This is the second appeal to this court in a diversity case under the Pennsylvania wrongful death and survival acts, 12 P.S. § 1601 et seq., 20 P.S. § 320.601 et seq. The first appeal is reported at 199 F.2d 930. The decedent met his death in a highway construction accident as the result of a premature explosion of dynamite, manufactured by the defendant and sold to the decedent's employer, a road building contractor, for use in blasting. At the second trial a jury found the defendant manufacturer responsible for the accident. The defendant has appealed.

The decedent was a loader of blasting charges. On the day of the accident his employer, C. J. Langenfelder & Co., was building a portion of the Pennsylvania Turnpike. To bring the roadway to grade it was necessary to blast out some very hard rock. For this purpose some 300 boreholes were drilled at five foot intervals in a customary pattern. Late in the afternoon the loaders followed close behind the drillers. Approximately five minutes would elapse between the completion of a borehole and the start of its loading. The decedent had loaded a hole and was coiling an unconnected wire about his hand. Another hole was being drilled five or six feet away. Without apparent cause, the charge decedent had just loaded exploded injuring him fatally.

At the first trial the defendant's motion to dismiss, made at the close of the plaintiff's case, was granted. At both trials the plaintiff relied upon the opinion of a witness qualified as an expert to prove that the accident was caused by heat generated in the borehole in the normal course of drilling, or by vibration from the nearby drilling, or a combination of the two. We think now, as we did on the first appeal, that the testimony of plaintiff's expert, particularly his answer to a detailed hypothetical question, was sufficient to create a jury issue on the cause of the explosion, although defendant now makes a substantial showing in support of its opposing theory that a hot broken drill tip left in the hole caused the explosion. But in order to recover plaintiff had to do more than introduce enough evidence to go to the jury on the cause of the explosion. Admittedly, the dynamite was not defective. Thus, the situation was one which frequently arises in negligence cases of various types where the plaintiff must show a known or foreseeable hazard against which the defendant should have guarded, by warning or otherwise, but did not. Maize v. Atlantic Refining Co., 1945, 352 Pa. 51, 41 A.2d 850, 160 A.L.R. 449; cf. Brabazon v. Belships Co., 3 Cir., 1953, 202 F.2d 904.

It is also true that a manufacturer of a very dangerous substance is held to a high degree of care in guarding against injury to those to whom the substance is supplied. Maize v. Atlantic Refining Co., supra; Restatement of Torts, § 388. But beyond this generality, every case must turn upon a determination of what would be reasonable behavior by the defendant, considering his actual or imputed knowledge both of a danger and of a need to warn or otherwise protect users against that danger. Brillhart v. Edison Light & Power Co., 1951, 368 Pa. 307, 82 A.2d 44. Koelsch v. Philadelphia Co., 1893, 152 Pa. 355, 25 A. 522, 18 L.R.A. 759. A review of the record indicates that the plaintiff's theory of the case and the court's charge comprehended these principles. What we have to decide is whether on the present record these issues should have gone to the jury at all.

We first consider the hazard alleged to have been inherent in the loading of a blasting charge in a fresh borehole before dissipation of the heat normally generated by drilling in hard rock. On the first appeal, we believed that a jury might reasonably have concluded that the defendant should have given some warning of this danger because it appeared that blasting workers and those supervising their operations, both generally and in this very case, were uninformed of the heat hazard involved in the customary procedure of loading almost immediately after drilling. However, on the present record we reach a different conclusion.

First, it is noteworthy that there is no testimony and no basis for an inference that a dynamite charge has ever before been detonated by the heat which drilling generates in a borehole. To the contrary, there was affirmative evidence of witnesses of many years experience that they did not know of any such occurrence. Plaintiff relied exclusively upon the statement of a metallurgist, testifying as an expert, that in his opinion the heat generated by drilling in this case made the temperature in the hole about 200° F. at the time of the loading, and that this was sufficient to cause and did cause the explosion which occurred. From this alone the jury was asked to draw the inference that defendant should have anticipated this danger, and the conclusion that defendant should have warned of it. But to establish liability for failure to give a warning, plaintiff also had to prove that the user of the dynamite, in this case a contractor represented on the job by a blasting foreman, would not and did not have adequate information about the increased hazard which in the opinion of plaintiff's expert existed when the hole was loaded.

With this aspect of the plaintiff's responsibility of proof in mind we examine the evidence as to the user's information or lack of information as to the particular hazard. The key witness was the blasting foreman himself. Called by the plaintiff, he stated at the outset that he never had known of an explosion caused by the normal heat of drilling a borehole in rock. However, in this respect there is no claim that he was less informed than anyone else, including the manufacturer, because it is not contended that anyone ever before experienced a premature detonation of dynamite under circumstances such as plaintiff undertook to establish in this case. Beyond this the foreman also testified that he had been informed by people whom he considered reliable that drilling in hard rock might produce sufficient heat to cause a premature explosion if the newly drilled hole should be loaded with dynamite before cooling. He also stated his own understanding that a temperature of 200° F. would be too hot for the safe loading of a blasting charge. Finally, he said that he had instructed his men to test each hole for heat, albeit by the rather crude method of inserting a wooden pole the full depth of the hole, leaving it there for three or four minutes and estimating the temperature by observing and touching the pole when it was withdrawn.

As already stated, the foreman was the plaintiff's witness and the issue of the user's ignorance of danger was one on which plaintiff had the burden of proof. On this issue we have found nothing in the record beyond that which is outlined above. We conclude, therefore, that although the trial judge carefully included the user's ignorance of the hazard which caused the accident as an element of the case which the plaintiff had to establish in order to recover, there is no evidence that the user was uninformed of the heat hazard said to be incidental to the normal drilling of boreholes. To the contrary, the only affirmative evidence on this issue showed that the foreman had such information, believed the very temperature which plaintiff claims to have existed here would have been dangerous and directed such precautions as he deemed adequate to prevent the loading of dynamite in a dangerously hot hole. It follows that the defendant cannot be held liable if this accident was caused by disregard of the heat hazard thus known to have been incidental to drilling in hard rock.

But plaintiff's evidence comprehended other possibilities. Plaintiff's expert stated as his opinion that heat or vibration, or both, caused the accident. And the jury may have concluded that defendant should have warned of the vibration danger involved in drilling too close to a loaded hole. But that would not help plaintiff. For plaintiff's own evidence of the cause of the accident merely presented alternative possibilities with no basis for a reasoned choice among them. And we already have shown that on the present record defendant cannot be blamed for at least one of these alternatives.

In this situation the law of Pennsylvania is clear. It starts with the general proposition that the plaintiff bears the burden of proving that negligent conduct of the defendant was the proximate cause of plaintiff's injury. Erbe v. Philadelphia Rapid Transit Co., 1917, 256 Pa. 567, 100 A. 966; Sandt v. North Wales Foundry Co., 1906, 214 Pa. 215, 63 A. 596. But where plaintiff's evidence supports an inference that excludes liability of the defendant just as strongly as an inference that comprehends liability, a jury would have to engage in speculation and conjecture in order to attribute the accident to one rather than the other. Dangelo v. Pennsylvania R. Co., 1930, 301 Pa. 579, 152 A. 743. Where plaintiff thus leaves the jury to guess as between suggested causes, there can be no recovery. Fix v. Pennsylvania Power & Light Co., 1943, 346 Pa. 598, 31 A.2d 114; Houston v. Republican Athletic Ass'n, 1941, 343 Pa. 218, 22 A.2d 715; Anderson v. Reading Co., 1932, 306 Pa. 246, 159 A. 450; Gausman v. R. T. Pearson Co., 1925, 284 Pa. 348, 131 A. 247. This is just such a case.

We understand the difficult position of the trial judge trying scrupulously to avoid any course of action such as this court had denominated error on appeal from the judgment on directed verdict at the first trial. However, we are now satisfied that there is no evidence in the record before us sufficient to go to the jury on the issues discussed in this opinion. Defendant's motion for judgment notwithstanding the verdict should have been granted.

The judgment will be reversed.