dissenting. Being of the opinion that the judgment of the Court of Appeals is based entirely on the question of the sufficiency or insufficiency of the evidence to support the ruling of the trial court, and that question not being one of public importance or gravity, the writ of certiorari was improvidently granted. I would dismiss the petition for the writ.

Louis Friedman filed his complaint against Royal-Crown Bottling Company, a partnership, seeking to recover damages for alleged personal injuries. Count 1 alleged that the defendants had impliedly warranted to the plaintiff that the bottle of carbonated beverage was merchantable and reasonably suited to the use intended and that the defendants knew of no latent defects undisclosed, to wit: that the carbonated beverage was a mixture suitable for human consumption and would not cause its container to explode or fragment. Count 2 alleged that the defendants impliedly warranted to the plaintiff that the glass bottles containing the beverage were merchantable and reasonably suited to the use intended. Count 3 alleged that the defendants impliedly warranted that the glass containers containing the carbonated beverage would not explode or fragment during their keeping, storage and handling by the plaintiff, *723under ordinary and customary circumstances and conditions. Count 4 alleged that the defendants were negligent in not properly inspecting and testing the bottle prior to filling it with a carbonated beverage.

On the trial of the case the motion of the defendants for a directed verdict was overruled and the jury returned a general verdict in favor of the plaintiff. On appeal to the Court of Appeals, that court held that the trial court erred in overruling the defendants’ motion for a directed verdict. Kennedy v. Friedman, 123 Ga. App. 105 (179 SE2d 566). The court, in its opinion (p. 108), said: "Here the plaintiff had the burden to produce evidence on these critical elements of his case. From the time the bottle left the manufacturer, only three people had access to it. Nevertheless, plaintiff failed to show that two of them had not tampered with, banged, or otherwise handled the bottle in such a way that it could have caused the explosion. For this reason, plaintiff failed to prove a prima facie case either in negligence or warranty” and in its headnote of the opinion ruled: "In a products liability action, whether based on negligence or breach of implied warranty, the plaintiff must account for the product from the time it leaves the defendant’s hands until the time of the occurrence in order to eliminate other explanations than the defendant’s negligence (if plaintiff is relying upon res ipsa loquitur) or to prove by inference that the product was defective when sold.”

We granted the plaintiff application for the writ of certiorari, and we now reverse.

It appears without dispute that the plaintiff purchased 24 bottles of Diet-Rite Cola beverage from the defendants; that he carried them to his home and placed one six-pack carton on a pantry shelf in his kitchen; that on May 21, 1962, he went into the pantry and as he removed a bottle from the carton, the bottle exploded causing injuries to his person.

There was evidence that the plaintiff’s wife and the housemaid had access to the pantry where the bottle that exploded was located, but the record is silent as to whether either one prior to the explosion had or had not handled or touched the bottle that exploded.

*724On the trial, glass fragments of the bottle that had exploded were introduced in evidence. Earl W. Toulouse, Regional Service Engineer for Owens-Illinois Company, on direct examination as a witness for the defendants, testified as follows: That he determined from the glass fragments of the bottle that the bottle was manufactured by Owens-Illinois. He described the method and standard employed by that company and the test to which the bottles were subjected. He stated that after he had reassembled the glass fragments of the broken bottle he had reached the conclusion that the cause of the bottle breaking was an impact or external pressure and not internal pressure.

Dr. Oscar G. Fryer, consultant for the Science Center of Drury College, a witness for the defendants, testified that he had examined the fragments of the broken bottle and in his opinion, the bottle was broken by reason of external pressure.

On cross examination the witness Toulouse testified as follows: "Q. What other kind of examination did you make of this glass bottle here, sir? A. This particular bottle? Q. Yes, sir. A. We examined the — well, of course, all of the pieces we examined for the detravel, we examined the — what might have been left of any stress that may have been in it, uneven stress. Q. Part of the glass was pulverized, was it not? A. That’s correct. Q. Did you determine anything on that account? A. No, other than we could tell that most of the pulverized glass came from the — directly behind the point of impact, by determining this because that was the only portion of the glass, of the bottle itself where the pieces, except for one or two places, where goodly size pieces were missing. Just two, I think two. And the other pulverized glass, of course, too small to put back together or anything like this, but knowing from other tests that we have made and knowing that the pulverized pieces that we get come from directly behind the point of impact where the stress is the greatest. Q. Yes, sir. Did you find any internal strains in' that bottle? A. It has what we classify as a number two, this is extra-polated to our standards, then compared with other bottles that had broken, the strain that remained — now, we did this with a polariscope when we did this in Toledo in 1966. We did *725not do that with this one right now, but at that time with what we call a number two. Q. You did find internal strains in the glass? A. That’s correct. This is normal. Q. Would a combination of internal strain and external scratching with carbonated beverage being contained there, wouldn’t that accentuate any one of those defects? A. It would add to it. Now, by accentuating, I might have to say relative there, too, because this amount of strain is insignificant in this respect because the other strains from impact are so high that this is lost, this amount. Q. All right. But you talked about, a while ago, of internal pressure by the carbonated beverages on the inside exerting pressure on the outside? A. That’s correct, yes. Q. Because of the pressure pushing out? A. Uh, huh. Q. All right. And then if you strike it on the outside it is more inclined to break or explode with these defects than it is without them, is it not? A. Right, correct. Q. Yes, sir. And it wouldn’t take as much of an impact with these defects in the bottle as it would with a good sound bottle without any of these defects, would it? A. It wouldn’t take as much, and then again you have to get back to the pattern itself and relate this to fracture analysis to determine whether or not — which was involved and how much. Q. But it is a relative term? A. It is a relative term, yes. Q. And the defects. . . A. And it depends on the individual bottle itself — that is the pattern of the fracture itself. Q. Did you ever experience any bottle exploding at your factory because of scratches internally caused during the manufacturing process? A. Yes, we have — where they are badly annealed, where they haven’t been annealed properly, you can scratch the inside surface and they will break. They won’t explode, but they’ll crack. Q. What do you mean by annealed? A. All glass that is cooled fast sets up a strain in the glass itself. Now this is where the annealing layers or annealing ovens as they are known in some places, we call them layers, where the bottle is reheated up to a point just shy of the melting point so it won’t slump and go out of shape, reheated to that point, and then gradually cooled through the annealing point, which is a particular temperature.”

"Where an action is brought to recover damages for an injury caused by the explosion of a bottle, the contents of which were *726manufactured, bottled, and sold by the defendant as a harmless beverage, an inference of negligence on the part of the manufacturer arises, when it is shown that all the persons through whose hands the bottle had passed were free from fault, and that the condition of the bottle and its contents had not been changed since it left the defendant’s possession.” Payne v. Rome Coca-Cola Bottling Co., 10 Ga. App. 762 (73 SE 1087).

The jury under the testimony of the defendants’ witness Toulouse, on cross examination, could have found that the bottle that exploded was defective when it left the hands of the defendants. The principal reason for granting certiorari in this case was that the complainant’s case was not only based on the doctrine of res ipsa loquitur but was based on the fact that there was direct proof going to show the defective character of the bottle. In such a case the rule is as stated by the Court of Appeals, through Judge Jenkins: "Under the state of facts testified to, the plaintiff in this case was not dependent upon the application of the doctrine res ipsa loquitur in order to establish his case, since there was direct proof going to show the defective character of the bottle.” (Emphasis supplied.) We think the Court of Appeals was right in holding that the evidence did not authorize a verdict under the doctrine of res ipsa loquitur for the reason it stated, but we think that the Court of Appeals was wrong for not ruling that a verdict was authorized under the theory stated above in Atlanta Coca-Cola Bottling Co. v. Danneman, 25 Ga. App. 43 (4) (102 SE 542).

The Court of Appeals erred in holding that the trial court erred in not directing a verdict on motion of the defendants.

Judgment reversed.