If a motor car runs off the road, that fact, standing alone and unexplained, provides some evidence that the driver was negligent. But here much more was known than that the vehicle ran off the road. The occurrence was not unexplained. That the steering mechanism had suddenly failed was not in dispute. Both sides accepted it as the fact. Therefore the only way in which any place could be found for res ipsa loquitur would be if negligence on the part of the driver could be inferred from the unexpected failure of the steering mechanism. What went wrong, what part became loose or broken was not said. No attention was given to this at the trial. No one suggested then that the defendant was in any way responsible for the failure of the steering. But afterwards it was argued, and the argument found favour with the majority of the Supreme Court, that this was something for which the defendant could have been held responsible; and that, as this was not put to the jury, there should now be a new trial. It seems that, when the trial took place, the decision of the Supreme Court in Priest v. Arcos Enterprises Pty. Ltd. [1] was generally accepted in New South Wales as defining the scope of the doctrine of res ipsa loquitur in a way later shewn to have been mistaken. The misapprehension prevailed, it is said, until the decision of this Court in Anchor Products Ltd. v. Hedges [2] . The trial, it is said, had thus proceeded under a cloud and, the darkness having been now dispelled, there should be a new trial in the light. The proposition is quite remarkable if it means that the verdict of a jury on an issue of fact, submitted to them without objection by either side, is to be set aside because one party might have put his case differently from the way he chose to do. It is no less remarkable if it means a new trial can be had on a suggestion that the senior and experienced counsel who appeared for the plaintiff at the trial had refrained from asking the judge to put a matter to the jury because he thought he would decline to do so. It may be that the Supreme Court in banc could, and in some circumstances should, in the proper exercise of its discretion to control trials at nisi prius, order a new trial when a trial had been conducted in accordance with a ruling that Court had given which was later reversed on appeal. But there is much more to this case than that; and I think we need not make any general statement; for, with all respect to the Supreme Court, I cannot accept the initial proposition that, if the steering mechanism of a car - an old car, if you like - suddenly fails, that, without more, is evidence that the driver was negligent.