HOWELL, J.

The plaintiff appeals from a judgment entered after an involuntary nonsuit in an action based on strict liability of the defendant for manufacturing a defective safety latch on a truck.

Plaintiff was a truck driver employed by Pihl Transfer Company. The truck, a “cab-over” vehicle with a van body, had been purchased new by Pihl Transfer from defendant International Harvester Company approximately five months before the accident in question. It had been driven 5,000 miles and had been serviced about five times. In a “cab-over” truck, the cab tilts forward on its nose in order to expose the motor, which is directly under the cab. The cab is secured in front by hinges, allowing the raising *385 and lowering of the cab. In back, the cab is secured to the chassis by a cab, or main, latch which operates similarly to the latch on the hood of a regular automobile. The main latch is supplemented by a safety latch which also operates similarly to the safety latch on the hood of an automobile.

On June 2, 1967,- plaintiff was driving the truck empty over the Broadway bridge in Portland. The truck hit a rough spot, and the cab became disengaged and tilted completely forward. Plaintiff collided head-on with another vehicle.

Plaintiff brought this action on the theory of strict liability of the defendant for manufacturing a defective safety latch on the truck. He alleged in his complaint that the “latching mechanism and safety latch that are designed to secure the cab from tilting failed to do so by reason of defect in manufacture by the defendant in that the safety latch was improperly assembled and in that the hook rivet was not properly installed nor properly headed, and the latching bar was misaligned.”

The primary question is the sufficiency of plaintiff’s evidence of a manufacturing defect. This issue depends upon the testimony of a consulting engineer who examined the safety latch, and whether such testimony created a jury question of defective manufacturing of the safety latch, causing plaintiff’s injuries.

A description of the safety latch and its method of operation is necessary for an understanding of the expert testimony. When the cab is in place, it is locked to the chassis by the main latch and the safety latch. The cab is released by pulling a lever, or handle, behind and outside the cab. Pulling the handle releases the main latch first. The safety latch then engages the *386 latching bar, a steel bar about one-half inch in diameter attached to the frame of the truck below the cab. The safety latch is then released from the latching bar by pushing the same outside lever forward. The cab may then be tilted forward by one man lifting the cab. The tilting-forward process is facilitated by rubber springs called torsion rubber springs. To lower the cab, it is merely pushed down into position, and normally the main latch and safety latch will engage automatically, similar to latching the hood of an automobile.

The safety latch, a strip of metal about eight inches long with a hook on the end, is mounted on the underside and rear of the cab by two bolts. It is held rearward by a spring so that it will engage the latching bar and keep the cab from raising. The latching hook is released by the operating handle on the outside of the cab.

The hook portion of the safety latch is attached to the latch by a large rivet. The hook rotates up and down on the rivet. When the cab is lowered, the top of the hook grazes the latching bar, passes under it, and then extends past the latching bar to a position allowing it to catch onto the latching bar when the cab is raised or tilted forward.

John Talbott, a consulting engineer who had made examinations of the safety latch mechanism, was called as an adverse witness by plaintiff. On his direct testimony he stated that an examination of the safety latch revealed several long striations or marks on the metal on the bottom, side of the hook, resulting from steel rubbing on steel. He concluded that the hook had been caused to ride over the latching bar instead of going underneath it. He stated the overriding occurred “because the safety latch was improp *387 erly assembled at the time of manufacture in that the hook rivet was not pulled up tight enough and was not properly headed. This condition was aggravated by the misalignment of the latching bar in relation to the safety latch.”

On cross-examination by defense counsel, however, Talbott testified that he and an associate had made another examination of the safety latch at a later time. He stated that he had secured “new information” and had made “new findings.” He concluded that the looseness of the rivet in the safety latch allowed only side-to-side movement of the hook but did not allow the hook to extend rearward, which extension would result in the hook overriding the latching bar. He stated that the marks or striations on the top of the hook indicated that the latch had previously allowed the hook to operate properly by grazing the latching bar and passing under it as it should have done. He concluded that the safety latch had operated properly upon manufacture. He found that the overriding of the hook on the latching bar was caused by displacement of the hook extending too far to the rear, or the latching bar being too far forward, which occurred sometime after the safety latch had been operating properly. ①

The plaintiff argues that Talbott’s testimony on *388 direct examination and on cross-examination was conflicting, because on direct examination Talbott stated that the looseness of the rivet constituted a manufacturing defect; on cross-examination he testified that the looseness of the rivet permitted only lateral movement, not forward and rearward movement, and therefore was not causative of the failure of the hook to latch properly. Plaintiff relies on the general rule that when a witness’s testimony on cross-examination conflicts with his testimony on direct examination, the probative value of the testimony is for the jury. 58 Am Jur 491, 492, Witnesses, § 863. However, the rule does not apply where the witness’s former testimony is corrected or explained on cross-examination. 58 Am Jur 492, supra; 88 CJS Trial 490, § 214. See Washburn v. Simmons, 213 Or 418, 323 P2d 946, 325 P2d 255 (1958).

When Talbott testified on direct examination, he was referring to a first examination and report that he had made. When asked if he had come to a conclusion “then” he answered, “Yes, at that time.” Later, on cross-examination by defendant’s counsel, he stated that he had subsequently made new findings based on new information, particularly a photograph showing the latching bar to be bent. ② He also concluded that the loose rivet allowed only lateral movement of the hook which was not causative of the failure of the hook to engage the latching bar.

Talbott’s testimony regarding his subsequent examination and findings adequately explained and *389 clarified Ms initial contrary conclnsion. The plaintiff failed to produce direct evidence of a manufacturing defect and was not entitled to go to the jury based on his expert’s testimony.

Plaintiff also argues that the case should have been submitted to the jury under the theory that the latch did not perform in keeping with the reasonable expectations of the user, citing Heaton v. Ford Motor Co., 248 Or 467, 435 P2d 806 (1967). In Heaton we stated:

“In the type of case in which there is no evidence, direct or circumstantial, available to prove exactly what sort of manufacturing flaw existed, or exactly how the design was deficient, the plaintiff may nonetheless be able to establish his right to recover, by proving that the product did not perform in keeping with the reasonable expectations of the user. When it is shown that a product failed to meet the reasonable expectations of the user the inference is that there was some sort of defect, a precise definition of which is unnecessary. If the product failed under conditions coneermng which an average consumer of that product could have fairly definite expectations, then the jury would have a basis for making an informed judgment upon the existence of a defect. * * 248 Or at 471-72. (Footnotes omitted.)

In the instant case there was direct evidence, available and produced by plaintiff; however, such evidence established that there was no manufacturmg defect causative of the accident. The purpose of the rule regarding the effect of the product’s failure to meet the reasonable expectations of the user is to establish an Mferenee “that there was some sort of defect, a precise definition of which is unnecessary.” Heaton v. Ford Motor Co., supra at 472. Any infer *390 ence of a manufacturing defect causing the accident was dispelled by plaintiff’s own expert witness. Here, the problem was not the strength of the latch nor a defective design. The latch had operated properly until some time after manufacture and before the accident when it started to override the latching bar instead of passing underneath the bar.

Finally, plaintiff contends that while the improperly-headed loose rivet may not have been the “sole” cause of the accident, it was “a” proximate cause of the accident. In this regard plaintiff argues that its proof showed that the loose rivet was a manufacturing defect which was a cause combining with another defect “probably from defendant’s negligence in making repairs” to proximately cause the accident.

The evidence did not establish that the defective rivet was a cause of the failure of the safety latch. It is true that Talbott testified that the looseness of the rivet was not alone sufficient to have caused a malfunction of the latch mechanism, but that the looseness of the rivet when combined with the misalignment of the latching bar would cause the latch to malfunction. However, he also stated, as did the shop foreman, that looseness of the rivet permitted only lateral movement, not forward and rearward movement of the hook, and that the loose rivet was not causative. There was no evidence that the misalignment of the latching bar was a manufacturing defect. It is not sufficient merely to show malfunction of the latch from the rivet —the malfunction must have proximately caused the accident. 2 Restatement of Torts (Second), § 402A; Annot. 13 ALR3d 1057, 1084, Products Liability: Strict Liability in Tort. "While it is clear from a consideration of all the testimony that the rivet should *391 not have been loose, it is equally clear that the looseness permitted only lateral movement which was not responsible for the safety hook overriding the latching bar. Consequently, the loose rivet was not a substantial factor in the cause of the accident.

We believe that the evidence was insufficient to establish a manufacturing defect which was a substantial factor in bringing about the accident. The trial court properly allowed the involuntary nonsuit.

Affirmed.

Tongue, J., concurs in the result.

Notes

①

The defendant’s shop foreman, called as a witness by plaintiff, also testified that the looseness of the rivet would allow only-lateral, not forward and rearward, movement of the hook. A mechanic working for the defendant testified that he had worked on the cab a month or two before the accident and had replaced the torsion bar rubbers used to assist in raising the cab. He also removed the safety latch mechanism and pounded the head of the loose rivet to tighten it. He found the hook was not engaging the latching bar so he bent the latching bar forward so the hook would engage. He stated that the hook operated properly thereafter.

②

Plaintiff contends that no weight should be given to the photograph because Talbott could not identify the truck from the photograph of the latching bar. However, the photograph was received without objection, and Talbott stated at the time the exhibit was offered that it was a photograph of the truck involved in the accident.