In the early morning hours of February 5, 1965, a house occupied by a family named Sugarman, whose personal representative is the plaintiff, was partially consumed by fire and the entire family burned or suffocated. Plaintiff's suit was commenced in the Massachusetts state court and removed to the district court on the ground of diversity. The now remaining defendants are appellees Rockwell Manufacturing Company and Worcester Gas Light Company, hereinafter Rockwell and the gas company. The case was tried to a jury. At the close of the plaintiff's evidence the court directed verdicts for these defendants, and plaintiff appealed.

The house had been built the previous year. Its furnace (known in the trade as a "boiler") heated water in a multiairtube heat exchanger and was gas fired, operating at 3½ inch (of water) gas pressure. The gas came in from the main at a 50-pound pressure, and was reduced to a pressure of 7 inches by a diaphragm reduction valve known as a Reynolds regulator, manufactured by Rockwell, and then to 3½ inches by a device of another manufacturer, Honeywell, not a party. Neither defendant had furnished nor installed the furnace, but the gas company had installed the Reynolds regulator, the meter and some of the line. It was plaintiff's theory of liability that the regulator had foreign matter in it, some of which was introduced during Rockwell's initial assembly, and some attributable to the gas company; that on servicing the furnace the gas company should have discovered the fault; and that this fault led to sporadic overgassing, which caused the fire. Plaintiff objected to the exclusion of various items of evidence as well as to the directing of the verdicts.

This is a factually complicated case. We may say at the outset that our path has been made more difficult by being continually obliged to refer back and forth between two substantial record appendices, a matter of which we complained, but apparently to no avail, in Moran Towing Corp. v. M. A. Gammino Construction Co., 1 Cir., 1966, 363 F.2d 108. In that case we did not assess the fault. In this case the fault is largely the plaintiff's, whose aversion to everything her witnesses said on cross-examination,1 and in some instances even to important parts of the direct, eventually required us, in order to obtain continuity, to read the entire original transcript.

The Sugarman home was a single family frame ranch-type house with no cellar. The furnace and some other appliances were located in an unfinished utility room. On October 6, 1964 the gas company "fired" the furnace. On November 26 a combustion odor was noted coming from the furnace. The gas was shut off, and on November 27 the furnace was cleaned by the gas company and a quantity of loose carbon removed that "obviously" indicated that "something was causing such a build-up."2 No odor from the furnace was ever noted again. Around the first of the year Sugarman's father-in-law installed outlets and electric wiring in the utility room. He was not a licensed electrician, and the wiring was not inspected. The jury could have found, however, that the work was competently done.

The fire of February 5 was a substantial one. One hundred pounds p. s. i. of water pressure was used in extinguishing it. The upstairs timbers fell into the utility room, knocking down the galvanized flue that ran from the furnace to the chimney and bending various pipes, including the one to which the regulator was affixed. The fire marshal was of the opinion that the fire started in or near the utility room, but could form no opinion as to its cause. Plaintiff conceded that there were no signs of an explosion.

Plaintiff's first complaint is that on November 27 the gas company serviceman, having discovered that some defect was producing the carbon deposit, should have done something about it. The witness testified that he did; that there was a slipped collar at the top of the furnace and that he straightened its legs and raised and fastened it with screws and that he cleaned the pilot, which was burning yellow from "building dirt" in its air passage. The pamphlet furnished with the furnace described the pilot as having no air passage. The furnace man testified that this model furnace did not have that kind of collar, and a witness who examined the collar after the fire testified that it was welded and that the welds were intact. In addition, the serviceman's work notes merely reported "cleaned boiler," and he acknowledged that any repairing should have been separately stated.

One Donohoe testified that he was a licensed plumber, a licensed gas fitter and a registered engineer who had been a Civil Service Examiner for inspectors and engineers; that he had had experience with gas regulators; that on February 8 he visited the Sugarman house; that the Reynolds regulator was still affixed at that time to the main and to the pipe which led to the meter; that there was no break in the pipe up to the point where the meter had been, and that there the end was turned down, but open; that on February 25 he removed the regulator, not allowing any dirt to get in, and that he wrapped it in a blanket and took it home in his car, where he wrapped it in polyethylene and stored it in a closet until about January 1966, when he placed it in a vise and took it apart and debris fell out. Some of the debris was offered in evidence and excluded. Donohoe testified that this debris did not enter the valve portion of the regulator before the fire. The court ruled his evidence insufficient to warrant such a finding. We have reviewed this testimony in detail, and are abundantly satisfied that the ruling was correct, even if some of the individual reasons given by the court and vigorously attacked here may have been incorrect.3 During the fire the regulator was subjected to enormous insult, to adopt a word used during the trial. With two openings in its top, see fn. 1, supra, it was found almost sitting amid extensive debris. In the meantime the heat, presumably before water reached it, had destroyed the diaphragm seal. With its integrity gone, and the gas outlet pipe open at the far end, anything could have entered the regulator and been moved about inside it during the fire. Although plaintiff's expert testified that this was a possibility only, and was not probable, we are compelled to find that there was no logical basis at this stage of the case for drawing this distinction.

Proof that debris could have entered the regulator during the fire did not foreclose the plaintiff from showing that in fact debris had been present before. She sought to establish its prior existence as a reasonable inference from other circumstances. First (by which we do not mean in the order of testimony) plaintiff produced evidence that tended to show4 that the November carbon deposit was due to overgassing, viz., gas reaching the burner at too high pressure. Higher pressure raises the flame, and if sufficiently, the flame makes contact with the metal above, which, in turn, produces carbon. Overgassing can result from malfunctioning of a regulator. It is true that the gas company serviceman testified that the November carbon was due to another cause or causes, but his testimony was substantially impeached, even in addition to the matters already mentioned. The jury would have been warranted in finding that these other causes did not exist and, by the same token, that the witness did nothing to perform a conceded duty.

Next, plaintiff produced evidence tending to show that there was also overgassing the night of the fire. This was a more complicated matter, requiring the testimony of several witnesses. Defendants' attempt to weaken their testimony by lengthy cross-examination designed to establish that, taken singly, their testimony was insufficient, in no way destroyed its cumulative effect. If this testimony was accepted it adequately warranted a finding that there was overgassing the night of the fire, before the fire.5

The question whether overgassing caused the fire has one troublesome aspect. If this case turned on the court's rejection of the testimony of the witness who "reconstructed" the metal flue from the furnace to the brick chimney, we might consider the testimony so weak as to show no error.6 However, there was other evidence of erroneous placement which we believe could have warranted a finding that the overheated flue caused the fire. This being so, we are unimpressed by defendants' indication of other possible sources of the fire. The burden that was on the plaintiff to show that the overgassing was the probable source of the fire did not require proof that it was the only possible source. Gates v. Boston & Maine R. R., 1926, 255 Mass. 297, 151 N.E. 320; Brooks v. Kinsley Iron & Machine Co., 1909, 202 Mass. 228, 88 N.E. 771.

This brings us to the final question of whether defendants are chargeable for the overgassing, assuming there was such. Plaintiff's evidence was that the overgassing was sporadic, and that this would be the reasonable result of debris in the regulator which would, or would not, depending upon where from time to time the gas circulation was placing it, interfere with the closing of the valve. Our study of this evidence satisfies us that the jury could have accepted it and have concluded, in other words, that the debris preceded the fire. On the other hand, the claim against Rockwell required a finding that the debris was present within the regulator at the time of its manufacture. While the size of some pieces of the debris precluded a finding that they came from the gas main, for reasons already given there was no adequate basis for a finding that these particular pieces did not enter at the time of the fire. Because of this, and also because there was evidence of debris in the Honeywell regulator,7 it could not be found that a defect attributable to Rockwell caused the fire.

The direction of a verdict for Rockwell gives us no pause. So far as the gas company is concerned, the jury could have found that its serviceman failed in his conceded duty to discover and remedy the cause of the carbon found in November. Compare Wadsworth v. Boston Gas Co., Mass., 1967, 223 N.E.2d 807; Trimbo v. Minnesota Valley Natural Gas Co., 1961, 260 Minn. 386, 110 N.W.2d 168. The fact that no recurrence of the odor was noted during the ensuing ten weeks before the overgassing, which could have been found to have preceded the fire, does not conclusively prove that there was not sporadic overgassing. Furthermore, in addition to this neglect, if it were such, we must suppose that a gas company could be found liable for furnishing gas with sufficient impurities to interfere with the operation of a regulator. The company has made no suggestion that it would not be. The direction for the gas company was error.

This decision requires us to consider the court's exclusion of some of the evidence offered to show conscious suffering on the part of members of the Sugarman family. Without ruling on the specific questions excluded, at least some of which seem clearly improper, we would think that a medical examiner could, under certain limits, express an opinion whether asphyxiation would be preceded by pain, or fright, and whether, if a deceased party were in a place where the jury concluded he or she was not before the fire, this movement was probably conscious. Cf. Campbell v. Romanos, 1963, 346 Mass. 361, 367, 191 N.E.2d 764; Markell v. Gahm, 1962, 343 Mass. 468, 470-471, 179 N.E.2d 587.

Judgment will be entered affirming the judgment for defendant Rockwell,8 but vacating the judgment for Worcester Gas Light Company and ordering a new trial against that defendant not inconsistent herewith.