JOHN R. BROWN, Circuit Judge.

The sole question here is whether the evidence was sufficient to sustain the jury verdict holding that Otis Elevator Company negligently failed to repair an elevator on which Robinson was injured.

Otis was not the owner of the elevator, nor was it the operator of it. Rather, the elevator was one of many actively in use in the new and modern Statler Hilton Hotel of Dallas. Otis had a maintenance contract with the Statler Hilton. No question is raised about the obligation or duty of Otis to persons who might be using the elevators. It is conceded that under its contract Otis had the duty to use reasonable care in the inspection, repair and maintenance of the elevators. And that duty would extend, of course, to employees of the Hotel as well as to guests.

Robinson, an employee of the Hotel, was engaged in the performance of his duties as a house security officer at the time of his injuries on June 1, 1956. He was injured when the No. 7 (middle) service elevator dropped at extraordinary speed from the 14th floor, suddenly coming to an abrupt stop at the 8th floor.

These were, of course, modern, high speed, automatically-controlled elevators. There was a bank of 6 elevators, 3 passenger and 3 service, back to back. The service elevators were numbered 6, 7 and 8 with No. 7 being in the middle. At this hour all of the elevators were on automatic control. Robinson pushed the usual "down" button in the corridor on floor 14. The car stopped. The door opened. He entered. The door shortly closed. He pushed the control button numbered for the street floor. As the car started moving, he heard a popping noise overhead, noticed unusually quick changes in the numbered floor signal lights over the door and felt the sensation of extraordinary speed. Since the car was dropping at excessive speed, he instinctly shoved the palm of his hand over a number of the floor signal buttons. The car stopped suddenly as though, he said, "it had hit a brick wall."

The weakness of the defense by Otis before the jury, as well as the development of its legal theory here, lies in the fact that in the final analysis it is really the assertion that the accident never happened. Testimony was offered by maintenance people describing the intricate control and built-in safety mechanisms of this modern elevator installation. Synchronized with a rated speed of so many hundred feet per minute, a succession of safety devices would automatically operate at successive stages. Had it occurred as claimed, the car would have been inoperable until certain of the safety devices were manually reset.

But the jury saw and heard Robinson. Not only that, it heard the testimony of another employee who had gone through a similar recent experience in that very same elevator. And it saw Otis service reports introduced as exhibits which reflected that complaints had been received respecting this or an adjacent elevator reporting substantially the same sort of an occurrence. It was entitled to, and did, find that the accident happened just as Robinson said it did.

Of course this is not the first time that something has occurred that simply could not have happened. The more complex becomes our machine age civilization the more this becomes evident. A little bird weighing but a few ounces, capable of flying at only a lumbering pace, may perhaps down a multi-ton, multi-million dollar supersonic airliner. Ships, visible on each other's radar screens on courses that will apparently clear as they plough through fog, come together in an awful crash. Indeed, it is the likelihood that the "impossible" will occur which augments the growing business of insurance, engages the professional attention of lawyers in behalf of victims and tortfeasors and furnishes to us and others the inexhaustible but well-nigh exhausting grist for the judicial mill.

In the face of this finding that the occurrence did actually happen, it is somewhat difficult to understand the position of Otis. Apparently it is that Robinson failed to prove specifically what was wrong with Elevator No. 7. In its thesis that seems to have been essential because in the absence of such proof, it would not be possible to determine whether or not Otis exercised ordinary care in its (a) discovery or (b) repair, i.e., adjustment or other corrective steps. To this is then apparently added the contention that unless such step-by-step process is followed and satisfied, it means that liability is being imposed under res ipsa loquitur on one not having exclusive custody and control of the instrument of harm, i.e., the elevator. But this misapprehends the nature of the duty which Otis owed and the legal approach which Texas takes to the question of proof including the specific problem of res ipsa loquitur.

There was evidence which the jury could credit that on this particular Elevator No. 7 another person had recently experienced a sudden over-speed dropping. And another similar incident had occurred though not reported to Otis. The records of Otis for May 29 and May 30, within 48 hours of Robinson's incident, reflected similar falling of these elevators. One incident referred to No. 7, the other to No. 8, but there were circumstances which would permit the jury to conclude that reference to No. 8 was erroneous and that No. 7 was intended. After each of these reports, a service call was made by an Otis employee. He could find nothing wrong and hence made no repairs. Thus it was in substance two more instances of Otis concluding that the reports of sudden falling were baseless — i.e., it did not happen because it could not have happened. In addition to the detailed service reports which showed the necessity for almost daily checking and the making of minor repairs or adjustments which tended to demonstrate that these elevators developed frequent trouble, it was acknowledged that due to design deficiencies in the mechanical room, there was a serious overheating problem which was not corrected for some months.

Otis was hired because it was an expert in elevator maintenance. Statler Hilton looked to it to take all necessary steps. Otis is the first to acknowledge that a falling elevator is indeed dangerous, and an elevator in proper condition simply won't do that. But this one did. Worse, it did it after Otis had received very recent reports that similar events had occurred either to this one, or to the adjacent one or both. Twice inspection "demonstrated" nothing was wrong. Twice no repairs were made. Shortly thereafter the same thing occurred for the third time. Cannot a reasonable man draw some inference from that?

Texas says he may. Proof by circumstantial evidence is not confined to the narrower limits of res ipsa loquitur which, as we have pointed out, "* * * is simply a facet of the general law that verdicts may rest upon circumstantial evidence." Dement v. Olin-Mathieson Chemical Corp., 5 Cir., 1960, 282 F.2d 76, 81. Circumstantial evidence may be used, and may often be a decisive factor on critical indispensable issues, in a variety of situations. Sam White Oldsmobile Co. v. Jones Apothecary, Inc., Tex.Civ.App. 1960, 337 S.W.2d 834 (error ref. n.r.e.); Roberts v. Texas Pacific R. Co., 1944, 142 Tex. 550, 180 S.W.2d 330; International G.N.R. v. Finger, Tex.Civ.App. 1929, 16 S.W.2d 132 (error dismissed); Comet Motor Freight Lines v. Holmel, Tex.Civ.App. 1943, 175 S.W.2d 464 (error ref. w.o.m.); Beaumont Iron Works v. Martin, Tex.Civ.App. 1945, 190 S.W.2d 491 (error ref. w.o.m.); McCray v. Galveston H. S.A. Ry. Co., 1896, 89 Tex. 168, 34 S.W. 95.

The jury must have concluded that three times is too much, once perhaps, twice maybe, but certainly no for the third time. Otis knew of the specific phenomenon. It did not remedy it the first time. It did not remedy it the second time. That, the jury could infer, shows a failure to exercise ordinary prudence in making the repair. Indeed, in many ways, the expertise of Otis may have convicted it in the jury's mind more by reason of this unexplained triple occurrence than had Robinson been able to prove the absence or defect of a particular screw or resistor, relay, vacuum tube or insulator. Proof would have had to come from an Otis employee. Otis was hired because it knew elevators. The jury may have reasoned that until Otis could come in and identify the cause (or all possible causes) and then show why this could not reasonably have been discovered, the failure of the expert to locate and correct the source of trouble showed neglect in the performance of the work in which it claimed pre-eminent competence. Otis never undertook to do this. It was content to urge that the occurrence never happened.

There was thus a solid basis for jury inference from circumstances which have a compelling attraction in the ordinary experience of men. That was enough.

Affirmed.