(concurring in part and dissenting in part). Even in the absence of direct proof, a jury may infer that a product was defective provided the plaintiff has “excluded all causes of the accident not attributable to defendant” (Halloran v Virginia Chems., 41 NY2d 386, 388; accord, Caprara v Chrysler Corp., 52 NY2d 114, 123). It is with respect to the contention that Driscoll’s testimony raised another possible cause of the defect, which plaintiff failed to eliminate, that I disagree. Driscoll stated that he “vaguely” remembered washing similar pots with a brush he described as a “curly handled type thing”, but couldn’t recall whether he had done so in the six months prior to the accident. Moroever, as the Trial Judge noted, the pot he cleaned was not necessarily the pot in question. Other evidence tended to show that this pot had never been washed with a wire brush. Plaintiff testified that she always washed the pots, and then not with a brush but with a sponge. Another waitress testified that she never saw anyone at the restaurant use a wire brush to clean the pots. All this testimony, which the jury and the court had the inestimable advantage of observing being uttered, simply presented a fact question for the jury as to whether the pot had been improperly handled. The jury’s resolution of that issue should not be disturbed. However, I agree with my colleagues that there was insufficient evidence presented on the issue of defective design of the coffee pot to submit this theory to the jury. Since certitude is lacking as to the principle of liability the jury relied upon, defective manufacture or defective design, in arriving at its verdict, I would reverse and remand for a new trial.