The opinion of the court was delivered by

The plaintiff seeks to recover damages for injuries caused by an explosion, which he attributes to gasoline. The defendant’s liability is rested upon the theory that it delivered gasoline in place of water-white kerosene oil, which the plaintiff had bought. The proof is circumstantial only, and the only question is whether the facts proven justified a submission of the case to the jury.

The oil was bought December 5th, 1902, of the driver of a tank wagon of the defendant, and seventy gallons were carried from the wagon through the storeroom of the plaintiff in open pails holding five gallons each; two pails were carried at a time, making seven trips necessary. The storeroom was twenty by thirty feet in size, heated by a stove. The oil Avas poured into a drip pan at the rear of the store, from Avhich it ran through a pipe in the floor to a tank of one hundred and thirty-five gallons capacity, located in the cellar. No attention was attracted by any odor of gasoline during the process of delivery, although the testimony shows that it has a marked and distinguishable pungent odor and *152was necessarily carried through a small storeroom and close to persons who were likely to recognize the odor of gasoline. Among other sales of oil from the tank made during the day were sales to Morris and Lodge, two customers of the plaintiff. Mrs. Morris testified that upon lighting her lamp it flamed up and there was a peculiar blue flame; -Mrs. Lodge that she used the oil in four lamps and the flame was such that seven lamp chimneys on two lamps broke in the course of a few days, but it ivas shown that her husband used the same oil for kindling fires, just as he had used kerosene, without noticing any difference, and the whole five gallons bought by them was used in the ordinary way without any difficulty other than the breaking of the seven lamp chimnejrs.

The seventy gallons delivered to the plaintiff was a part of six hundred and thirty-five gallons in the tank wagon, allot which was sold to customers of the.defendant on the same day ra varying quantities, as their needs required. Several of these customers were produced and testified that they had had no complaints of the oil.

After the explosion a considerable quantity of the oil remained in the tank in defendant’s cellar, apparently unaffected by the explosion, and this oil was taken shortly after the explosion to a new store in charge of the plaintiff’s brother; it was tried and seemed all right, and part of it was used by them for lighting and part sold to their customers for the same purpose. There is no proof of any injurious effects resulting from the use of this oil. No one seems to have taken a sample for analysis.

The storeroom was lighted by two gasoline lamps and one kerosene lamp; the latter was at the rear of the store near the drip pan through which the oil was delivered to the tank in the cellar, and near a pump used to draw oil from the tank when needed for sale. The plaintiff dealt also in gasoline, which was stored outside the building. Upon shelves in the rear of the storeroom were small quantities of black, powder and of smokeless powder.

The explosion occurred late in the afternoon, just as the *153plaintiff was in the act of lighting the kerosene lamp with an ordinary match. One of the gasoline lamps had been' lighted shortly before — the plaintiff says three minutes— and a naked flame from alcohol in a little cup under the burner had been used for a minute or a minute and a half for the purpose of heating the gasoline and no explosion had occurred. All the lamps hung about five and a half feet from the floor.

AAre think the evidence is sufficient to justify an inference that the explosion was due to gasoline, but not sufficient to justify an inference of liability on the part of the defendant, and that, on the contrary, the evidence shows, almost to the point of demonstration, that what was delivered by the defendant’s driver was not gasoline, but kerosene. The salient points are as follows:

1. The positive evidence of the driver and the man who helped him load the tank wagon that it was loaded with kerosene only.

2. The absence of proof of complaint from any other customer to whom oil was sold out of the same bulk of six hun-dred and thirty-five gallons, and the proof from several customers that they had sold it as kerosene without having a complaint. Had it been gasoline, someone would surely have detected the odor.

3. The subsequent use of what was left of the oil in- the plaintiff’s tank and its sale by the plaintiff’s brother to customers as kerosene is very cogent evidence that it was actually kerosene.

4. The use of.some of the oil by Lodge for kindling fires without his discovering any difference from the kerosene hehad previously used, and without any injury from its use, and the use of the rest of the five gallons for lighting without any trouble except the breaking of seven lamp chimneys.

5. The absence of testimony- of any noticeable odor as the oil was carried in and emptied into the drip pan or at any time during the day.

6. The circumstances of the explosion itself. The testimony of experts on both sides is that the vapor of gasoline *154is heavier than the atmosphere and sinks. If it had been possible for the vapor which caused the explosion to rise from the tank in the cellar by diffusion through the pump and the drip pan, which seems very improbable in view of the specific gravity of the vapor, it is difficult to explain why the explosion did not occur until the plaintiff was in the act of lighting the lamp, or why it did not run to the tank and there cause combustion. The gasoline lamp had been first lighted, and if gasoline had been diffused through the atmosphere the explosion would probably have occurred then. If, however, the gasoline vapor came from the kerosene lamp which might have been filled with gasoline by mistake, the vapor in falling would meet the lighted match and the explosion would he accounted for.

The failure of either party to take a sample for analysis of the oil left in the tank is unaccountable. If any argument is to he drawn from that failure it must be adverse to the plaintiff, for he had the opportunity and the right to take the sample. The defendant, if it had the opportunity, which does not clearly appear, had no right to take the plaintiff’s property.

Upon the whole ease, we think the plaintiff failed to prove any liability on the part of the defendant. The rule must therefore be made absolute.