This is an action of tort to recover for personal injuries sustained by the plaintiff in an explosion alleged to have been caused by the defendant’s negligence. The plaintiff had a verdict which was recorded under leave reserved. Thereafter the judge entered a verdict for the defendant. The plaintiff brings the case here on exceptions challenging this action and various rulings made during the trial. The defendant also comes here on exceptions but presses them only if the plaintiff’s exceptions are sustained.
A summary of the pertinent portions of the evidence most favorable to the plaintiff is as follows: On February 6, 1942, the defendant installed a gas burner, controlled by a thermostat, in the furnace of a house which was owned and occupied by the plaintiff’s parents with whom the plaintiff lived. The furnace was connected with a chimney, about six feet away, by means of an eight inch flue. This flue “as installed by the defendant ran from the furnace to the chimney in a horizontal line without pitch or rise and sagged in the middle.” The burner contained a safety device designed to prevent gas from flowing into the burner in the event the pilot went out.
On September 24, 1945, the defendant installed an automatic gas hot water heater near the furnace. It had a three inch flue which was joined by “a short right angle connection” to the eight inch flue, mentioned above, at a point about midway between the furnace and the chimney.1 In the basement there was also a coal stove which was connected with the chimney by a flue which entered the chimney about eight inches below the flue to which the furnace and hot water heater were joined.
In March of 1945 the defendant cleaned the pilot in the furnace and also the flue, and in May turned off the gas for the summer. On August 1, 1945, “there was a general *626cleaning of the furnace by the defendant” and “it was turned on for the season on October 26, 1945.” Thereafter down to the date of the explosion on February 3, 3946, “there had been no complaints with respect to the furnace or the hot water heater.” The mother of the plaintiff testified that “so far as she knew nobody other than the defendant did anything in connection with the house heater or the hot water heater.”
About ten o’clock on the morning of February 3, 1946, the plaintiff’s father smelled gas and went downstairs to the basement. He looked through a crack between the door and the casing of the furnace and could see no flame, but he “could hear the sound of gas going into the burner of the furnace.” He called the plaintiff who came down to the basement, and while the plaintiff was in front of the furnace “he heard the hot water heater go on and then there was an explosion which blew out the part of the flue pipe which was between the furnace and the hot water heater, the firebox door, and also the small clean-out door at the top of the furnace.” At seven o’clock on the morning of the accident a wood fire had been lighted in the stove next to the heater and there “was still a little fire ... [in it] after the explosion . . . but ... it was low and there were no sparks.” As a result of the explosion the plaintiff was severely injured. Evidence in addition to that set forth above will be recited hereinafter as occasion requires.
The plaintiff’s exception to the entry of the verdict for the defendant under leave reserved raises the following questions: (1) whether the defendant has violated any duty of care owed by it to the plaintiff, and (2) whether, if there was such a violation, it caused the injuries of which the plaintiff complains. The plaintiff contends that, irrespective of what caused the excessive accumulation of gas in the furnace and flues, the explosion was caused by the right angle or “T” connection and that the installation of this type of connection by the defendant was a breach of duty owed to the plaintiff. The plaintiff also contends that the explosion could have been found to be caused by the defend*627ant’s improper cleaning of the burner in the furnace. But, as will presently appear, we are of opinion that the evidence fails to show that the explosion was caused by the defendant’s improper cleaning of the burner, and the issue narrows down to whether the evidence would warrant a finding that the connection installed by the defendant was a breach of a duty owed to the plaintiff, and, if it was, whether it was causally related to the explosion.
We shall deal with the latter question first. The evidence on this issue came entirely from experts. One expert, Hesselschwerdt, called by the plaintiff, testified that “There is a possibility that with the ignition of this gas heater . . . a tongue of flame, or an incandescent spark of carbon could have gone up conceivably through . . . [the] stack pipe, out into the main breaching ... and caused an explosion somewhere in . . . [the] line,” that if “the right angle connection were not there, there would be a longer travel for that spark to become extinguished, and only the products of combustion could have gone into the breaching.” On cross-examination this witness qualified this testimony by saying that on the facts assumed by him as the basis for his opinion it was not possible that a tongue of flame could extend up through the heater and into the stack, and that there would be no more “than a speculative possibility that any spark would go up the stack from the water heater” if the heater was operating normally. (There was no evidence that the water heater was not operating normally.) It may well be that this evidence leaves the issue of causation in such an atmosphere of guesswork and speculation that it would not afford the basis of a finding that there was a causal relationship between the installation and the explosion. The opinion, as qualified, would seem to be no more than an “expression of conjecture.” Brownhill v. Kivlin, 317 Mass. 168, 170. But this evidence did not stand alone. Another expert, one Goldberg, who had had forty to fifty years experience as a master gas fitter, testified that in his opinion the right angle or “T” connection was the cause of the explosion. His explanation of how the explosion occurred *628may be summarized as follows: Escaping gas filled up the furnace and got into the flue. Because of the right angle or “T” connection it was forced down into the hot water heater. When the hot water was turned on upstairs (which started up the hot water heater) this gas was ignited and this in turn ignited the gas which had accumulated in the flue and furnace, and caused the explosion.
Whether the installation made by the defendant was one which would be likely to cause an explosion of the sort occurring here was not a matter of which the jury could be expected to know from their common knowledge and experience. The subject was a proper one for expert testimony. Jackson v. Anthony, 282 Mass. 540, 544. Flynn v. Growers Outlet, Inc. 307 Mass. 373, 376. On the basis of Goldberg’s testimony the jury would be warranted in finding that the connection installed by the defendant caused the explosion by permitting the escaping gas to flow down the flue into the hot water heater. We cannot say that this opinion is so “contrary to common sense or to known natural laws of which the court can take judicial notice” that it must be disregarded. Ruschetti’s Case, 299 Mass. 426, 430-431. The defendant contends, however, that Goldberg’s testimony should be put out of the case since it rested on an assumption that the water heater was constructed with an air space that continued “up the sides of the storage tank” whereas in fact the heater was constructed with a flue through the middle of the boiler. This contention is without merit. It is true that Goldberg gave his opinion on the assumption that an air space existed on the side of the tank. And it is also true that after making a second inspection of the water heater during the trial he was recalled and admitted that it was constructed with a flue in the center. But he stated that this would not change the opinion which he had previously given.
The defendant owed a duty to the plaintiff to install the water heater in “a workmanlike manner, with reasonable judgment, skill and care, according to the approved usages of . . . [[the] trade.” Kelley v. Laraway, 223 Mass. 182, *629184. In determining what is reasonable skill and judgment in the circumstances the highly explosive quality of gas and the serious consequences which may ensue if proper precautions to guard against an explosion are not taken are to be considered. The degree of care should be “commensurate with the danger that would probably result if such care were lacking.” Barbeau v. Buzzards Bay Gas Co. 308 Mass. 245, 247. See Edgarton v. H. P. Welch Co. 321 Mass. 603, 610. There was evidence from which the jury could have found that the “T” connection would cause a “turbulence” in the flue; that it would interfere with the smooth flow of gas along the flue; and that the gases coming out of the small flue were directed by that flue in a direction at right angles to the flow in the big flue. In addition to the foregoing the plaintiff offered in evidence certain “gas-fitting” regulations (the genuineness of which was not questioned; see G. L. [Ter. Ed.] c. 233, § 75, as amended by St. 1943, c. 190, § 1) promulgated by the building and health commissioners of the city of Boston pursuant to the provisions of St. 1938, c. 479, which included the following: “Appliances shall be installed so that their continued operation shall not in any way create a hazard to persons or property” (§ 13). “No flue piping of any appliance shall be connected to the flue piping of any other appliance by the use of right angle connections; such connections shall be made with a Y or 45-degree fitting” (§ 19). This evidence was excluded subject to the plaintiff’s exception. The evidence ought not to have been excluded.1 The validity of the regulations is not challenged. It is not disputed that the water heater was installed with a right angle or “T” connection rather than with a “Y or 45-degree fitting.” This was a violation of § 19 quoted above, and we *630are of opinion that such violation could be shown as evidence of negligence. The prevention of explosions was undoubtedly one of the consequences at which the regulation was aimed. Cases such as Richmond v. Warren Institution for Savings, 307 Mass. 483, and Greenway Wood Heel Co. Inc. v. John Shea Co. 313 Mass. 177, relied on by the defendant, are not controlling. In those cases no common law duty of due care was involved and the statutes under consideration created no new duty of care. Accordingly it was held that the violation of the statutes had no effect as evidence of negligence. Here the defendant, as stated above, owed a duty at common law to install the appliances with reasonable skill and judgment. As evidence that this duty was not properly performed the plaintiff, was entitled to show that the defendant violated the regulations in question. It is but another instance of the familiar principle which permits proof of the violation of a valid statute, ordinance or regulation as evidence of negligence as to all consequences which such statute, ordinance or regulation was intended to prevent. See Guinan v. Famous Players-Lasky Corp. 267 Mass. 501, 516, 517; Baggs v. Hirschfield, 293 Mass. 1, 3; Kralik v. LeClair, 315 Mass. 323, 326.
Considering the evidence which was introduced together with that which was erroneously excluded, we think that the plaintiff was entitled to have the jury determine whether the defendant was negligent and, if so, whether such negligence was causally related to the explosion. It was, therefore, error to enter a verdict for the defendant under leave reserved, and the plaintiff’s exceptions must be sustained.
It becomes necessary to consider certain exceptions taken by the defendant. The judge in his charge, among other things, told the jury in substance that they could find for the plaintiff if the defendant cleaned the burner in the furnace improperly, provided such improper cleaning caused the explosion. The defendant excepted to this portion of the charge. This exception must be sustained. It is true that there was evidence that the improper cleaning of the pilot could be an adequate cause of an explosion of the *631sort that occurred here. We assume that, if the defendant undertook to clean the burner and did so improperly, it would be liable to the plaintiff for injuries attributable to such improper cleaning. Ferguson v. Boston Gas Light Co. 170 Mass. 182, 191. Barbeau v. Buzzards Bay Gas Co. 308 Mass. 245. D’Entremont v. Boston Consolidated Gas Co. 320 Mass. 582, 583. But there was no evidence that would warrant a finding that the defendant's cleaning of the burner was improper. The defendant cleaned the burner on August 1, 1945, more than six months prior to the accident. The burner apparently functioned all right from the time it was started on October 26, 1945, until the explosion on February 3, 1946. Mere proof of an explosion plus evidence that an explosion could be caused by an improper cleaning of the burner — and no more than that was shown here — established only that there was a possibility that the burner was not properly cleaned. Verdicts must rest on more solid foundations. Sargent v. Massachusetts Accident Co. 307 Mass. 246, 250-251. Gilmore v. Kilbourn, 317 Mass. 358, 362-363. Smith v. Rapid Transit Inc. 317 Mass. 469, 470. The instructions of the judge permitted the jury to pass on an issue which should not have been submitted to them, and we cannot say that this error was not harmful to the defendant. We have no way of knowing whether the jury rested their verdict on improper cleaning or on improper installation. The case must be retried. Other exceptions argued need not be discussed since they relate to questions that are not likely to arise at another trial.
Plaintiff’s exceptions sustained.
Defendant’s exceptions sustained.
This type of connection was sometimes referred to in the testimony as a “T” connection.
Another regulation (part of § 19) which was offered in evidence provided that “No horizontal flue greater in length than twenty feet shall be allowed and for long runs a pitch or rise of at least one half inch to the foot shall be maintained from the appliance to the chimney.” This was rightly excluded. If we assume in favor of the plaintiff that a violation of this regulation was shown — a matter by no means free from doubt — it would have no relevancy here because it was not shown beyond conjecture that there was any causal relation between the lack of a pitch or rise and the explosion.