reject Collins’ version.
. Collins stated that a different type of forklift would have been needed to move the plywood. But that machine could have been brought on without approval of any officers of the vessel. Further, the need for other equipment was no less true when the request to move the plywood was made in the morning. Thus the requirement of additional equipment was not a significant obstacle to this safety precaution.
. This standard is currently under reexamination by the Supreme Court. Scindia Steam Navigation Co. v. De Los Santos, 446 U.S. 934, 100 S.Ct. 2150, 64 L.Ed.2d 786 (1980). In Scindia, the Ninth Circuit endorsed the “reasonable care under the circumstances” standard. 598 F.2d 480, 485-89 (9th Cir. 1979). In adopting that standard in Griffith, this court specifically approved Scindia. 610 F.2d at 125. This dissent is written with the assumption that the standard established in Scindia and Griffith is the correct one.
The Second Circuit, in a recent decision, Evans v. Transportation Maritime Mexicana SS “Campeche”, No. 80-7081 (2d Cir. Jan. 5, 1981), followed a different standard, based on Restatement (Second) of Torts § 343A (1965): “a vessel is not liable for injuries resulting from known or obvious dangers unless the ship owner should anticipate harm despite the obviousness of the danger.” Slip op. at 899. This court rejected that standard in Griffith. 610 F.2d at 125. This split between circuits may be resolved in Scindia. If the Second Circuit standard is adopted by the Supreme Court, then under Evans I believe that Export should not, as a matter of law, have been expected to anticipate the harm to DiRago from the negligent condition of the stowage, a danger known to both DiRago and the stevedore company, I.T.O. My analysis in this case, however, does not depend on such a change of law.
. The LHWCA voids any warranties by employees to vessel owners establishing employer liability for the negligence of vessel owners. 33 U.S.C. § 905(b) (1976).
. The majority considers the lapse of time, see typescript at 10, separately from the use of the unsafe procedure, see id. at 11. I believe a court normally would and should consider all facts together in deciding whether an instruction on superseding cause is needed. Specifically, although lapse of a few hours might, by itself, be insufficient, that lapse should not be ignored when it combines with a reckless failure to use safe procedures especially in the face of a known danger.
. As the majority notes, see typescript at 16, stevedore companies have judicially created lien rights in any recovery a longshoreman obtains in his suit against the vessel for the amount paid to the longshoreman by the stevedore as workmen’s compensation. Edmonds v. Compagnie Generale Transatlantique, 443 U.S. 256, 269-70, 99 S.Ct. 2753, 2760-61, 61 L.Ed.2d 521 (1979). See 33 U.S.C. § 933(f) (1976) (statutory compensation reduced by recovery).
Because of this lien, a stevedore may be able to avoid all or much of any financial loss, regardless of its negligence. I concur with Judge Friendly’s observation that
we are compelled by the combination of statute and Supreme Court decisions to sanction a result wherein if a properly charged jury were to find in [an injured longshoreman’s] favor in [an amount not significantly greater than the amount owed to the longshoreman by the stevedore as workmen’s compensation], the less negligent ship would bear the entire load, the more negligent employer would emerge scot-free, and [the longshoreman] would take nothing, so that the beneficiaries of the recovery would be [the longshoreman’s] lawyer and his employer’s compensation insurer. Our judicial system should not be obliged to countenance so palpably unjust a result, and the maritime industry should not be exposed to such costs.
Evans v. Transportation Maritime Mexicana SS “Campeche”, supra, slip op. at 916-17.
To me, the decision of the majority compounds the injustice of this situation. In holding that this evidence is insufficient to support a finding of superseding cause, I believe they are requiring the shipping industry to make conditions on its vessels foolproof. For an industry already beset with enormous problems, the cost of such foolproofing may be unbearable.
. I agree with the majority that the OSHA regulations do not relieve vessel owners of any pre-existing responsibilities or duties, 29 C.F.R. § 1918.2(b) (1980), but that does not mean they should be ignored. Rather, I believe these particular regulations merely codify preexisting duties of the stevedore and that violation of those duties, even before the OSHA, is relevant to whether the negligence of the stevedore was a superseding cause of a longshoreman’s injury. As the majority notes, see typescript at 10 n.3, the regulations are propeUevidence of the stevedore’s duties to its employees. To recognize this duty as a further factor supporting the *875grant of an instruction on superseding cause in no way “preempt[s]” common law negligence, as the majority would have us believe. See typescript at 9.
. The majority suggests that this case might be different if an express warning had been given. To me, that distinction is specious in a case such as this when the third person, a contractor and employer, owed duties to both defendant and plaintiff and had knowledge of the dangerous condition, yet chose to disregard it because of the economic incentive in using the fast, unsafe method made possible only by taking advantage of the very defect. The message to vessel owners by such a decision is plain: Expressly “warn” every stevedore of all known aspects of all cargo to be loaded or stowage to be unloaded. This seems to me to resurrect a perverse warranty system that section 905 meánt to abolish.