This is a products liability case arising out of a one-car accident. Recovery is sought against the manufacturer and the dealer on grounds of breach of warranty, negligent misrepresentation, and negligence in refusing to repair. We hold that neither defendant is legally responsible for the accident, because the evidence is insufficient as a matter of law to show: (1) that the representation relied upon was false, or (2) that the accident was caused by the failure to repair or by any defect existing at the time the automobile was delivered.
Daniel J. Puskar, plaintiff, sustained personal injuries while driving alone in a 1959 Thunderbird automobile manufactured by Ford Motor Company. He had purchased the vehicle new from Jack Roach-Bissonnet, Inc., hereinafter referred to as Roach. Plaintiff brought this suit against Ford and Roach to recover damages for his injuries. The trial court rendered judgment on the verdict in his favor and against Ford and Roach, jointly and severally, for $60,815.33. The Court of Civil Appeals affirmed as to Roach, but reversed the judgment of the trial court and rendered judgment that plaintiff take nothing as to Ford. Ford Motor Co. v. Puskar, 394 S.W. 2d 1.
The accident occurred shortly after midnight at the T-intersection of Roark and Alief Roads in Harris County. On the occasion in question, plaintiff was driving the Thunderbird in a northerly direction on Roark Road, a graveled road which intersects but does not cross Alief Road. When plaintiff was some 125 to 200 feet south of the intersection, he saw the stop sign controlling vehicles entering Alief Road. He thereupon made a heavy application of his brakes, but there was little or no response. When the automobile reached the intersection, plaintiff attempted to turn to the right but the steering wheel would move only a few inches. He was unable to turn the vehicle, and it continued north across Alief Road, which is about 33 feet wide, across a ditch and into an embankment. Plaintiff was seriously injured in the accident.
When the automobile was purchased on May 2, 1959, plaintiff received from Roach its dealer’s warranty that each part of the vehicle would be “free under normal use and service from defects in material and workmanship for a period of ninety (90) days from the date of delivery of such product to Purchaser, or until such product has been driven, used or operated for a distance of four thousand (4,000) miles, whichever event first shall occur.” The warranty was expressly made subject to certain limitations and conditions which are not material under our view of the case. Plaintiff also received from Roach an owner’s manual issued by Ford and describing various features of the automobile. The relevant portions of this manual are as follows:
“MASTER-GUIDE POWER STEERING
“Up to 75% of the effort needed to steer your Thunderbird is taken over by Master-Guide Power Steering. Yet, this optional hydraulically operated steering system allows you to retain the natural feel of the steering wheel, particularly when you’re driving your car along the open highway.
“Master-Guide Power Steering provides a power assist only when your Thunderbird’s engine is running. However, even if your engine is stopped, or if the power system should not be operating normally, you’ll have safe steering and full control of your car with conventional steering.”
“SWIFT SURE POWER BRAKES
“The low-level pedal for the optional vacuum-operated Swift Sure Power Brakes system will operate with approximately one-third less pressure than the *264conventional brake pedal for most normal stops.
“If, for any reason, Swift Sure Power Brakes should completely lose their brake-assisting power, your Thunderbird’s conventional hydraulic brake system will remain fully effective and it will take over to permit you to stop the car safely when you push down the brake pedal.’’
Plaintiff read the dealer’s warranty and the owner’s manual. He understood, believed and relied upon them. After delivery of the automobile, he proceeded to drive it and found that it operated normally with a few exceptions. The motor died once while he was backing out of the driveway, and the same thing happened on several occasions when the brakes were applied or when the motor was idling while the car was standing in traffic. Plaintiff noticed that extra effort was required in applying the brakes when the motor died, but he was moving so slowly that this did not make much difference. There was no reason to turn the steering wheel on any of these occasions.
Plaintiff telephoned Roach several times to complain about the performance of the automobile. He was assured that it was operating normally and was told to keep a list so the necessary adjustments could be made when the car was brought in for the 1000-mile checkup. On Saturday, May 30th, plaintiff telephoned Roach again and complained vigorously because nothing had been done. He asked to be allowed to bring the car in, but was told that Roach could not service it then or on Monday. Roach instructed him to bring the car in on Tuesday, which he agreed to do. The accident occurred in the early hours of Sunday, May 31, 1959. This was 29 days after Roach delivered the automobile to plaintiff, and it had been driven less than 1,000 miles at the time.
By their answers to the primary liability issues, the jury found: (1) that in purchasing the car, plaintiff relied upon Roach’s warranty that the vehicle was free from defects, (2) that it was not free from defects, and (3) this was a proximate cause of the accident; (4) that Ford and Roach represented to plaintiff that when the engine of the automobile stopped or the power steering was not operating normally, he would have safe steering and full control of the car with conventional steering, (5) that plaintiff relied upon such representation, (6) which was untrue as applied to the steering system at the speed at which the car was moving just before the accident, and (7-8) the making of the representation was negligence and a proximate cause of the accident; (9) that Ford and Roach represented to plaintiff that if the power brakes of the automobile lost their brake assisting power or were not operating normally, conventional brakes would remain fully effective to permit him to stop the car safely, (10) that plaintiff relied upon such representation, but (11) the same was not shown to be untrue by a preponderance of the evidence; (14) that the engine of the automobile died just before the collision, and (15) when that occurred, the braking and steering efficiency was reduced to such an extent that plaintiff was unable to safely control the car under the circumstances then existing, but (16) it was not established by a preponderance of the evidence that this was known to or should have been known by Ford and Roach; (20) that after delivery of the car, the engine died upon full application of the brakes, (21) that plaintiff requested Roach to make repairs or adjustments to prevent such dying, and (22) that Roach refused to make such repairs or adjustments, (23-24) which was negligence and a proximate cause of the accident.
The jury further found (42) that the “collision” was not the result of an unavoidable accident, (25) that plaintiff was acting under an emergency and (26) acted as a person of ordinary prudence would have under the circumstances, (37) that there was some risk involved in operating *265the car, but (38) that it was not established by a preponderance of the evidence that plaintiff had full knowledge and appreciation of the risks. It also refused to find: (29) that the plaintiff failed to keep a proper lookout or (27) failed to make proper application of his brakes, (31) that he was driving at excessive speed or (33) under the influence of liquor, or (35) that his failure to stop at the stop sign or (40) bring the vehicle to Roach for inspection and adjustment was negligence.
The judgment of the trial court against Roach thus rested upon breach of warranty as established by the jury’s answers to Special Issues Nos. 1-3, negligent misrepresentation as established by the answers to Special Issues Nos. 4-8, and negligent failure to repair as established by the answers to Special Issues Nos. 20-24. The judgment against Ford was based upon negligent misrepresentation as established by the answers to Special Issues Nos. 4-8. Plaintiff insists that the judgment against Ford may also rest upon breach of an implied warranty of fitness, but the Court of Civil Appeals concluded that he had not relied upon that theory in the trial court. The intermediate court held that there is no evidence to show any misrepresentation regarding the steering mechanism of the automobile, and that the limiting language in the dealer’s warranty was effective to exonerate Roach from liability for plaintiff’s injuries based upon breach of such warranty. It upheld the judgment against Roach upon the basis of the jury’s findings that the accident was proximately caused by the dealer’s negligent failure to repair.
An accurate and rather full statement of the testimony given by the various witnesses will be found in the opinion of the Court of Civil Appeals. In upholding the judgment against Roach, the intermediate court began with the premise that there is circumstantial evidence of probative force to support the jury’s finding that the engine of the automobile stopped firing just prior to the accident. It then observed that motor failure was the only possible explanation afforded by the evidence for plaintiff’s loss of power brakes and power steering, and concluded that the jury was entitled to find that the accident probably would not have occurred if Roach had made the necessary adjustments and repairs to prevent the engine from dying. In our opinion the latter conclusion is not supported by the record in this case.
There is no direct testimony that the engine died before the automobile ran into the embankment. In an ex parte statement which was introduced in evidence, plaintiff said that he did not know whether the motor died but felt at the time that it had. Three expert witnesses testified. Plaintiff offered Professor James O. Melton of the University of Oklahoma. Ford and Roach countered with the testimony of Mr. Richard J. Chwalek and Professor Robert E. Martin. None of these witnesses undertook to say, in response/to a hypothetical question or otherwise, that the facts as related by plaintiff indicated that the engine had ceased firing. Professor Melton expressed the opinion that the difficulty plaintiff experienced probably was due to “an additional input of air into the system so that you have a loss of vacuum in some manner.” 1 He did not express any opinion based upon reasonable probability as to the mechanical cause of the additional input of air or the loss of vacuum.
The conclusion that the engine failed is itself an inference which must be drawn from plaintiff’s previous experience with the vehicle and, more importantly, the loss of power on the occasion in question. Once that conclusion is reached, plaintiff would have us go full circle and say that engine failure caused the loss of the power assist for the brakes and steering, but even that is not enough. To sustain a recovery against Roach on the theory of negligent failure to repair, we must take one further step and *266say that engine failure was the cause of plaintiff’s inability to stop or turn the vehicle. It is here that the proof fails.
The undisputed evidence shows that a power-equipped 1959 Thunderbird has a vacuum reservoir which allows the driver to make two or three power-assisted applications of the brakes after the engine stops turning. Even when this reservoir is filled with air at outside pressure, the driver has conventional brakes which, if he applies to the brake pedal the force normally required for a vehicle not equipped with power, will enable him to slow or stop his automobile as quickly and effectively as with the power assist. It also appears from the undisputed evidence that the steering gear ratio of a 1959 Thunderbird equipped with power steering is the same as that of one not so equipped. In either vehicle one pound of force applied at the rim of the steering wheel produces 30 pounds of force at the end of the pitman arm. When the automobile is moving at a speed of 30 miles an hour, approximately 150 pounds of force at the end of the pitman arm is required to turn the front wheels. If the vehicle is equipped with power steering, turning the steering wheel in either direction causes the linkage to move and actuates the valve assembly. This permits fluid under pressure to flow to the appropriate side of a cylinder, which supplies part of the force required to turn the front wheels.
As pointed out by the Court of Civil Appeals, there is no evidence of any defect in the conventional braking or steering systems or the power assist of either that would cause the same not to operate. Plaintiff had no difficulty with the brakes or the steering prior to the accident. A friend who was driving the automobile on one occasion when the motor died said that he had trouble steering, and then went on to explain:
“It wasn’t the first time I had driven it, and I was used to the power steering. All of a sudden when it goes out, it was quite a change, and I could feel the weight of the car and the weight of the engine and so on; whereas, before, I could not. There was very little effort to turn the corner when the engine was running, but when the engine stopped, evidently the power steering went out and it just caught me off guard and, of course, nothing happened. I had control of the car. There were no other cars around, but there was a very definite difference in the pull of the car.”
An appraiser who examined the automobile after the accident noticed a dent in the power steering valve assembly. When asked to describe the dent, which the witness assumed was caused by the wreck, he answered:
“It was damaged from a blow. If memory serves me correctly, it was damaged from a blow of concrete or — anyway, it was marked like if you run into a curb or concrete, you will skin it. It is not a smooth skin but it is an indentated skin.”
The witness thought the valve assembly would have to be replaced. It appears from the undisputed evidence, however, that no repairs were made to either the power braking or power steering system after the accident, and yet the automobile operated satisfactorily with no difficulty in the brakes or steering. The only defect in either material or workmanship fairly raised by the evidence as having been present when plaintiff purchased the automobile was an improper adjustment of the idle-speed screw on the carburetor. The jury could reasonably conclude that this caused rough idling and engine failure at low speeds, but our question is whether the evidence will support the further conclusion that it caused the accident and plaintiff’s injuries.
The motor of plaintiff’s automobile had died on five or six previous occasions. This occurred when he was backing out of the driveway at his home, pulling up to a stop sign, applying the brakes or standing in traffic with the motor idling. According to his testimony, “two instances that *267I can think of for sure in traffic was during idling,” and “when this would happen when the car was in motion, it was such a minute motion that it didn’t make too much difference.” On the basis of plaintiff’s experience then, the motor died when it was idling or when the vehicle was moving very slowly. Neither of these conditions obtained at the time the engine is said to have stopped just prior to the accident.
Plaintiff testified that he was driving about 30 or 35 miles an hour when he observed the road intersection. He may have been somewhat high in this estimate, but it is clear that the engine was turning, and the automobile was moving, at a fairly rapid rate. These are not the circumstances under which motor failure had occurred in the past. Of even greater significance is the fact that plaintiff got only “a slight response” when he first applied the brakes. Although plaintiff pushed on the brake pedal as hard as he could, the wheels of the car did not lock and skid; the speed of the vehicle was not reduced. Under the facts as related by plaintiff, the motor almost certainly was turning and firing at the instant the brakes were applied, and there is no reasonable basis for concluding that the initial malfunction of the brakes was caused by engine failure. The motor may well have stopped firing shortly after the brakes were first applied, but the record is devoid of any evidence to warrant the belief that this would or even could have.impaired or affected the operation of the conventional brakes. It is clear, moreover, that a person of plaintiff’s size and physical condition would have no difficulty in making a full application of the conventional brakes even without help from the power assist. A dying engine does not account, therefore, for his inability to stop the automobile before it ran into the embankment.
When the automobile reached the intersection, plaintiff attempted to turn to the right. He was about 5 feet 8½ inches tall, weighed 155 or 160 pounds, and was in good physical condition. He had both hands on the steering wheel and used both in attempting to turn the automobile. Although he “turned that steering wheel with everything” he had, exerting the maximum effort of which he was physically capable, the wheel would move only some two or three inches. It then locked and would not turn further. Plaintiff was thus able to turn the car only “a very slight bit,” and it went across Alief Road and into the embankment.
The witness Chwalek is an experimental engineer employed by the Bendix Corporation, which manufactured the power steering system used on the 1959 Thunderbird. He testified that when a power-equipped 1959 Thunderbird is moving in gear at a speed of 30 to 35 miles an hour, the engine will be turning even though it is not firing and the power steering will thus be operating. Professor Martin, who is associate professor of Mechanical Engineering at Texas Technological College, had made a number of tests with a 1959 Thunderbird. He found that when the vehicle was moving in gear at speeds of from 20 to 50 miles per hour, turning off the ignition did not affect either the power brakes or the power steering. Professor Melton said that while the engine was likely to be turning when the ve- • hide was moving 35 miles an hour, it might or might not be turning at 30 miles per hour. When the evidence is viewed in the light most favorable to plaintiff, we think the jury was entitled to conclude that the power steering was not operating when he attempted to make his turn. This does not, however, explain his inability to make the turn with conventional steering.
Mr. Chwalek was responsible for the development of the power steering system used on the 1959 Thunderbird. He testified that in developing the equipment, consideration was given to problems that might *268be encountered if the engine stopped firing, Tests were made to determine what effect this would have, and the results of these tests were as follows:
Speed of vehicle 30 mph 10 mph 5 mph 0 mph
Effort required to turn steering wheel with power steering operative 3-6 lbs. 3-6 lbs. 3-6 lbs. 6-10 lbs.
Effort required to turn steering wheel with power steering on vehicle but not operative 4-8 lbs. 8-15 lbs. 12-18 lbs. 50 lbs. approx.
Effort required to turn steering wheel on Thunderbird not equipped with power steering 4-8 lbs. 8-15 lbs. 12-18 lbs. SO lbs. approx.
According to the tests conducted by this witness, there is no perceptible difference in the amount of force required to turn the steering wheel of a 1959 Thunderbird equipped with power steering that is not operating and the amount required to turn the steering wheel of such a vehicle not equipped with power steering. This was confirmed by tests made on a 1959 Thunderbird by Professor Martin. He found that the power-equipped 1959 Thunderbird, when moving with the gear in neutral and the ignition turned off, “operated just like a car with conventional steering and conventional brakes.” To determine whether additional force was required to move the power steering fluid through the linkage when the system was not operating, he placed the vehicle on a grease rack with its wheels raised off the floor. He was able to spin the steering wheel from one extreme to the other with his little finger when the engine was running and when it was not running. According to his testimony, there was no measurable difference in the amount of force required.
On the other hand, Professor Melton testified that it is more difficult to steer a power-equipped automobile when the engine is dead than one which does not have the power assist. He stated that this is due to the additional force required to move the power steering fluid through the linkage. The witness also said that he had tested a 1959 Thunderbird raised on a grease rack with its motor off, and that the steering wheel could not be turned with the little finger. He “believed” that he had measured the force required, but did not “remember the numbers. In fact, I believe these numbers went into our statistical analysis.”
According to Professor Melton’s testimony, “the normal force necessary on a conventional car is about 12 pounds at the wheel, about 7 pounds with power steering and with the power steering assisting, and that will run up, depending on the automobile itself, as much as 18 to 36 pounds and even beyond when the power steering is inoperative.” These figures are the averages of a series of tests made by the witness upon various automobiles. He had no records of *269the tests with him, and was unable to give the readings for any particular vehicle at any particular speed. The automobiles tested included a 1957, a 1959 and a 1960 Thunderbird, but they also included an undisclosed number of vehicles made by General Motors. It appears from the evidence that some automobiles equipped with power steering do not have the same steering gear ratio as the identical make and model when not equipped with power steering. Aside from the three Thunderbirds, the facts concerning the steering gear ratios of the vehicles tested by Professor Melton are not shown. And while his testimony as to the speed at which the tests were made is quite vague and contradictory,2 it is clear that at least some and perhaps all of the readings were taken while the vehicles were standing still.
Professor Melton’s tests and figures are thus of no value in determining how a power-equipped 1959 Thunderbird might perform when moving at 30 or 35 miles an hour with its engine dead. He was unable to express any opinion as to the additional force required to turn such a vehicle under those circumstances. When interrogated on that point, he answered “I would guess it could take up to maybe twenty or thirty extra pounds, but I don’t know.” The jury could only speculate then as to what the witness meant when he said that “considerably more effort” is required to steer a power-equipped vehicle with its engine dead than a similar vehicle not equipped with the power assist. Be that as it may, there is nothing in his testimony or elsewhere in the record to support the conclusion that engine failure would probably render a 1959 Thunderbird, moving at any speed approaching 30 or 35 miles an hour, so difficult to steer that a driver of plaintiff’s size and physical condition would be unable to guide and turn the same with conventional steering. Under the evidence, a dying engine simply will not account for plaintiff’s inability to turn the steering wheel more than two or three inches.
*270The idle-speed screw of the carburetor may not have been properly adjusted when the automobile was delivered to plaintiff, but that is the only defect existing at that time which the jury could reasonably find under the present record. In our opinion there is no evidence to warrant the belief that the accident was caused by this defect or by Roach’s failure to repair the same. Without proof as to what did cause the brakes and steering not to function, the jury had no way of knowing whether the failure of either was due to a defect in the car at the time of its delivery. Liability may not be predicated then upon any theory of either breach of warranty or negligent refusal to repair.
The jury also found in response to Special Issue No. 6 that the representation in the owner’s manual concerning the steering was false. In discussing such representation, the Court of Civil Appeals stated that the manual made it plain that 75% more effort would be required to operate the steering when the power system was not working. A fair and more accurate construction of the manual is that the driver will find it necessary to exert up to four times as much effort to steer when the power is lost.3 Subject to this relatively minor correction, we agree with the Court of Civil Appeals for the reasons set out in its opinion that there is no evidence to support the jury’s answer to Special Issue No. 6. This leads to the conclusion that judgment should have been rendered in favor of both Roach and Ford.
The judgment of the Court of Civil Appeals is accordingly modified so as to reverse the judgment of the trial court and render judgment that plaintiff take nothing. As so modified, the judgment of the Court of Civil Appeals is affirmed.
SMITH and POPE, JJ., dissenting.I. Emphasis throughout this opinion is supplied.
. Q. What speeds were you running these tests at?
A. We ran these through a series of different speeds. I probably could make this much more clear to you with somewhat of an explanation.
Q. Well, sir, you will be afforded an opportunity to do that. What I am asking you is: What speeds were you using when you got these figures of twelve, seven and eighteen to thirty-six that you mentioned a little earlier?
A. I believe that these speeds were essentially at — Well, there were, in some cases, vehicles that were stopped. In other cases, they were somewhere between thirty and forty miles an hour. There were a series of tests.
Q. These figures you have given us: twelve, seven and eighteen to thirty-six, those were done with vehicles stopped, weren’t they?
A. Not necessarily.
Q. Well, were they?
A. Not all of them.
Q. And is the truth of the matter that those figures of twelve, seven and eighteen to thirty-six, were obtained with the automobile sitting dead still?
A. In some cases.
Q. Well, sir, you have given me one set of figures: twelve, seven, eighteen to thirty-six. Were those three figures obtained at different speeds?
A. These are average figures.
Q. Average of what?
A. Of a series of tests.
Q. What are you averaging?
A. I am averaging values.
Q. Now, all these figures are different, depending on what speed the automobile is going, aren’t they?
A. Yes, sir.
Q. What I am asking you is a very simple question: How fast was the automobile going when you got those figures?
A. These figures essentially are fixed figures. In other words, the vehicle isn’t moving.
Q. That is exactly what I have been trying to find out. This is when the automobile is standing still.
A. That is not necessarily true.
Q. Didn’t you just say they are figures that are fixed, when the car is fixed and the engine not running?
A. Essentially, yes.
. The manual states that “up to 75% of the effort needed to steer your Thunderbird is taken over by the Master-Guide Power Steering.” If 75% of the required effort is supplied by the power steering, the driver supplies the remaining 25%. When the power assist is lost, the driver must furnish 100% of the required effort, or four times as much as with the power assist operative.