delivered the opinion of the court:
The plaintiff, M & G Provision Company (M.& G), is a purveyor of meat and occupied the building at 185-187 South Water Market Street in Chicago. The building .had an ammonia refrigeration system which refrigerates the basement and first two floors. Tire plaintiff, United Potato Company, owned. the adjoining premises at 183 South Water Market. The defendant, Midwest Engineering & Equipment Company (Midwest), pursuant to a service contract, had undertaken to maintain, service and repair M & G’s ammonia refrigeration system. On December 23, 1965, an explosion occurred at the premises occupied by M & G; and this suit was brought, alleging in Count I specific acts of negligence. Count III was based on the doctrine of' res ipsa loquitur. Count III was stricken by the court and the case submitted to the jury on Count I only. A verdict was returned for the defendant; and this appeal is based solely on the plaintiffs’ contention that the court erred in refusing to submit the theory of res ipsa loquitur to the jury.
The refrigeration system involved is “closed,” that is, the refrigerant, ammonia, has been put in and the system sealed. In operation, the am*899monia changes intermittently from a liquid to a gas and back to a liquid. The ammonia, when located in the compressors, is a hot gas which goes through lines to the condenser on the roof of the building where it is condensed from a gas to a liquid and cooled. The liquid is then piped down the system to two receivers located in the compressor room. The receivers are connected to two accumulators where the liquid collects and is reserved for later use to cool the budding. The liquid ammonia lines run from the accumulators to the cooling coils located in the basement and on the first and second floors. Ammonia goes to these cods as needed and directed by an “expansion” or solenoid valve which expands the liquid ammonia to a semi-gaseous state. There were five solenoid valves at M & G: one in the basement, two on the first floor and two on the second.
The ammonia system was in the building when it was purchased by M & G in 1965; it had been installed in 1956 or 1957 and was too complex for M & G personnel. They never repaired or serviced the system. Service was delegated to the defendant, Midwest, which was the “exclusive repairer” of the equipment. It did not instad the system. Under the terms of its agreement, Midwest was required to service the plant once a month and to make any other cads M & G wanted. M & G was bided once a month on a time-material basis. One of the monthly functions to be performed was the draining of oil from the accumulator tanks. The compressors required lubrication, and oil was fed through a crankshaft to lubricate the bearings and pistons within the compressors. Although a portion of the oil stayed within the system and was recycled, part of it settled in the accumulators with the result that the level of oil increased. That od required periodical draining through the valve at the bottom of the accumulators. Removal of the oil, which increased the volume of ammonia in the accumulators, was a corrective measure. There was no automatic drain system because the oil did not accumulate quickly.
On December 23, at about 4 or 5 A.M., Fred Ginsberg, the vice president of M & G, walked through the plant on his dady inspection. He noticed the smell of ammonia on the second floor. Suspecting a leak, he immediately called Midwest and arranged for a repairman.
Lou Braunlin was the Midwest employee who normady serviced the unit and, coincidentally, had installed the unit in 1956 or 1957 while the employee of another service company. He arrived at approximately 7 A.M. and was directed by Ginsberg to the area of the leak, a valve on the second floor. Braunlin testified that he repaired the leak, which was in a solenoid valve on the second floor, and was directed by his dispatcher to give the system a regular check and drain the oil.
Oscar Ginsberg, the president of M & G, had arrived at the premises *900at 6:30 A.M. and saw Braunlin in the boiler room. At the time Braunlin was alone and was draining the oil; he was the only one working on the system on that particular morning and was completely in control of whatever was being done in the refrigeration system. During the normal draining process, oil came out like an oily gelatin and it took hours to clear the system. Braunlin had opened the valve through which the oil drained and it was coming out very slowly. Normally, the draining operation caused a slight smell of ammonia.
The compressor room within which Braunlin was working was located directly below the men s washroom on the second floor. The employees working on that floor began to complain of the smell to Oscar Ginsberg. The odor got progressively worse, and at 9:30 A.M. the work crew informed Oscar Ginsberg that the smell was getting exremely strong. Oscar then directed his brother Fred to check the compressor room. At the same time, Oscar Ginsberg saw Braunlin walk by and told him that the men reported they were getting a strong ammonia smell. Braunlin testified that he thought that “there was nothing to be alarmed about.” He told Ginsberg he would be right back.
Fred Ginsberg reached the compressor room about 5 minutes after he had talked to his brother. When he opened the door to the room he saw that it was filled with ammonia. He knew it was ammonia because of the smell; he could not breathe. He described the atmosphere in the room as a “terrible white fog.” He immediately closed the door and ran upstairs screaming for the fire department. As he reached the top of the stairs, an explosion blew him through the swinging doors which led to the loading dock. He was familiar with the smell of ammonia and there was no doubt in his mind that there was an ammonia haze in the compressor room. The explosion took place within minutes after Braunlin had left the premises.
Edward McLean was a qualified consulting engineer who testified for the plaintiffs. His experience included many years in gas and electrical engineering; and he had been involved in the design of gas systems. He arrived upon the premises 4 days after the occurrence. Repair work after the explosion had been done by Midwest, and they had immediately begun disassembling parts of the system. Consequently, McLean had to concede that he was “not personally aware of anything that Midwest did or did not do to the ammonia refrigeration system.” It was his opinion that the cause of the explosion was an ammonia accumulation in the room up to the explosive limit. The probable source of ignition was the pilot light on the boiler. One of the purposes of his inspection was to rule out any other causes. He checked all natural gas lines, sewer sources, and available fluids. Of the various fluids used upon the premises, none *901was flammable; sewer gas was eliminated because there was no odor of it and, additionally, a check of the sewer track indicated it was full of water and was working. The boiler gas lines and gas meter were undamaged and working. Thus, he ruled out all causes other than ammonia.
He also concluded that the ammonia leak had not been caused by a breakdown of any portion of the ammonia system. He described the possibility that the ammonia came from the drain valve on the accumulator tank which Braunlin was draining: A valve is opened to drain the oil. After the oil is drained, an interphase between the oil and ammonia is reached. At that point there is a smell of ammonia. Thereafter, the ammonia rushes out because the accumulator is under pressure. Ammonia then makes a swishing sound, picks up moisture from the air and rises from a liquid to a gaseous vapor. When there is ammonia present in an open-valve situation, it has the appearance of a white fog.
Caryl Reaver at the time of the explosion was the service manager for Midwest and was Braunlin s supervisor. He had been in the refrigeration business for about 25 years; 8 of them in ammonia refrigeration service. He was personally familiar with the refrigeration system at M & G but did not have a schematic design of it. He visited the plant 2 days after the explosion, and a check valve was discovered under pipe covering and in a wall. (The record does not reflect who discovered it.) It was found between the solenoid and the accumulators. The check valve was not in the system when the plant first went into operation, nor was it installed by Midwest. A check valve acts as a block; it allows a substance to flow in one direction only. It is conceded by the plaintiffs that the check valve was improperly located. Reaver testified that Midwest found a cracked strainer located in the ammonia line to which the check valve had been attached. (The record does not reflect who discovered that either.) It was his opinion that the strainer had been damaged by pressure within the ammonia line.
Braunlin testified that he opened the quarter-inch purge valve and allowed the oil to run from the accumulator through a copper line which he had attached to the drain valve on the accumulator and which he had run into the bottom of a barrel which contained about 25 gallons of water. The process took approximately an hour and a half and the oil stopped draining before Braunlin left for his break. He disconnected the copper tube, shut off the purge valve and replaced the steel plug in the end of the purge valve to seal it. On explaining his reaction when Ginsberg told him that a strong odor of ammonia was detected by the men in the area directly above where he was working, he testified as follows:
“I assumed that he was referring to this ammonia odor from *902draining the oil, I didn’t — couldn’t think of any other thing that it could be because the plant was operating 100 per cent, everything was in order, there [were] no leaks or anything, so I assumed that he was referring to this odor from the draining of the oil, so I went on across the street and went by the counter and got a cup of coffee and a roll * * *."
He also testified that after they found the check valve “it immediately became clear to [him] the reason for this broken strainer, because this check valve isolated this whole area, liquid system from its release valve.”
The elements necessary for application of the doctrine of res ipso loquitur and the duties of court and jury with respect to its application are set out in Drewick v. Interstate Terminals, Inc., 42 Ill.2d 345, 348-349, 247 N.E.2d 877:
“Assuming that die plaintiff was in the exercise of due care for her safety, two additional conditions must be established for the application of the doctrine of res ipsa loquitur: first, the accident must be the kind which does not ordinarily occur without someone’s negligence; and second, the instrumentality which caused the injury must have been .within the management or control of the defendant. [Citations.] The purpose of this second condition is, of course, to limit the application of the doctrine to those cases' where the negligence, if any, must reasonably have been that of the defendant. It is not enough for the plaintiff to show that she was- injured because of the negligence of an unknown party.[Citation.]
Whether the doctrine applies in a given case is a question of law which must be decided in the first instance by the trial court. [Citation.] Once this decision is made, however, it becomes the function of the jury to weigh the strength of the inference- of negligence arising from the facts in evidence. This inférence does not vanish upon the introduction of contrary evidence by defendant but remains to be weighed with all the other evidence.”
It is the defendant’s position that the- doctrine is not ápplicable because the evidence establishes, as a matter of law, that the ammonia explosion could have occurred without negligence on the part of anyone and that 'the ammonia refrigeration system was not sufficiently within the control of the defendant.
In support of its first contention, the defendant argues that the evidence shows that the system had been subject to other leaks in tire past and, therefore, “the existence of a subsequent leak * * * does not *903support a presumption that someone was negligent.” It is, of course, true that if the leak resulted from the constant stress to which the system was subject, no liability could be attributed to the defendant. But the jury had before it the testimony of Fred Ginsberg that they had experienced only occasional minor leaks, and “nothing that strong.” Braunlin had checked the system and found only the leak on the second floor and he repaired it. He testified that, when he left the plant, it was “operating 100 per cent, everything was in order, there [were] no leaks or anything.” It is particularly significant that the explosion took place in the very room in which Braunlin was working, The compressor room was filled with a “terrible white fog.” McLean testified that ammonia, in the draining process, has the appearance of a “white fog” in an “open valve situation.”
The jury was not required to believe Braunlin’s testimony that he closed the valve. The job of draining the oil was a tedious one; he had done it many times before with no adverse results. It would be understandable, therefore, that he-would leave while the oil was still draining. Why would he intend to return if he had repaired the leak, checked the system and drained the oil? We conclude that the circumstantial evidence was sufficient to present a jury question of whether the “terrible white fog” resulted from another leak inherently caused by the system itself or was caused by the negligence of the defendant.
The defendant’s second argument is that the evidence shows, as a matter of law, that the refrigeration system was not sufficiently within the control of the defendant. It bases its argument on the fact that it had only monthly contact with the system and had no way of knowing what was done to it “either intentionally or unintentionally during the day-today occupation of the premises by the plaintiff”; that the system was not installed by the defendant; nor was the crucial check valve.
The evidence shows that the check valve would have been proper if it had been in a different location. No one could explain when it was installed, and no one could givé a reason why it would have been hidden behind a wall. The existence of the chéck valve, as well as the strainer, depended solely on the testimony of the witnesses of the defendant, which assumed control of the scene of the explosion immediately after it occurred. The plaintiffs have not challenged the credibility of those witnesses; therefore, we shall accept their testimony as true.
The defendant’s contention that the explosion was caused by something not under its control may be summarized: The check valve was improperly placed in the system by someone else; it caused a blocked system; all five solenoids, which opened or closed depending on the *904temperature, closed at the same time; the temperature of the ammonia was raised and thereby increased the pressure within the pipes to the point that the strainer ruptured and the ammonia spewed out.
The answer to the defendant’s argument is apparent: If the jury believed McLean’s testimony that the explosion itself, and not internal pressure, caused the break in the strainer, it could reject all the testimony concerning control over the check valve as immaterial.
In addition, because the defendant came only on a periodic basis, its control is not negated as a matter of law. The defendant was the “exclusive repairer”; neither the plaintiffs nor any third party worked on the system; Braunlin was alone in the compressor room, and the smell of ammonia became stronger while he was in the building; he left, but was to return; and, as we have emphasized before, the explosion took place in the very room in which he had been working within minutes after his departure. We believe these facts provide more than ample evidence to require the submission of question of control to a jury. See Drewick v. Interstate Terminals, Inc., 42 Ill.2d 345, 247 N.E.2d 877; Metz v. Central Illinois Electric & Gas Co., 32 Ill.2d 446, 207 N.E.2d 305; Oakdale Building Corp. v. Smithereen Co., 322 Ill.App. 222, 54 N.E.2d 231.
In the factual setting of this case the language of the Restatement of Torts is appropriate (Restatement (Second) of Torts, § 328 D, comments at 159 (1965):
“The plaintiff’s burden of proof (see § 328A) requires him to produce evidence which will permit the conclusion that it is more likely than not that his injuries were caused by the defendant’s negligence. Where the probabilities are at best evenly divided between negligence and its absence, it becomes the duty of the court to direct the jury that there is no sufficient proof. The plaintiff need not, however, conclusively exclude all other possible explanations, and so prove his case beyond a reasonable doubt. Such proof is not required in civil actions, in contrast to criminal cases. It is enough that the facts proved reasonably permit the conclusion that negligence is the more probable explanation. This conclusion is not for the court to draw, or to refuse to draw, in any case where either conclusion is reasonable; and even though the court would not itself find negligence, it must still leave the question to the jury if reasonable men might do so.”
For these reasons we judge that Count III based on the doctrine of res ipsa loquitur should have been submitted to the jury for its determination. The judgment of the Circuit Court of Cook County is reversed *905and the cause remanded with directions to reinstate Count III and to proceed in a manner not inconsistent with this opinion.
Judgment reversed and cause remanded with directions.
BURKE, P. J., and HALLETT, J., concur.