The Lummus Company appeals from a judgment entered upon a $146,926 verdict for the plaintiff in a diversity action in the United States District Court for *869the .Southern District of New York. The appeal turns on whether there was sufficient evidence to support the jury’s finding that Lummus was negligent and that its negligence was the proximate cause of Lawrence Iacurci’s death. As we find that the evidence, under governing Pennsylvania law, was not sufficient to support the jury’s verdict, we reverse the judgment for the plaintiff and direct that judgment be entered in favor of the defendant Lummus Company.
Lummus was retained by the Beryllium Corporation, a producer of beryllium and beryllium products, to design and construct a new plant at Ashmore, Pennsylvania. Lummus did not itself manufacture the equipment for the plant but was responsible for its design and selection. One item of equipment, and the immediate cause of Iacurci’s death, was a skip hoist — a bucket on an inclined track — used to transport a mixture of beryllium and other chemicals between two operations in the plant.
At the lower end of the track is a pit, 8 feet long, 6 feet wide and 4 feet deep, where the bucket is loaded; at the other end, some 10 feet above and 5 feet to the side of the pit, the bucket is emptied into a briquette-making machine. Beryllium is a highly toxic material. In order to prevent contamination of the plant, the skip hoist machinery is enclosed, and the pit itself is roofed with steel plates, leaving an opening just large enough for the bucket to pass through. The pit may be entered only by removing some of the plates, and a person working there must wear a dust respirator, glasses and gloves.
No one is in the pit during normal operations; the skip hoist and other machinery are remotely controlled from an instrument panel, known as a Thayer control panel. The jury could find, however, that Lummus should have anticipated that someone would go into the pit twice a week to lubricate the machinery and also that it would be necessary to go into the pit in order to obtain samples of the mixture. Such samples would be required only in calibrating the machinery which measured out the chemicals and presumably would be taken only infrequently.
Lubrication of the skip hoist seems not to involve any particular danger. To take a sample from the bucket, however, a workman must bend over it so that his body is between it and the roof of the pit. Iacurci was working in this posi-, tion when the bucket was accidentally put in motion, crushing him against the steel plates.
The skip hoist is controlled by three switches, one on the Thayer control panel and two elsewhere. All three must be turned on in order to activate the machinery, and each switch can be padlocked in the off position. However, Lummus seems to have taken no steps to call the locking device to Beryllium’s attention beyond sending it the manufacturer’s general instructions, and warning tags rather than locks were supplied when the controls were installed.
The warning tags bore the words “Do Not Operate” in large letters. Prior to the accident Beryllium had put into effect safety regulations requiring that employees attach a tag to the switch controlling equipment on which they were working, and it instructed its employees generally in the use of the tags. In addition, appropriate posters were displayed in the plant. There was no direct evidence that Iacurci was instructed as to the use of the tags, and there even was dispute as to whether the safety regulations by their terms applied to a laborer such as he. However, the undisputed testimony of Robert Mensinger as to Iacurci’s statements just prior to the accident, quoted below, appears conclusively to establish at least that Iacurci was aware of the tagging-out system.
The accident occurred on September 28, 1957. Lummus then had only clerical employees at the plant. The skip hoist had been turned over to Beryllium approximately one month earlier, and Beryllium had accepted the entire plant by September 20. During the shakedown period, Beryllium had experienced some difficulty in attaining the proper propor*870tion of chemicals in the mixture. Iacurci, who was a laborer unassigned to any permanent department, was directed by Simon Morana, a supervisor, to remove the contents of the skip hoist for analysis. Iacurci proceeded to remove the plates covering the pit and descend into it.
Mensinger was employed by Beryllium as a lab technician at the time of the accident. He testified that just prior to the accident he had seen Iacurci leaning over the skip hoist bucket and had inquired whether the machinery had been tagged out. Iacurci replied, “Oh, hell, yeah, it must be.” Mensinger was not satisfied, however, and he told Iacurci, “Wait a minute, I want to go over and double-check the switch.” He started off for the main control panel but he was delayed on the way by a request to attend to some other business.
At the same time two other employees, Anthony Kasarda and William Hill, were working under the supervision of John Heffner on the briquetting device. It was necessary for them periodically to activate the skip hoist, and Heffner, unaware that Iacurci was working in the pit, gave Kasarda and Hill permission to turn it on. When the skip hoist jammed, they discovered that Iacurci had been pinned between the bucket and the metal top of the pit. He was rushed to a hospital but was dead on arrival.
On the record before us, there can be no doubt that Beryllium was negligent, and the jury so found in answering one of the court’s interrogatories. However, this suit was not brought against Beryllium,1 and the only question to be decided is whether Lummus should be held liable for lacurei’s death.
As we view plaintiff’s contentions in the light of the evidence, they offer two general theories as to Lummus’ negligence. The first is that Lummus was negligent in designing the skip hoist and controls. This theory was given to the jury in some detail on special interrogatories. They found Lummus negligent in designing the equipment, but of the five subquestions relating to the specific manner in which the design was faulty, they answered only one. They found that the design should have permitted removal of the bucket’s contents without requiring a workman to put his head over the bucket.2 In view of the court’s instructions, the jury’s failure to answer the other subquestions must be inter*871preted to mean that they found that defendant’s negligence had not been established in those other respects.
A second theory for recovery is that Lummus was negligent in failing to instruct Beryllium in the use of the locking device. None of the interrogatories dealt directly with this but, since plaintiff’s counsel made timely objection to that omission, we must consider whether the verdict in favor of the plaintiff could have been supported by findings under this theory.
, We hold that, as a matter of law, Lummus was not required to design the pit so that a workman could stand upright beside the bucket. We also hold that, even if it was negligent for Lummus not to call the locking device to Beryllium’s attention, this omission was not a cause of the accident and therefore is no basis for imposing liability. This suffices to dispose of the case, and we need not consider the jury’s findings on the issues of intervening cause and contributory negligence, all of which were favorable to the plaintiff. Accordingly, we reverse the judgment of the district court and direct that judgment be entered for the defendant.
I.
The Pennsylvania law as to manufacturers’ liability has expressly fol- ' lowed the Restatement of Torts.3 Where the alleged negligence lies in the design of the product, § 398 governs:
“A manufacturer of a chattel made under a plan or design which makes it dangerous for the uses for which it is manufactured is subject to liability to others whom he should expect to use the chattel lawfully or to be in the vicinity of its probable use for bodily harm caused by his failure to exercise reasonable care in the adoption of a safe plan or design.”
The question before us, however, is not the broad one whether the jury could have found that Lummus did not exercise reasonable care in designing the skip hoist but the relatively narrow question whether they could have found the design unreasonable in not allowing a workman to stand upright beside the bucket. In other interrogatories the jury was asked whether the design was unreasonable in failing to include various safety features suggested by plaintiff’s evidence. They did not answer these other interrogatories, and, in the light of the court’s instructions, we must take it that they found that Lummus’ negligence was not established in those respects. And the plaintiff’s evidence and argument do not suggest any respect in which the design might have been improper other than those covered by the interrogatories.
The only evidence favorable to the plaintiff with respect to design of the pit itself consisted of brief statements by two experts. The defendant’s witness, Herbert I. Hollander, .conceded that the pit would have been safer had a workman been able to stand upright beside the bucket. Alfred F. Buff, testifying for the plaintiff, apparently assented to the description of the pit by plaintiff’s counsel as “a dangerous condition”; however, Mr. Buff’s recommendations for making the skip hoist safer included no suggestion as to construction of the pit.
That the design of the pit created a danger for a person emptying the bucket cannot be doubted; Iacurci’s death is conclusive evidence of this. However, it does not necessarily follow that Lummus was negligent, and we do not think that there was an adequate basis for the jury’s finding that it did not exercise reasonable care in designing the pit.
First, the danger was obvious. It required no knowledge of electrical or mechanical engineering to perceive the dan*872ger in wedging one’s body between fixed steel plates and a movable 450 pound bucket or to realize that safety lay in insuring that the bucket would not be put in motion. We do not conclude that, under Pennsylvania law, a manufacturer need provide safety devices only where the danger is latent; Pennsylvania appears not to have drawn any sharp line between obvious and latent dangers. But the obviousness of the danger is one factor in determining negligence; a foolproof safety device is less likely to be required if the danger is obvious and the manufacturer can expect some cooperation from the user in mitigating the danger. The risk in this case was apparent, and it was proper for Lummus to design the machinery in the light of that fact.
Second, the pit was not normally a place of work. At most, employees were to go into it twice a week to lubricate the machinery and less frequently to sample the mixture, and only the latter appears to involve a risk that would be substantially lessened by allowing a workman to stand upright. Just as a grade crossing may suffice .for a country road but be inadequate for an interstate highway, the safety devices guarding an area used only for occasional maintenance may reasonably requix*e the worker’s eoopex-ation even though automatic devices, or the elimination of the danger altogether, are considered necessary on the production line.
Third, the locking device, if used, would have protected the workman as effectively as deepening the pit. (Indeed, the plaintiff’s own expert, Mr. Buff, recommended only more effective safety switches.) Where the danger was obvious and arose only under exceptional cix’cumstances, we believe that Lummus could design the pit on the assumption that the locking device would be used. Given that assumption, it was not negligent in designing the pit as it did.
II.
The second theory suggested by the plaintiff’s evidence and argument before the district court is that Lummus did not take reasonable steps to inform Beryllium of the importance of using locks with the switches. Plaintiff may not recover on this theory because Lummus’ omissions in this respect, whether or not they were negligent, were not the cause of Iacurci’s death. Had the question been put to the jury, they could have found that Lummus made no effort to inform Beryllium of the locking device beyond supplying it with the manufacturer’s general instructions and that the equipment was installed with warning tags rather than locks attached. Although Lummus did not itself handle the installation, it presumably knew of it and was responsible for it. Since defendant’s own argument was in part directed at establishing that locks were a satisfactory substitute for other safety devices, there may have been an adequate basis for finding Lummus negligent in this respect, and for purposes of this discussion, we assume that there was.
But, leaving aside the rules governing the liability of concurrent wrongdoers, it is elementary tort law that even a negligent defendant will not be liable if the accident would have occurred absent his negligence. If Lummus failed adequately to inform Beryllium of the importance of using locks, the effect of this failure did not extend beyond causing Beryllium to use tags instead. Plaintiff thus is entitled to recover under this second theory only if the jury could properly have found that the use of locks leather than tags would have prevented the accident.
The plaintiff’s dilemma is this: She must show that the jury could have found that the machine was tagged by or for Iacurci, for even a lock is effective only if used. But she must reconcile this putative use of the tag with the fact that someone did turn the switch.
The difficulty of plaintiff’s position can be seen by comparing this issue with the issue of contributory negligence. In finding Iacurci not contributorily negligent, the jury might have believed that he had not been properly instructed in *873the tagging-out system; that he had asked another employee to tag the machine and the latter failed to do so; that the machine had already been tagged and Iacurci reasonably but mistakenly did not think it necessary to add his own tag; or even if he had negligently failed to tag the machine, that this negligence was superseded by reasonable reliance on Mensinger’s correcting the situation. But in each case the decedent’s due care would have been no less futile if locks rather than tags had been in use.
Only two hypotheses avoid plaintiff’s dilemma: The switch was turned despite the presence of a tag, or the tag had fallen off or been removed. There was no evidence directly supporting either hypothesis, and that someone would overlook or maliciously disregard a conspicuous warning tag is too unlikely to be inferred from the fact of accident alone; this is equally true of removing another employee’s tag.
The remaining possibility, that the tag had simply fallen off, requires more detailed consideration of the switch mechanism. The tags were to be hung from the same notch in the switch handle that would have been used for a lock. The notch itself formed three sides of a square, and plaintiff’s counsel presumably thought that the fourth side was open when he argued that the tag might have fallen off. A replica of the switch, however — introduced only at the close of the trial — reveals that the fourth side of the square is in fact closed off by a slide mechanism, and in no position does this slide leave a gap through which the string of a tag might slip. Thus, the tag could fall off only if the string somehow came untied. This seems at least as unlikely as the possibility that someone deliberately removed the tag or turned a tagged switch, and the sum of these possibilities, unsupported by anything more than speculation from the fact of the accident, does not provide an adequate basis for finding that the use of tags rather than locks was a cause of Iacurci’s death.
The judgment of the district court is reversed; the distinct court is directed to’ enter judgment for the defendant.
. No common law action could have been brought against lacurei’s employer, Beryllium, if ’ Iacurci was covered by the Pennsylvania workmen’s compensation statute. Jackson v. Gleason, 320 Pa. 545, 182 A. 498 (1936).
. “1. Are you persuaded by a fair preponderance i of' the credible evidence that the defendant," The Lummus Company, negligently designed the skip hoist pit and the location for the switches serving the skip hoist at the Beryllium Corporation’s plant at Hazleton, Pennsylvania?
X
YES NO
“If your answer to Question 1 is in the affirmative, please indicate which, if any, of the following findings you have made in reaching this conclusion:
“(a) There should have been a control button in the pit.
YES NO
“(b) There should have been an automatic interlock on the cover of the pit which would have cut off the power to the skip hoist.
YES NO
“(e) There should have been an unobstructed view between the controls for the skip hoist and the skip hoist bucket.
YES NO
“(d) The electrical controls for the skip hoist equipment should have included locks so as to permit the locking of switches in a closed position.
YES NO
“(e) The contents of the skip hoist bucket should have been removable without the necessity for a person to put his head over the bucket and into the skip hoist shaft.
X
YES NO”
. See Mannsz v. Macwhyte Co., 155 F.2d 445, 450 (3 Cir. 1946) and the cases cited therein. ' Although the court was concerned only with § 395, § 398 is merely a special application of the former section. Restatement, Torts § 398, comment a.