OPINION OF THE COURT.

McENTEE, Circuit Judge.

This is a diversity suit for injuries sustained by plaintiff as the result of an explosion that occurred on June 22, 1962, in the sander dust disposal system at the Plywood Products plant in North Stratford, New Hampshire. 1 The defendant, M. C. Carlisle & Co., Inc. is a Boston sheet metal work firm that deals especially in dust collection systems. Plaintiff, a Plywood employee, charges that this explosion and his consequent injuries were caused by Carlisle’s negligence 2 in designing, manufacturing and installing the disposal system in which the explosion occurred. The case was tried to a judge and jury. After plaintiff had put in his case on the issue of liability, the court directed a verdict for the defendant and entered judgment dismissing the suit.

Plaintiff contends that on the evidence presented, he was entitled to go to the jury on this issue — hence the trial court erred in directing the verdict. This is the only question raised in his appeal.

The operation of the dust disposal system in question revolves principally *950 around a blower (enclosed fan), a separator and a downspout or tail pipe that goes into a furnace where the dust is burned. 3 Two other devices should be mentioned. At the point where the feeding duet enters the separator there is a weighted damper that is supposed to close and seal off the feeding duct when the blower is not in operation. On the tail pipe, about ten feet up from the floor, there is a ball and socket joint that enables the pipe to be swung away from the fire holes, a distance of three or four feet. 4

First, let us review the evidence with reference to the work Carlisle did on this system. This firm became involved in Plywood’s wood waste problems in 1955 when it was called in to work on the system in which this explosion later occurred. 5 Almost every year since then defendant has been involved in work on wood disposal systems at the Plywood plant. An employee who has worked for Plywood since 1932 testified that the first job Carlisle had at the plant was to take out the Bannister blower system and replace it with a larger system. He said that from then on this firm did all the blower work. This included the installation of the tail pipes that were in the plant when this accident happened in 1962 as well as the blowers and the duct work to take care of the sander dust.

The maintenance foreman, who has also been employed at the plant for many years, said that Carlisle installed the balance of the sander system of which the Bannister blower was a part; that this firm also installed all the tail pipes that were in the boiler room in 1962; that it put in the first sanding equipment in the plant; did all the duct work in the woodworking system from 1955 on and replaced the separator after this explosion. He added that Lavin, the president of Carlisle, and his son came to the plant from time to time and gave advice on duct work.

A third Plywood employee testified that he was present when Carlisle installed the tail pipes that were in the system in 1962 when plaintiff got hurt. Also, that Carlisle’s men repaired the separator after an explosion in March 1962.

Lavin testified that the work of designing dust collection systems in the Carlisle firm was done by him or under his supervision. He denied that Carlisle had designed the sander system in question but admitted that in his deposition, taken prior to trial, he stated that Plywood had told him what they wanted with reference to this particular system; that on that information he prepared plans and recommendations of what should be done and that he designed the system. At the trial he tried to explain this by saying that in 1955 Carlisle worked on two systems at the plant; that he designed one of them in its entirety except for the tail pipe but that on the system in question he designed the suction side only and not the blower, the separator or the tail pipe. Although he admitted that Carlisle supplied and installed all the tail pipes in use at the plant at the time plaintiff was injured, he stated that these tail pipes were designed and requested by Plywood. Also, Lavin confirmed the testimony given in his deposition that his firm manufactured and installed the parts that went into the sander system in question; that he recom *951 mended the installation of the ball and socket joint in the tail pipe and also the installation of the fire damper near the entrance to the separator.

At the trial Lavin also testified that Carlisle did most of the duct work and the wood disposal work at the plant between 1955 and 1962. He further stated that most of this was replacement work.

Carlisle’s employee Kamilewiez, who spent considerable time at the Plywood plant, 6 testified that his company’s first job at Plywood was putting in “the sanding system — sanding dust collector” in 1955 and that it was a full installation, 7 He indicated that in 1955 Carlisle replaced the separator and the duct work running from the separator to the fire box but only to the roof line. 8 He also testified that the entire system that is in the plant now is what Carlisle put in in 1955, the only exception being that Plywood has more sanders there now than it had at that time.

At the time this explosion occurred the sander dust system had been shut down for at least three days for repairs. The Dutch ovens were in operation and the tail pipe on this system had been swung away from the fire hole. Plaintiff and two other members of the maintenance crew were on the platform in the boiler room working on the blower. 9 Explosions in this plant were not uncommon. Over the years there had been a number of small ones in the separators— one about every six months. 10 As recently as March 1962 there was an explosion, in this system that damaged the separator and the weighted damper. 11 Also it should be noted that over the years numerous small fires were burning on the boiler room floor “most all the time” and were allowed to burn themselves out. 12

The specific negligence alleged is that in designing, manufacturing and installing the tail pipe, Carlisle failed to equip this pipe with proper safety devices. 13 Furthermore, that Carlisle did not use reasonable care in designing and installing this system because the fire damper installed at the entrance of the feeding duct into the separator was not sufficient to prevent flames from travelling to places where, as here, workmen were likely to be located.

Plaintiff’s experts testified that when the system was shut down and the tail pipe was out of the fire hole, a draft was created in the tail pipe; that this draft sucked burning embers from the boiler room up through the tail pipe and into the separator where they ignited a highly combustible mixture of air and fine dust that had collected there — and that this caused the explosion. One of plaintiff’s experts also testifed that the flames from the explosion flashed through the weighted damper (fire damper) at the entrance to the separator and back through the eye of the blower where plaintiff was working at the time. He further stated that blowing the dust directly into the Dutch oven, as was *952 done here, is “absolutely wrong”; that “it is standard practice all over the world” to have a storage tank 14 between the separator and the furnace. This expert recommended that in the existing system there should be two check valves on the tail pipe, one near the bottom of the separator and the other down from the ball and socket joint. He said these valves would prevent sparks from going up the tail pipe 15 and would have prevented this accident.

This expert also testified that the weighted damper at the entrance to the separator was not sufficient to prevent the flames from the explosion from travelling back through the system and out the eye of the blower because it was not tight. There was evidence that Lavin recommended the use of this damper as a safety device. In the opinion of plaintiff’s expert a rotary valve instead of the fire damper at the point where the feeding duct enters the separator would have prevented the accident because “it’s more accurate and shuts off better.” He also testified that such valves are in common use.

There is no dispute that in determining the question of liability in this case the substantive law of New Hampshire must be applied. In Russell v. Arthur Whitcomb, Inc., 100 N.H. 171, 121 A.2d 781 (1956), the New Hampshire Supreme Court adopted the view:

“that independent building and construction contractors should be held to a general standard of reasonable care for the protection of third parties who may be foreseeably endangered by the contractor’s negligence even after acceptance of the work. This rule is subject to the following qualification: ‘ * * * The employer’s failure to discover the defect will not relieve the contractor of liability; but * * * if he discovers the danger, or it is obvious to him, his responsibility supersedes that of the contractor.’ * * * ”

This was a case where a landowner sued a road contractor who constructed a sewer line for the city. Plaintiff alleged that in doing so defendant negligently failed to refill an excavation which he made within a few feet of plaintiff’s property which resulted in damage to his building. The New Hampshire Supreme Court said that in the absence of evidence that the defect in the highway became obvious to the city or was discovered by it, plaintiff was entitled to have the jury consider whether the work of the defendant contractor caused the damage to plaintiff’s building. 16

Applying the law enunciated in Russell to the facts in the instant case, Carlisle is held to a general standard of reasonable care for the protection of employees in the Plywood plant who may be foreseeably endangered by Carlisle’s negligence. Although failure on the part of Plywood to discover any defects in this system 17 would not relieve Carlisle of liability, 18 if there was evidence that Plywood had actually discovered the danger 19 or if there was evidence that the *953 danger was obvious to Plywood, then Carlisle’s responsibility is superseded by that of Plywood. In the absence of evidence that the danger became obvious to Plywood or that Plywood discovered it, plaintiff is entitled to have the jury consider whether the work done by Carlisle caused the injuries to the plaintiff 20

We now turn to the final question of whether the evidence presented by the plaintiff in the instant case was sufficient to entitle him to go to the jury on the liability issue. In this circuit a scintilla of evidence is not enough. Magnat Corporation v. B & B Electroplating Co., 358 F.2d 794 (1st Cir. 1966). Conclusions of fact must be supported by substantial evidence. This means such relevant evidence as a reasonable mind might accept as adequate to support a conclusion. United States v. Krumsiek, 111 F.2d 74, 78 (1st Cir. 1940). In Rainey v. Gay’s Express, Inc., 275 F.2d 450, 451 (1st Cir. 1960), we quoted with approval the principle to be applied by this court in reviewing the record on the question of the sufficiency of evidence on a motion for a directed verdict. This principle was stated as follows:

“ ‘ * * * In determining whether or not the evidence in a given case is sufficient to take the case to the jury over a motion for directed verdict, the evidence must be viewed in the light most favorable to the plaintiff, giving the plaintiff the benefit of every inference favorable to him which may be fairly drawn. It is not for the court to weigh the conflicting evidence or to judge the credibility of witnesses. Whenever the evidence is such that fair-minded men may draw different inferences therefrom, and reasonably disagree as to what the verdict should be, the matter is one for the jury. * * * > »

Initially, we find that there is evidence that Plywood had actual knowledge at the time of the accident of the danger resulting from the alleged defect in the tail pipe. 21 Thus, under the rule in Russell, Carlisle’s responsibility with reference to this particular alleged defect was superseded by that of Plywood.

There is no evidence, however, that Plywood knew of the danger caused by the alleged deficiency in the damper or that this danger or alleged defect had become obvious to Plywood at or before the time of this accident. Therefore, we cannot say that Carlisle’s responsibility was superseded by that of Plywood with reference to this alleged deficiency.

Applying the test set forth in Rainey, supra, it is our opinon there is sufficient evidence to go to the jury on the issue of whether Carlisle was negligent in designing, manufacturing and installing the damper and that this negligence was the proximate cause of plaintiff’s injuries. Carlisle’s president, Lav-in and its employee, Kamilewicz both admitted that Carlisle designed, manufactured and installed the damper, the separator and the ducts running to and from the separator. Nor is there any question that Carlisle designed, manufac *954 tured and installed the system above the roof line or that the explosion which caused plaintiff’s injuries occurred in the separator. 22 There is substantial evidence that the flames caused by the explosion travelled back through the system from the separator, through the damper and out the eye of the blower where plaintiff was injured. This would indicate that the damper may have been deficient as a safety device. 23 Plaintiff’s expert testimony supports the contention that this damper was not an adequate safety device and that the use of a rotary valve at this point would have prevented this accident.

New Hampshire has adopted the Restatement Rule, 24 that negligent conduct is a proximate or legal cause of harm, if the defendant’s conduct is a substantial factor in bringing about the harm and if there is no rule of law relieving the defendant from liability because of the manner in which his negligence has resulted in the harm. Maxfield v. Maxfield, 102 N.H. 101, 151 A.2d 226 (1959).

The only pertinent rule of law that could relieve the defendant here is the foreseeability rule. 25 On the evidence before us, we think a jury could reasonably find that plaintiff was fore•seeably endangered by this alleged defect in the dust collecting system. Defendant is not excused because he could not foresee the exact manner in which the injury to plaintiff might occur.

New Hampshire has consistently taken the position that the question of whether the defendant’s conduct was a substantial factor in bringing about the harm is a question for the jury. Maxfield v. Maxfield, supra. It is our opinion that on the evidence in this case the jury could reasonably find that the designing and installation of this damper was a substantial factor in bringing about the injury to this plaintiff. See New England Tel. & Tel. Co. v. Reed, 336 F.2d 90, 95 (1st Cir. 1964).

The judgment of the district court will be vacated and the case remanded for a new trial consistent with this opinion.

Notes

. Plywood Products was a woodworking firm and as an essential part of its operation had several wood waste disposal systems in its plant. The one involved in this case was designed to collect and dispose of dust resulting from wood sanding, i. e., the dust made by the sanding machine in processing the wood.

. The specific negligent acts and omissions upon which plaintiff relies are discussed later in this opinion.

. The dust is pulled upward from the sanding apparatus through a suction hood and a metal duct into the blower located on a platform about sixteen feet up from the boiler room floor. The blower propels the dust through another duct into the separator which is' mounted on the roof. There the air is separated from the dust. The air escapes through the top of the separator and the dust goes down from the bottom of the separator through the tail pipe, supposedly by the force of gravity, into a Dutch oven in the boiler room below. There are three Dutch ovens. At the time of this accident tail pipes from two other disposal systems ran into the other ovens.

. The tail pipe is supposed to be removed and swung away from the fire holes when the system is not in operation.

. This system had been installed originally by another Boston company named Bannister.

. He stated that except for one year he did work at the Plywood plant once or twice a year from 1955 to 1962 for a period of two, three or maybe four weeks at a time. In fact, he was at the plant working on another system when this explosion occurred.

. Later he stated that in speaking of full installation he referred to the separator as distinguished from the entire system.

. He admitted that since 1955, Carlisle replaced the tail pipe from the roof line down to the fire box itself.

. Flames flashed through the eye of the blower, engulfed the plaintiff and he was severely burned.

. These explosions always occurred when the tail pipe was put into the fire hole before the blower was started.

. But no previous explosion was as violent as the instant one.

. Frequently lighted embers popped out of the fire holes and ignited wood waste that had been dropped on the floor. This waste consisted of chips and long strips of wood that were fed into the Dutch ovens through the fire holes.

. The tail pipe which is about forty feet long is completely open. There is no damper or other device on this duct which would close it off when there was no downward movement in it.

. There was no storage tank in this system.

. According to the expert this is more apt to happen where, as here, there is an open top Dutch oven. Lavin testified that he told Plywood many times to keep the end of the tail pipe away from the boilers when the system was not in use and agreed that fine dust in these collecting systems is a highly combustible substance.

. The test as stated in Russell is an objective one. The evidence would have to show actual discovery of the defect or that it was in fact obvious to the contractor. It is not sufficient that the defect in question would be discovered by or was obvious to a “reasonable man.”

. The alleged defects consisted of (1) the open tail pipe, and (2) use of a simple weighted damper (rather than an electrically controlled rotary valve) where the feeding duct entered the separator.

. This is because the failure of Plywood to discover the defects is within the foreseeable risk of the contractor Carlisle.

. It would appear that in Russell “danger” and “defects” are used interchangeably. However, the “danger” is the foreseeable risk which exists as a result of the defect. For example, in the instant case *953 one of the alleged defects was the open tail pipe. The danger inherent in such a tail pipe is that sparks from the boiler room may freely travel the length of the pipe and enter the separator.

. Although couched in terms of “duty” the Russell case really turns on the question of causation.

. Plywood knew, for example, of the numerous explosions in the separator over the years which occurred about every six months and of the damage therefrom. This is especially true with reference to a violent explosion that occurred in this system only two months prior to this one, in which the separator and the damper were badly damaged. It also knew that these explosions were caused by sparks going up the open tail pipe and into the separator when the system was not in operation. It was common knowledge in the plant that numerous small fires were allowed to burn almost constantly on the boiler room floor and that frequently lighted embers popped out of the fire holes near the end of the tail pipe. See footnotes 4, 12 and 15. Furthermore, it should be noted that there is specific evidence that this tail pipe was designed and requested by Plywood — not Carlisle.

. The fact that the inside and the roof of the separator were damaged and its sides split open indicates that the explosion occurred there.

. There was evidence that the damper was damaged in the explosion of two months earlier and that it was repaired, straightened and put back at the entrance to the separator.

. Restatement (Second), Torts, § 431(a), (b) (1965).

. We have already considered the question of whether Carlisle’s responsibility was superseded by that of Plywood.