—Appellee is a farmer, living about four and a balf miles north of tbe town of Lyons. Tbe buildings on tbe farm are located on a hillside. Tbe north side and west end of the barn are against tbe bill. Tbe bouse is about a third tbe size of the barn, and stands about three or four feet south of tbe barn, and a little to tbe west. Tbe bouse consists of a basement, built in tbe side of tbe hill, and a second floor, which is about even with *722the ground. The barn is about 18x30 feet, having a gable in the east end, with a hay door therein. At the time of the fire, the door to the haymow was open, and the haymow was nearly full of-hay, that had been put in about ten days or two weeks before the fire, and was damp at that time. The lower portion of the barn is 12 feet to the eaves, and the upper portion about 10 feet, making the haymow 22 feet above the ground. The kitchen is in the east end of the house. The appellant’s railway track extends north and south past the premises in question, and the center of the track is 140 feet east of the barn. Between the railway fence and the barn is an old shed, close to the fence, and there was also an accumulation of grass and weeds between the fence and the barn at that time. ' About 4:30 o’clock on the morning of October 23, 1916, an engine pulling a train of 12 cars, part of which were loaded and part empty, passed north on the appellant’s right of way, past appellee’s premises. About an hour later, the appellee came downstairs in his house, and entered the kitchen. At that time, there was a strong wind blowing from the northeast. The appellee testified that he cleaned the fire out in the cookstove that morning, and took the stovepipe, down, and cleaned out the stove; that, while engaged in this work, he noticed smoke coming into the house, and looking out, saw smoke coming from under the eaves of the barn, and noticed that the northwest corner of the barn was on fire. The evidence shows that, on the day of the fire in question, there were no .signs that there had been any fire in the grass and weeds between the right-of-way fence and the barn. The appellant’s railway track past the appellee’s buildings is on a down grade, and the evidence tended to show that the engine on appellant’s train at the time in question was not working hard, when it passed appellee’s premises. The engine, at the time, was burning Illinois coal, which the evidence tends to show is the best grade of coal for engine use. The evidence also tends to show that the engine, a-t said time, was equipped with what is known as the “Slater box,” or “spark arrester,” which is a netting, or box, placed in the engine in such a way that sparks or cinders coming from the fire box into the smoke box in the front of the engine must pass through the netting. The smokestack connects with the top of the Slater box. The *723evidence tends to show that the Slater bos on the engine at said time was in good condition, and that it had been recently inspected, and that none but small sparks or cinders could pass through the same.
The foregoing is a general statement of the substance of the evidence necessary for consideration of the question involved in the case. On this state of facts, the court directed the jury to return a verdict in favor of the appellant, and subsequently sustained a motion for new trial, on the sole ground that there was error in directing said verdict.
At this point, the question for our determination is whether or not the court erred in the first instance in directing a verdict in behalf of the appellant.
Appellee’s cause of action is predicated upon negligence. The appellee was required to produce evidence sufficient to go to the jury on the question as to whether or not the fire which destroyed his property was caused by a spark from the appellant’s engine. Was there sufficient evidence in behalf of the ap-pellee to carry this question to the jury? It is apparent from the foregoing statement of facts that the fire in appellee’s barn may have originated from one of two independent and separate causes. It is possible that it may have originated from a spark from appellant’s engine, which may have been carried by the strong Avind that was bloAving from the northeast, a distance of 140 feet, and entered the haymow, 26 feet above the track ■ and it may be possible that this could have been done, notwithstanding the equipment of the engine with the spark arrester, and notwithstanding the fact that there was no evidence of any fire in the dry grass and weeds between the track and the barn. It also may be true that, if such a spark was carried from the appellant’s engine into the east end of the haymow of the barn, which was 30 feet long, and was nearly full of hay, more or less damp, the evidence of fire would first be obvious near the eaves in the northwest corner of the barn. It is also possible that such fire may have originated from the passing train, approximately an hour before it was discovered.
But it is also possible that the fire in question may have originated from sparks from the stovepipe of appellee’s kitchen stove. The evidence shoAvs that appellee was engaged in raking *724the fire from said stove shortly before he discovered the fire jin the barn, and that the stovepipe from said stove was located but a few feet from the eaves of the barn where the fire was first discovered.
Which of the two theories as to the origin of said fire is the more plausible or reasonable is not for us to determine.
Before the appellee was entitled to recover from the appellant, the burden rested upon him to establish the fact that a fire was set by the locomotive of the appellant, and that his injury resulted therefrom. Upon the record as set forth, we do not think there was sufficient evidence to carry the case to the jury on the question whether the fire originated from sparks thrown from appellant’s engine, in view of all of the facts in the case.
In Neal v. Chicago, R. I. & P. R. Co., 129 Iowa 5, we said:
“If other conclusions may reasonably be drawn as to the cause of the injury from the facts in evidence than that contended for, the evidence does not support the conclusion sought to be drawn from it. Verdicts must have evidence to support them, and the jury will not be permitted merely to conjecture how the accident occurred. In matters of proof, they are not justified in inferring from mere possibility the existence of facts. ’ ’
In Gibson v. Iowa Cent. R. Co., 136 Iowa 415, we said:
“It is not enough for him to show that this might perhaps have been the cause of the injury. His case, so far as the second count.of the petition is concerned, rests wholly upon circumstantial evidence, and the rules applicable thereto are well understood. The circumstances shown must negative every other reasonable hypothesis save that of defendant’s negligence. Verdicts should not be based upon mere theory or supposition.”
In Eisentrager v. Great N. R. Co., 178 Iowa 713, we cited authorities in this and other states, and said:
“We agree, of course, that, when facts and circumstances are such that reasonable men, unaffected by bias or prejudice, may disagree as to the inference or conclusion to be drawn from them, there is a case for a jury. But it is one thing to have a state of facts from which differing conclusions may reasonably be drawn; quite another, to hold that one who has the burden *725of proving a given conclusion lias discharged the burden by-showing that a theory which sustains him is a possible one, if it also appear that a theory upon which his adversary would not be liable is just as possible.5 ’
It is unnecessary that we cite other authorities. A number are referred to in the Eisenirager case.
The burden rested upon the appellee to prove that the appellant set the fire that destroyed appellee’s barn. There was no direct evidence of this, and he relies wholly upon circumstantial evidence. The appellee did not carry the burden resting upon him to establish his theory that the fire originated from the appellant’s engine, under the circumstances proven, when it may fairly be said that the established circumstances are equally consistent with the theory that the fire originated from the appellee’s stove. It is not enough for the appellee to produce evidence that creates a surmise or conjecture that the fire may have originated as claimed by him, while, at the same time, the proof also creates a surmise or conjecture that it may have originated in another way. "Where the proof is equally balanced, or the facts are as consistent with one theory as with the other, a plaintiff has not met the burden which the law casts upon him. This is the well established rule of our previous cases; and, applying that rule to the facts of this case, it must be said that the appellee did not carry the burden of proof to sustain his claim that the fire that injured his property originated from the operation of appellant’s engine. The trial court was, therefore, right in the first instance, in directing a verdict in behalf of the appellant, and it was error to grant a new trial on the sole ground that the previous ruling had been erroneous.
In view of our holding on this branch of the case, it is unnecessary that we pass upon other alleged errors urged by the appellant. The cause will be remanded to the district court, with orders to set aside and vacate the order granting a new trial in said cause.- — Reversed.
Stevens, C. J., Evans and Arthur, JJ., concur.