Rhodes, Justice:
Actions for wrongful death and conscious pain and suffering, arising out of a fatal boating accident on Lake Moultrie in Berkeley County, resulted in jury verdicts in favor of the plaintiff-respondent (Young) against the defendant-appellant (Tide Craft). Tide Craft appeals. We reverse.
I
The tragic mishap which is the subject of these actions occurred on March 30, 1972 shortly after the respondent’s husband, Novel Young, had launched his boat. According to the only testifying eyewitness, Young, who was alone in the boat, had just passed him and was proceeding on a straight course at approximately 30-35 miles per hour. The witness testified that he had glanced down to check his fuel gauge and when he looked up Young’s boat was circling and Young could not be seen. Upon drawing closer to Young’s boat, he observed that Young was hanging off the side of the boat with his head submerged in the water, one of his feet tangled in the lines of the trolling motor, and one of his hands grasping the rail which runs along the side of the boat. Before Young could be rescued he had drowned.
The boat in question was a flat bottomed, sixteen foot, 1972 “Deluxe Bayou” bass boat manufactured by Tide Craft, Inc. of Minden, Louisiana. The boat had been purchased by Young in December of 1971 from Danny Bell’s Winter Park in Eutaw Springs, S. C. When the boat was shipped to Bell, it contained two swivel seats, one located near the bow and the other near the stern. The boat was also equipped with a “pop-up stick” steering system, which had been patented by Tide Craft. The “pop-up stick” steering system differs from the conventional systems in that a “stick” or lever attached to a large pulley wheel is substituted for the conventional steering wheel and is mounted on the side of the boat to the left of the forward seat. Con[459]*459trols for the engine are mounted on the side of the boat to the right of this seat. This arrangement enables the bass fisherman to operate the boat from the forward seat and has the advantage of allowing him to fish from the forward seat without the interference of a steering wheel. The swivel seats allow him to fish in any direction.
As patented, the remainder of the steering system was to consist of what is commonly referred to in the boating industry as a “cable pulley” system. This system consists simply of 3/16 inch plastic-covered steel cable which, on the type of boat in question, was threaded through the stick steering pulley wheel, led aft through a 'series of pulleys attached to the side of the boat, and connected with the engine through a series of springs, pulleys and other hardware.
On Tide Craft’s 1972 model boat, the cable pulley system was only partially installed at the factory. Installation of the system on the boat purchased by Young was completed by Bell at the time the boat was fitted with the outboard motor Young had selected. Prior to the 1972 model, Tide Craft had completely installed the steering system, including cable and pulleys, leaving the dealer only to connect the cable with the motor. Due to complaints from dealers that complete installation was creating problems when it came time to connect the system to the various sizes and makes of motors available, Tide Craft only partially installed the steering system on its 1972 model boats.
After his purchase of the boat in December 1971, Young used the boat for several months without any problems in the steering system. Then in the latter part of February 1972 he began to experience difficulty in manipulating the steering stick. On February 25th, he took the boat for repairs to Henry R. Hegel, the owner and operator of Berkeley Marine Center in Moncks Corner. Hegel determined that the problem with the steering was that the cable had slipped off of a pulley because of slack in the steering system, and was resting between the pulley and pulley housing. As a result [460]*460of friction, the plastic sheathing on the cable had become frayed making it increasingly difficult to move the steering stick. Since Hegel did not have sufficient cable to rewire the system, he made temporary repairs which consisted of stripping away the frayed portion of the plastic sheathing, resetting the cable in the pulley, and adjusting the tension to take the excess slack out of the system. He advised Young that the system would have to be rewired since, without the plastic coating, the exposed portion of the steel cable would eventually corrode from exposure to the weather. However, Hegel told him it was perfectly safe to use the boat as it was until it could be rewired.
On March 9th, Young returned to Berkeley Marine Center and left the boat with Hegel to have the boat rewired. The boat was to be ready on the 14th and Young returned on that date. However, Hegel told Young he had mistak-ingly given him the wrong pickup date and the boat would not be ready until the following day. At this time Hegel still did not have sufficient cable to rewire the system and the possibility of splicing in a portion of new cable to replace the frayed portion was discussed. According to Hegel, Young wanted to use the boat for the approaching weekend and, if he were to have use of the boat, the only alternatives were to splice the cable or return the boat to Danny Bell for his attention. In any event, splicing was discussed and the decision to splice was made. Accordingly, at Hegel’s instruction, his mechanic spliced ten feet of cable into the steering system by means of tiller clamps. When Young picked up the boat on the 16th, Hegel, by his own admission, considered it “operable” but “dangerous” and the repairs “temporary”. The testimony reveals it was common knowledge in the boating industry that splicing is a dangerous practice. Young used the boat that weekend without incident. The following weekend he met his death.
The only reasonable inference from the evidence is that the cause of the boat going into its sudden, sharp turn was [461]*461a complete loss of steering which resulted from disengagement of the steering cable from one of the tiller clamps installed by Hegel’s mechanic. The clamp had been fitted over both the steel core and plastic sheathing of the cable. As a result, considerable stress was placed upon the plastic sheathing. Consequently, the plastic gave and the cable pulled free of the clamp.
In 1975 the respondent instituted a wrongful death action and a survival action for conscious pain and suffering. The suits were combined for trial. Joined as defendants in each of the two actions were Tide Craft; Danny Bell, d/b/a Bell’s Winter Park; and Henry H. Hegel, d/b/a Berkeley Marine' Center. Both actions were based on alternative theories of negligence, breach of implied warranty, and strict liability in tort, all of which were submitted to the jury. The jury, in the action for conscious pain and suffering, returned a verdict against Tide Craft in the amount of $30,000 actual damages. In the wrongful death action, a verdict was returned against Tide Craft in the amount of $160,000 actual damages and $10,000 punitive damages. Bell and Hegel were absolved of liability.
II
There are two separate and distinct aspects of claimed liability in this case. The respondent first contends that Tide Craft is chargeable with the splicing and resulting disengagement of the steering cable from the tiller clamp and is, thus, liable for the damages alleged. Second, respondent contends that, even if Tide Craft cannot be charged with the splicing of the steering cable, Tide Craft is, nevertheless, liable because certain alleged defects of the boat which became operative after the cable parted, were contributing proximate causes of Young’s death.
Proximate cause is an essential element common to all three theories of recovery advanced by the respondent in this case. Prosser, Law of Torts, § 103, pp. 671-2 (4th Ed. [462]*4621971) (common elements) ; Royal v. Black & Decker Manufacturing Co., 205 So. (2d) 307 (Fla. App. 1968) (common denominators) ; and Williams v. Brasea, Inc., 497 F. (2d) 67 (5th Cir. 1974) cert. den. 423 U. S. 906, 96 S. Ct. 207, 46 L. Ed. (2d) 136 (1975) (elements of strict liability in tort). With regard to the first aspect of this case, we conclude that Tide Craft cannot be held liable since the intervening acts of Hegel constitute, as a matter of law, the sole proximate cause of the disengagement of the steering cable. As to the second aspect, with one exception to be hereinafter disposed of on other grounds, we conclude that the respondent has failed to establish proximate cause.
Ill
Under the first aspect of respondent’s case the following are alleged to constitute defects and/or lack of due care on the part of Tide Craft: (1) improper design of and inherent difficulties with the pop-up steering system; (2) allowing installation of the steering system by dealers, and/or not furnishing instructions on proper installation; and (3) failure to warn against improper methods of repair in the event of difficulties with the system. The respondent asserts that the actions of Hegel were foreseeable and that the alleged defects and/or lack of due care on the part of Tide Craft were proximate causes of the injuries complained of.
A reading of any of the host of decisions in this State clearly discloses that the touchstone of proximate cause in South Carolina is foreseeability. “Foreseeability of some injury from an act or omission is a prerequisite to its being a proximate cause of the injury for which recovery is sought. Kennedy v. Carter, 249 S. C. 168, 153 S. E. (2d) 312.” Stone v. Bethea, 251 S. C. 157, 161, 161 S. E. (2d) 171 (1968). The standard by which foreseeability is determined is that of looking to the “natural and probable consequences” of the complained of act. E. g. McQuillen v. Dobbs, 262 S. C. 386, 204 S. E. (2d) 732 (1974); Childers v. Gas Lines, Inc., 248 S. C. 316, 149 [463]*463S. E. (2d) 761 (1966); Matthews v. Porter, 239 S. C. 620, 124 S. E. (2d) 321 (1962). While it is not necessary that the actor must have contemplated or could have anticipated the particular event which occurred, e. g., McQuillen, supra, liability cannot rest on mere possibilities. The actor cannot be charged with “that which is unpredictable or that which could not be expected to happen.” Stone, supra, 251 S. C. at 161-2, 161 S. E. (2d) at 173. In determining whether a consequence is one that is natural and probable, the actor’s conduct must be viewed “in the light of the attendant circumstances”, Stone, supra, at 161, 161 S. E. (2d) at 173.
Where there is no contention that the intervening actions of a third person broke the causal chain, the probable consequences test consists simply of determining whether the wrongdoer’s actions were such that they “would probably cause injury to something or someone,” Childers, supra, 248 S. C. at 325, 149 S. E. (2d) at 765. In other words, there is only one consequence to be taken into consideration — the injury.
Where there is a contention that an intervening agency interrupts the foreseeable chain of events, there are two consequences to be tested: (1) the injury complained of, and (2) the acts of the intervening agency. If the acts of the intervening agency are a probable consequence of the primary wrongdoer’s actions, i. e., “foreseeable”, the primary wrongdoer is liable. Stone, supra. However, even if the intervening acts are not foreseeable, the primary wrongdoer is nevertheless liable if his actions alone “would have caused the loss in natural course.” Benford v. Berkeley Heating Co., 258 S. C. 357, 365, 188 S. E. (2d) 841 (1972).
Turning to the instant case, we must determine: (1) whether the actions of Hegel were a probable consequence of the alleged defects and/or negligence attributable to Tide Craft; and (2), if the answer to the first question is “no”, [464]*464whether the injury complained of would have followed in natural course absent Hegel’s actions.
In making these determinations it is elementary that we view the evidence in the light most favorable to the opposing party, e. g. Legette v. Smith, 265 S. C. 573, 220 S. E. (2d) 429 (1975). It is also elementary that when the evidence permits but one reasonable inference, a question of law is presented for the court. Horne v. Atlantic Coast Line R. Co., 177 S. C. 461, 181 S. E. 642 (1935).
In light of the attendant circumstances in the present case, the only reasonable inference to be drawn from the evidence is that Hegel’s actions were not foreseeable. Hegel was aware that splicing was dangerous and far removed from good practice and yet instructed his mechanic to splice the steering cable. By his own testimony, he admitted that splicing was “not accepted practice” and “not advisable”. When the boat left his place of business on March 16 with the spliced cable, he considered it “dangerous” but “operable” and the repair “temporary”. He testified that he was “certain he [Young] knew it was unsafe.” The effect of this statement regarding Young’s knowledge of the danger was an admission on Hegel’s part that he [Hegel] fully realized the danger involved in splicing the steering cable.
The only evidence which would tend to support foreseeability on the part of Tide Craft was testimony by its witnesses that there were different levels of competence among the dealers. However, this condition is insufficient to charge Tide Craft with foreseeing that a repair method would be chosen in the face of knowledge by the repairman that it was inherently dangerous.
Even disregarding Hegel’s knowledge of the danger, the other circumstances point out the high improbability that a repair would be effected by splicing instead of rewiring. The only reasonable inference from the evidence is that it was common knowledge in the trade that splicing was unsafe [465]*465and far removed from good practice. Hegel said this was the only splice he had ever authorized both prior to and after the accident in question. Hegel’s mechanic testified that other than Young’s boat he had never heard of a tiller clamp being used in the midst of a steering system. The only reasonable inference to be drawn is that the likelihood of a steering cable being spliced is highly remote.
Adding together the remoteness of the possibility that splicing would take place under any given set of circumstances with Hegel’s knowledge of the danger involved, it can not be seriously contended that Hegel’s actions were a probable consequence of any wrongdoing on the part of Tide Craft. “The actor’s conduct may be held not to be a legal cause of harm to another where after the event and looking back from the harm to the actor’s negligent conduct, it appears to the court highly extraordinary that it should have brought about the harm.” Restatement (Second) of Torts § 435(2) (1965). As was stated in the case of Steward v. Cox, 55 Cal. (2d) 857, 13 Cal. Rptr. 521, 362 P. (2d) 345 (1961): “[Negligent conduct with full realization of the danger may properly be considered highly extraordinary”. 13 Cal. Rptr. at 525, 362 P. (2d) at 349. We agree. The only reasonable inference from the evidence is that Hegel had full realization of the danger. This is not a case where ignorance on the part of the intervening third party led to the creation of a highly dangerous condition, but one where the risk was unleashed with full realization of the dangerous consequences. In looking back on the event, we are appalled at the highly extraordinary character of Hegel’s actions.
Although the actions of Hegel were unforeseeable, the question remains whether the injury would have occurred in natural course absent the actitons of Hegel. There is no evidence which would reasonably support an inference that the cable would have parted or that there would have been a complete loss of steering absent the splicing. The only rea[466]*466sonable inference from the evidence presented is that the defects and lack of due care alleged by the respondent would have, at most, resulted only in a constriction in the steering system and inability to manipulate the steering stick.
The respondent’s expert did testify that with binding in the system “eventually if it goes far enough you’ll get a parting.” Assuming this to be true, it is highly improbable that such a parting of steel cable could occur without one having been aware of the binding for a very considerable period of time. Indeed, Young did discover the binding and, as the respondent’s expert testified, “[with binding] usually you can’t move the steering wheel but that doesn’t mean you’re thrown out of the boat where you can’t shut the engine down and, in terms of safety, I would much rather have something bind up where I could shut the engine down and find out what the problem is.”
Given the above, the only reasonable inference that can be drawn from the evidence is that the damages here sustained would not have occurred in natural course absent Hegel’s unforeseeable actions.
IV
Turning to the second aspect of this case, the respondent asserts that, even if Tide Craft cannot be held responsible for the splicing and subsequent parting of the steering cable, certain features of the boat and its design constituted defects, or were manifestations of a lack of due care, which came into play at the time the cable parted and were contributing proximate causes of Young’s death. The particulars which are alleged to constitute the basis for recovery under each of the three theories advanced can be summarized as follows: (1) a propensity of the boat to eject an occupant, and (2) the failure to install a kill switch, a safety device which cuts power to the motor if the operator should be thrown from his seat. We hold that the respondent has failed to establish that the propensity [467]*467to eject was a proximate cause of the injuries complained-of, and that the failure to install a kill switch does not constitute a defect under either the implied warranty or strict liability theories, nor does it constitute negligence on the part of Tide Craft.
With reference to the alleged propensity of the boat to eject, respondent’s theory is that the size of the outboard motor on the boat, the height of the seats, the absence of locking devices to prevent the seats from swiveling and the location of the “pop-up” steering stick were factors which, because of the dynamic forces operating upon an occupant of the boat while in motion, resulted in an enhanced risk of ejection. This alleged propensity of ejection, the respondent contends, was a contributing proximate cause of Young’s being thrown overboard. As pointed out, the only reasonable inference from the evidence is that the precipitating cause of the boat going into its sudden, sharp turn was the disengagement of the steering cable from the tiller clamp. Because of this, in order for the respondent to establish the propensity to eject as a contributing proximate cause, it was necessary that she present evidence from which it could reasonably be inferred that, in the absence of the allegedly defective features of the boat, the forces generated by the loss of steering alone would not have thrown Young overboard.
The respondent sought to establish the alleged causal connection through her expert witness Frank Douglas Fowler. Fowler’s theory of causation was what he termed a “domino theory” with the loss of steering and each of the defects which allegedly resulted in the propensity to eject constituting “dominoes”. Fowler concluded that each of these dominoes were contributing causes of Young’s ejection and that, if any one of them had been removed, Young would not have met his death.
Fowler’s conclusion was not elicited by means of a hypothetical question. Neither did Fowler set forth the underlying factual basis for his conclusion. Thus, [468]*468his conclusion was simply a surmise, lacking in probative value and inadequate to establish the alleged causal connection.
“Opinion testimony of an expert witness may be based upon facts within his own knowledge or upon hypothetical questions embracing facts supported by the evidence and relating to the particular matter upon which the expert opinion is sought. ...” 31 Am. Jur. 2d, Expert and Opinion Evidence § 36 (1967). The opinion of the expert “must be based upon facts . . . sufficient to form a basis for an opinion. . . . Expert opinion is inadmissible if its factual foundation is nebulous.” Id.; accord, Chapman v. Foremost Dairies, Inc., 249 S. C. 438, 154 S. E. (2d) 845 (1967). The probative value of expert testimony stands or falls upon an evidentiary showing of the facts upon which the opinion is, or must logically be, predicated. Chapman, supra. It is, of course, elementary that the factual or underlying basis for the expert’s opinion be set out, otherwise the opinion lacks probative value. Benford v. Berkeley Heating Co., 258 S. C. 357, 188 S. E. (2d) 841 (1972); see Chapman, supra.
The requirement that an underlying basis be established is no more than a corollary to the settled rules of evidence for unless there is a showing that the expert’s special knowledge has “been brought to bear upon the facts of the case being tried . . . [his opinion] would be irrelevant,” Dreher, A Guide to Evidence Law in South Carolina 10 (1967), and without probative value.
The requirement of an underlying basis is by no means intended to be a requirement that the expert set out his calculations or detail the scientific or professional knowledge leading to his conclusions. The scientific facts, calculations, or methodology upon which his opinion is based “may be so exclusively within the domain of scientific or professional knowledge that their significance would not be understood by the jury.” 31 Am. Jur. 2d, [469]*469supra, § 38. Neither should it be implied, however, that the complexity of the expert’s area of knowledge relieves him of establishing, before the jury, the underlying basis for his opinion. However complex or esoteric the specialized knowledge the expert draws upon, he must show that in formulating his opinion, he has taken into consideration the material facts of the case being tried which was necessary to the formation of an intelligent opinion.
As previously stated, respondent’s theory under the second aspect of her case was that, given the forces which were operative when the boat was in motion, and given the alleged defects in the boat, a propensity of ejection resulted. Implicit in the conclusion by Fowler that these defects were contributing causes of ejection is the further conclusion that, when the tiller clamp and cable separated, ejection would not have occurred in the absence of these alleged defects because, otherwise, they would not constitute contributing proximate causes.
It is apparent that both the initial forces which came into play as a result of the steering loss and any enhancement of these forces attributable to the alleged defects were material considerations in forming an opinion as to proximate cause. The importance of these forces is evidenced by Fowler’s own testimony. He testified that “when you look at what forces act upon a human when there is loss of steering you can predict whether he’s going to go out or stay in the boat. ...” It follows, then, that to predict whether a person will be thrown out of a boat given a sudden loss of steering, it is necessary to determine “what forces act upon” him. Ascertainment of the nature and extent of the dynamic forces acting upon the boat and its occupant was especially critical in view of Fowler’s testimony that different boats have different forces acting upon them, and in view of the very strong inference which arises from the circumstances of the mishap itself that considerable centrifugal force was generated solely by the loss of steering. Despite all of this and [470]*470.despite Fowler’s testimony that he could have mathematically rated each of the “dominoes” as to its potential for danger, he made no attempt to ascertain the nature and extent of the operative, dynamic forces, nor did he attempt to scientifically determine the contribution, if any, of the individual “dominoes” to ejection. Not having done so, it was impossible for Fowler to say with any reasonable degree of accuracy that the dynamic forces attributable to the alleged defects contributed to Young’s ejection to the extent that ejection would not have occurred in the absence of the alleged defects.
Having failed to establish an underlying factual basis taking into consideration the material facts necessary to the formation of an intelligent opinion, Fowler’s testimony on causation is devoid of probative value and cannot support a verdict. Otherwise, the verdict “would be allowed to rest on conjecture and speculation, which is prohibited.” Horton v. Greyhound Corp., 241 S. C. 430, 438, 128 S. E. (2d) 776, 781 (1962).
Although there is no need to consider them in detail, we also note that the respondent has set out two other particulars upon which she bases recovery under this aspect of her case, these being a failure to warn of the propensity to eject and a failure to conduct tests to determine the maximum safe speed at which the boat could be operated before the alleged ejection propensity comes into play. Since the respondent has failed to show that the ejection propensity was a proximate cause of the injuries complained of, it follows that the failure to test and the failure to warn cannot be deemed proximate causes of the injuries in question.
V
Tide Craft’s failure to install a kill switch warrants separate consideration. A kill switch is a safety device which cuts electrical power to a boat’s motor whenever the operator is thrown from his seat and is no longer in a position to [471]*471control the boat. With the electrical power disconnected, the motor will cease functioning and the boat will come to a stop. This device is attached to the motor and ignition system neither of which is manufactured by Tide Craft. Even assuming that Tide Craft had responsibility in this area, it is clear to us that the absence of this device did not constitute a breach of warranty, an unreasonably dangerous defect, or lack of due care on the part of Tide Craft.
To warrant recovery under the strict liability in tort theory, the absence of the kill switch must constitute a defect “unreasonably dangerous to the user or consumer,” S. C. Code § 15-73-10 (1976). As pointed out in Skyhook Corp. v. Jasper, 90 N. M. 143, 560 P. (2d) 934 (1977), the test of whether or not the “failure [to incorporate a safety feature or device in a product] constitutes a defect is whether the product, absent such feature or device, is unreasonably dangerous to the user or consumer or to his property”, 560 P. (2d) at 938.
The question that presents itself is whether the absence of the kill switch per se rendered the boat “dangerous to an extent beyond that which would be contemplated by the ordinary consumer who purchases it, with the ordinary knowledge common to the community as to its characteristics,” Restatement (Second) of Torts, § 402A, Com. i (1965). In applying this test to a situation in which the manufacturer of a winch used on fishing boats failed to include a brake on the winch, the court in Williams v. Brasea, Inc., 497 F. (2d) 67 (5th Cir. 1974), cert. den. 423 U. S. 906, 96 S. Ct. 207, 46 L. Ed. (2d) 136 (1975), stated: “The danger posed by the lack of a brake could scarcely be beyond the contemplation of crewmen who knew of its absence and worked with the winch in that condition on a daily basis.” Id. at 79. It is common knowledge that a normal risk of boating is that of being thrown overboard. While the test set out above is an objective one and knowledge common to the community must be attributed to Young, [472]*472there can, nevertheless, be no question of his awareness of this risk. His wife testified that Young had fished almost every weekend for the past ten years. He had often used boats and his wife felt he was relatively familiar with them. Young being aware of the normal risks of boating, the danger posed by the obvious lack of a kill switch could hardly be beyond his contemplation. Accordingly, the lack of a kill switch does not constitute a defect within the meaning of the strict liability in tort statute.
Since the main thrust of the parties’ arguments have been directed to the strict liability theory, we see no need to discuss in detail the theories of negligence and implied warranty as they relate to the failure to install a kill switch. Given the facts of this case, we feel it follows from the above discussion, and so hold, that the failure to install a kill switch does not constitute a breach of any implied warranty nor does it evidence a failure to exercise due care.
In view of our disposition of this case, there is no need to reach the other issues raised by the parties.1
Reversed.
Lewis, C. J., and Littlejohn and Gregory, JJ., concur.
Ness, J., dissents.