This appeal from a directed verdict for the defendant in a negligence action presents the question of how clear a case a plaintiff must show by circumstantial evidence in order to establish a jury question, when there is no direct evidence available to show what the actual events in dispute were. Early in the morning of February 6, 1952, J. E. Smith, Jr., left his Leesburg, Florida home for work in Orlando, Florida, in his 44-day old 1951 Oldsmobile 88 Super DeLuxe sedan. At approximately 5:30 A. M. he stopped at the Sunny Side Drive-in Restaurant in Leesburg, where he bought a pack of cigarettes and had two cups of coffee. About ten minutes later he left, and was never again seen alive.

Some time after 5:30 that morning J. M. Collins, bridgetender of the Dead River Bridge, between Leesburg and Tavares, Florida, heard a crash as he was dressing. He hurriedly pulled on the rest of his clothes and ran from his home at the east end of the 600-foot long bridge toward the west end. About 47 feet from the west end of the bridge lay the wreckage of Smith's car, facing in an almost southern direction, with Smith under the engine, dead. The right front wheel had been driven back further than the rear wheels. Hundreds of pieces of wreckage were scattered over the bridge. The generator had been hurled 288 feet down the bridge, and the car was so completely demolished that the wreckers were unable to salvage a single salable part out of it. It took two trucks to haul away the wreck, because it could not be towed. The wrecker driver testified to having handled over 2000 wrecked cars, and to have found only three or four in any more contorted condition. He also testified to the position of the speedometer hand, and that flesh like that of a finger or hand was caught in the radio panel, about two inches above the buttons.

Shortly after 6:00 o'clock, Douglas Sewell, Deputy Sheriff of Lake County, Florida, received a call at his home to investigate the accident. When he arrived at the scene, he found cars lined up for quite a distance back from the bridge, the wreckage having blocked traffic. As soon as the "investigating jury" finished its investigation, he and a group of bystanders began to clear the road, to get traffic going. In so doing, they picked up the pieces of the wreckage, and threw some into the wrecker's trucks and some off the bridge.

Neither the road approaching the bridge nor the bridge itself revealed any skid marks. However, on the shoulder of the road, about six inches off the pavement, and four or five feet from the bridge abutment, was a place where the grass and dirt had been torn up. Next to the abutment was a guard post to which a guard rail was attached. This ran alongside the road, about a foot from the pavement. The guard rail was completely destroyed at a point beginning about eight feet from the abutment. The abutment had been hit, and was left with the imprint of a headlight, about 18 inches from the edge thereof. The part of the abutment which had been hit had been knocked ten feet down a slope, and the concrete guard post next to it was pulverized. Wreckage was strewn from where the car hit to where it was found (a door from the right side of the car lay next to the guard rail), but there were no other marks except a rubber smudge on the north side of the bridge, behind the wreck.

It was dark at the time of the crash. There was some evidence that there were patches of fog in the vicinity.

That afternoon, William Alsobrook, out of curiosity, went down under the bridge and examined the pieces of wreckage that had been thrown there. He found a part of the steering relay rod, with the end adjustment plug partially screwed out, and the cotter pin gone. This member holds the idler arm ball stud and the tie rod ball stud in place. The tie rod was never found, but other parts of the assembly were. The purpose of this assembly is to assure the uniformity of action between right and left wheels, since there is no front axle. These pieces of physical evidence, plus expert testimony to interpret them, are the heart of the plaintiff's case.

The complaint originally alleged that by virtue of the end plug becoming loosened, the idler arm became disengaged from the relay rod as Smith approached the bridge, rendering the automobile uncontrollable and proximately causing the accident. The defendant's answer denied that there was any failure in the steering mechanism, and specifically that any part of the steering mechanism became disengaged prior to the accident. It also alleged contributory negligence generally, and in a separate defense contributory negligence in driving at a speed in excess of 90 miles per hour. Plaintiff then amended her complaint to add an additional count, alleging that the loosened end plug failed to hold the tie rod ball stud in its proper place, resulting in a loss of control permitting "a sudden and uncontrollable movement of the right front wheel," causing the accident. In an amended answer, the defendant denied this allegation, and reiterated its other defenses.

The cause then proceeded to trial, and at the conclusion of the plaintiff's case, the defendant moved for a directed verdict. This was granted as to the first count, but denied as to the second. The court reserved judgment on the motion as to the entire case. Plaintiff then asked leave to amend the first count, so as to eliminate the allegation that the idler arm had become disengaged from the relay rod. This was permitted, and the plaintiff alleged instead that the idler arm became loose to an extent rendering the steering uncontrollable.

The case was later submitted to a jury, and one of the five points in plaintiff's brief attacks an instruction of the court in its charge. The jury was unable to agree, however, and the court then considered the motion for a directed verdict, which it granted after argument. The major question presented on the appeal is the correctness of this ruling. However, plaintiff also alleges error in the exclusion of some of her proffered expert testimony, and in the inclusion of some of the defendant's. Appellant's evidence tended to show that the end plug was not loosened by the accident, but had been in a partially1 unscrewed position for at least some time prior thereto. There was also evidence by which plaintiff sought to show that there never had been a cotter pin inserted to secure the plug, or that if it had been inserted, it had not been properly spread, and later fell out. Plaintiff's evidence also sought to show that this could have escaped detection at the factory and in checks by the dealer, and that the unscrewing occurred gradually, so that Smith would not have been aware of the looseness before the time of the accident. Plaintiff's experts also testified that looseness in the steering mechanism could lead to loss of control, and could cause a front wheel "tramp" or hop. Plaintiff argues that this is what happened:

"One hop of this violent tramp, threw this car, uncontrollable, into a right hand dart and the metal guard railing then channeled this car and driver-occupant directly into the bridge abutment with the resultant tragic consequences."

He argues that this is the only reasonable hypothesis, pointing to the absence of skid marks, and to the scuffed-up shoulder. He attempts to show this by reference to the expert testimony adduced by both sides, and relies also on a physical equation to show the degree of looseness that obtained. The trial court was unimpressed with this evidence, finding that to submit it to a jury would be to permit the jury to speculate too freely on what actually happened when the crash occurred and then as to what might have brought these happenings about.

In its memorandum granting the defendant's motion, the court did not say exactly where it thought that the plaintiff's case failed. Its remarks at the argument indicate, however, that it assumed for the purpose of its action, that a looseness in the mechanism had been proved, and further that such looseness could cause a lack of steering control, which could cause such an accident. It held that no jury question was presented because there was no evidence showing that there probably had been a lack of steering control at the time of the accident, and that, if so, this lack of steering control had probably caused the accident.

Because of the absence of any eye-witnesses to the accident, the plaintiff's case here rests solely on circumstantial evidence. Taking pieces of demonstrative evidence from the wrecked automobile, she seeks by the interpretative testimony of various experts to reconstruct the events of the accident. The events urged as proved by inference do not encompass only the ten or fifteen-minute period following Smith's departure from the Sunny Side Drive-in and preceding Collins' discovery of the wrecked car. They begin earlier, with the assembly of the automobile, and continue throughout its entire 44-day life, including its previous operation by the deceased and periodic maintenance checks by Triangle Motors. The plaintiff's right to recover depends upon an answer in her favor by the fact-finder on each of the evidentiary propositions she would submit to it, and a failure to establish any link in the sequence of events from assembly to destruction, as she views those events, would disprove her case and expose her theory as false.

This is not to indicate that there is any rule which forbids a jury from reconstructing such a series of events or from finding "an inference on an inference." With each added factual proposition following only problematically from the one before it, however, the overall theory becomes more and more tenuous.

Florida law requires that the circumstantial evidence in such a case

"amounts to a preponderance of all reasonable inferences that can be drawn from the circumstances in evidence to the end that the evidence is not reasonably susceptible of two equally reasonable inferences." Mutual Life Ins. Co. of New York v. Hamilton, 5 Cir., 143 F.2d 726, 732.

This case and the rule as to circumstantial evidence there stated was cited with approval in the subsequent case of General Acc. Fire & Life Assur. Corp. v. Schero, 5 Cir., 160 F.2d 775, in which this court reversed the judgment of the trial court for failing to direct a verdict on circumstantial evidence which we held did not meet the test. In the Schero case the court cited as leading Florida decisions King v. Weis-Patterson Lumber Co., 124 Fla. 272, 168 So. 858, and Stigletts v. McDonald, 135 Fla. 385, 186 So. 233, 235. These cases all enunciate the same test as to the strength of the case that must be made out by a plaintiff in relying upon circumstantial evidence alone.

Turning now to the evidence, we must decide how likely plaintiff's hypothesis of the accident is. If we conclude, as the trial judge did, that his hypothesis is only a guess or a possibility, or is no more probable than one of several others, then we must affirm, because the court should then not have submitted the case to the jury.

While, as we have said, there is no absolute rule of law that forbids a jury from finding "an inference on an inference," here, however, the inference of the ultimate fact of causal negligence by the defendant, which plaintiff must prove to recover, is dependent upon not one other inference but four or five. The ultimate negligent act is that the manufacturer left out or negligently fitted, a cotter pin in the plug and barrel. There is no direct evidence of this fact; it must be inferred from the absence of a pin and the position of the plug after the accident, together with opinion evidence. The next fact that must be found in the chain of causation is that the absence of a cotter pin resulted in the unscrewing of the plug. There is no direct evidence of this fact but it is testified to by expert witnesses based on circumstantial evidence. The next fact is that the unscrewing of the plug caused a gap between the operating members in the right steering assembly. This was proven by direct evidence. The next fact is that such a gap would cause a looseness which created erratic action by the right wheel known as "tramp"; there is no direct evidence on this point but plaintiff sought to prove it by opinion evidence that under certain circumstances looseness of this kind would cause tramp; the next fact necessary to plaintiff's case is that this tramp caused the automobile to veer out of control and into the bridge. There is no direct evidence to this fact. The plaintiff sought to prove it by opinion evidence that if tramp occurred "anything could happen; you might have to stop the car completely to bring it under control."

Thus, it is apparent that the jury was expected to infer from circumstantial evidence alone (1) that the cotter pin was missing or was insecurely fitted when the car was built; (2) having inferred that fact, then that absence of the pin caused the plug to become partially unscrewed; (3) having inferred these two facts, then that a looseness resulted, which actually caused an erratic movement of the right front wheel at the time of the accident; (4) having assumed these three facts, that such movement was so violent that (5) it forced the car out of its course and into the bridge.

It may be helpful to analyze the logical process through which plaintiff says the jury should have been permitted to go for us to consider the type of proof offered to support each of the facts necessary to establish the chain of causation from the injury back to the alleged defective manufacture.

First, there is no proof that the decedent's automobile veered suddenly from the road. All that appears is that, beginning at a point seven or eight feet from the bridge, the metal guard rail, which parallels the road at a space of about one foot2 was completely torn away, a concrete post was pulverized and there was a torn-up place on the shoulder about four or five feet from the bridge.

Second, there is no direct evidence that the car was out of control of the driver just prior to its striking the guard rail. The only evidence as to this feature of the case is circumstantial. It will be discussed under "third" below.

Third, there is no direct evidence that the right wheel was in "tramp" or "shimmy" or was making any erratic movement just prior to striking the guard rail. The strongest evidence as to these two points was that given by witness Darnell. His testimony is not unambiguous or clear, but a careful reading of it discloses that the most he said was (a) that looseness in the tie rod connection affects the condition of tramp; (b) that such looseness would decrease the degree of control by the driver; that if a car went into a tramp or shimmy one would have very little control over the automobile; (d) one might have to stop the car completely to bring it under control; and (d) speed would affect that.3

For a full understanding of the lack of positiveness in Darnell's testimony, it is necessary that the exact language used by him be read in full.4

Here there is not even a clear and unambiguous opinion that if a car is driven at a stated speed with a looseness of the proven fraction of an inch that tramp follows or is even likely to follow. Moreover, Darnell testified that in his experience he found that one out of every five or six cars has a looseness in its steering assembly.

Fourth, there is, of course, no direct proof of the fact that the plug came unscrewed to the extent of three threads at any particular time or by reason of any particular circumstance. It is, however, fair to say, if it was proven that the cotter pin was left out at the factory, then the circumstantial evidence was strong enough to permit a fact finder to draw the inference that its absence permitted the plug slowly to "unwind"; it would also be sufficient to negative either an intentional or accidental unscrewing after the accident and before it was submitted to examination by the experts.

Fifth, there is no direct evidence that the cotter pin was left out of the plug and barrel during manufacture. The circumstantial evidence on this point included: (a) The absence of the pin after the accident; (b) The dissimilar appearances between the barrel in which the pin would be inserted and spread as to the left end of the barrel, in which the pin was still intact, and the right end of the barrel, in which the pin was still intact, and the right end of the barrel from which it was missing.5 Plaintiff's expert Lincoln, after testifying to the differences appearing in the metal at the two ends of the tube, was unable to give an opinion either that a cotter pin had been inserted or that it had not been inserted at the time of manufacture. He said:

"It was my conclusion that there had been no cotter key in it for some time prior to the accident, and that if there had been one in, it wasn't spread out as far as the other one, and it wasn't possible to say whether there ever had or had not been a cotter key in this hole."

(c) The fact that records as to the work done on deceased's car at the 1000 and 2000 mile checkup (the latter of which occurred some 1000 miles and 14 days prior to the accident) did not include any repair work that called specifically for adjustment in this plug or assembly. Among the circumstances which militated against the inference sought by plaintiff were (1) the fact that there was no positive proof that the car had not been serviced or repaired elsewhere than at the time of the two free checkups, (2) the testimony of the factory representative as to the inspection procedures at the assembly plant; (3) the presence of the slight marks on the barrel in juxtaposition to the hole intended for the cotter pin, and the opinion evidence of defendant's witness Dorschimer that the cotter pin is of softer metal than the barrel, and that it could be inserted and spread without necessarily leaving the type of marking found on the left end of the barrel; (4) the testimony of the witness Burleson. Although the testimony of this witness was sharply attacked by appellant on the basis of credibility, he did testify directly and positively that in lubricating this automobile, which he did on the pre-delivery, 1000 and 2000 mile checks, he inspected all the cotter pins in the steering mechanism and that all four of them were present at the time of the 2000 mile check.

We have discussed the nature of the evidence supporting the separate links by which plaintiffs sought to prove the alleged negligent omission of the cotter pin with the alleged mechanical failure of the automobile at the time of the collision with the bridge. We have done so with considerable particularity, and possibly at too great length, because plaintiff's counsel has presented a record, including physical exhibits and expert testimony, that is impressive in its completeness. Such a record and such obvious zeal and diligence on the part of counsel merit, and they have received, the most careful scrutiny and consideration by this court. Nonetheless the weakness of plaintiff's case is inherent in the absence of any direct proof or any circumstances tending to show by any preponderance that the death of plaintiff's husband was caused by a malfunctioning of his automobile. Appellant has pointed us to no case, and we know of none, in which the courts of Florida have found that a prima facie case of actionable negligence is made out by the plaintiff where he is injured or loses his life under such circumstances that no one can say what were the physical facts relative to the occurrence. In Mutual Life Ins. Co. of New York v. Hamilton, supra, this court recognized and gave effect to, as it was required to do, the Florida rule which provides that such a case can go to the jury only if

"the circumstantial evidence amounts to a preponderance of all reasonable inferences that can be drawn from the circumstances in evidence to the end that the evidence is not reasonably susceptible of two equally reasonable inferences." Mutual Life Ins. Co. of New York v. Hamilton, 5 Cir., 143 F.2d 726, 732.

Clearly, here, no one could say that the likelihood that Smith met his death as a result of a mechanical defect is any greater than that the collision with the bridge abutment resulted from human failure which, as reasonable men, we all know greatly preponderates as the cause of automobile accidents.

Even if this was not so, the same test would have to be applied as to each link in the chain of causation, and finally on the total hypothesis resulting from piling one inference on another. The criticism in the field of logic, even though such criticism does not result in an absolute legal prohibition against, piling inference on inference, stems from the very remoteness of the conclusion from the known facts. No more remote or tenuous conclusion could be imagined than that presented to support the theory of the plaintiff here, unless, in fact, a case were presented in which even the element of the partially unscrewed plug was lacking. To permit a recovery here would be but slightly more permissible than permitting recovery in a case in which no physical evidence even tending to show a mechanical defect was found, and the cause of the collision was unknown. From even such meagre circumstances could the plaintiff ask that the jury infer from the mere fact of a collision, unexplained by the defendant, that mechanical defects may cause erratic operation, that such defects may occur in the manufacture of an automobile, that the defendant as manufacturer may have caused such defects, and hence that he is liable to the plaintiff — in effect, relieving the plaintiff of his burden of proof?

We must find that the plaintiff here failed in her proof because she was unable to show that what might possibly have happened did probably happen to cause Smith's collision with the concrete abutment of the bridge. We also find that the necessity of piling inference on inference where there is no clear preponderance in favor of plaintiff's theory as to the existence of each of the links in the chain, makes this case on its facts one in which the trial court properly directed a verdict for the defendant.

Having reached this conclusion, the court finds it unnecessary to consider the other allegations of error specified by appellant. None of the evidence that appellant sought to introduce would have supplied the lack that we find existed. No alleged error in the court's charge had any effect on the proof of plaintiff's case.

The court correctly applied the law as to circumstantial evidence, and was compelled to find that there was no issue of fact to be presented to the jury.

Affirmed.