I. On May 31, 1925, George V. Friel owned and occupied a tract of about 116 acres south of Grand Junction in Greene County. The right of way of the M. & St. L. Railway Company over which appellant operates trains crosses the said tract north and south a short distance east of the dwelling and other improvements then situated thereon. Shortly after 1:30'P.M. on the date referred to, all of the improvements, except the dwelling and a small pump house, consisting of a barn, hog house, coal house, garage and granary, together with fences and trees adjacent to the barn, were totally destroyed by fire. The day in question was hot and dry and a very hard wind was blowing. The barn, which was 32x50 feet, was located 80 feet directly east of the dwelling house. Immediately east and adjoining the barn was a cow yard 46x32 feet. The southwest *151portion of the barn was occupied by a corn crib and driveway. Immediately east of the driveway was the hay barn and east of that an alley and stalls for horses. The north side of the barn was divided into two stalls and a cow bam. At the time of the fire a stallion and jack belonging to appellant occupied the stalls referred to. The stalls on the east side of the barn were occupied by two mares and there were one or more cows in the cow yard. There was a door on the north side of each of the stalls and cow barn and also to each of the other openings referred to. All of the doors on the south side of the barn were closed. The tool house, granary and garage were located south of the southwest corner of the barn. There was also a double corn crib east and northeast of the cow yard. The pump house was located a few feet south and east of the southeast corner of the cow yard. The shortest distance from the barn to the right of way was 650 feet. There was a small quantity of hay in the barn immediately south of which was a load of straw. All of the animals referred to were destroyed by the fire.
It is claimed by appellee that the fire originated from a cinder or other burning substance emitted from the smoke stack of one of appellant’s locomotives which went south immediately preceding the discovery thereof. When first observed, the fire was confined to the top of the load of straw.
At the conclusion of appellee’s testimony, appellant moved for a directed verdict upon the grounds that the evidence was insufficient to charge it with liability for the fire. The motion was renewed at the close of all of the evidence and again overruled.
The direction from which the wind was blowing is an important factor in determining the origin of the fire. Friel, who assigned all of his claims for damages to appellee, testified that he saw a freight train of appellant’s going south past his premises just as he entered his dwelling house. He observed that the engine was working hard. Immediately after going into the house, and within three minutes, he discovered that the load of straw on the south side of the barn was on fire. Appellee was present in the house. Friel and appellee immediately went to the barn, which was then observed to be on fire, for the purpose of releasing the stock confined therein.' Appellee testified that he ran at once to the northwest corner of the barn for the pur*152pose of releasing the stallion and jack from their stalls; that when he arrived at that point, which was a trifle over eighty feet from the residence, the flames swept over the building to such an extent that he was unable to open the door to the stall. Friel went through the gate leading to the barn to the southeast corner of the building, but the heat was so intense at that time that he was unable to open the door to the horse stalls, or the gate to the cow yard. The west and northwest portion of the barn was the first to fall. The barn was old and the weather dry and hot. The velocity of the wind was very great, possibly forty-five miles per hour, and the jury may well have found from the testimony that it was blowing from the southeast. The witnesses, however, vary somewhat in their understanding as to the direction of the wind and there was considerable testimony to the effect that it was variable and some of the time blowing from the south. The fire was communicated to the residence, but quickly exterminated. Witnesses other than Friel testified that the locomotive, which was passing over a slight grade, was working hard and throwing out great volumes of black smoke. One witness, Van Gundy, was working at a windmill on the Melvin Howard farm, something over eighty rods north of the Friel place, when the train passed. He testified that the locomotive was chugging hard and that he noticed coal was being shoveled into the engine as it passed; that the smoke came across to the windmill and that the locomotive was emitting substances nearly as large as a walnut.
Upon the discovery that the barn on the Friel place was burning, Van Gundy and the others present went immediately to the scene of the fire. Shortly after they left, Mrs. Howard discovered a fire in the bottom of an old straw stack near the windmill where the men had been working.
The distance a cinder or other burning substance would have to be carried by the wind to set fire to the load of straw could not have been less than 650 fee.t and probably a considerably greater distance. The bottom of the straw stack situated on the, Howard premises was around 600 feet from the right of way.-- No direct evidence as to the condition of the locomotive was introduced by appellee. The testimony on behalf of appellant was to the effect that it was equipped with a MudgeSlater spark arrester and that it was in good condition. The *153fuel used was mine-run Iowa coal which comprised a large percentage of slack.
Appellee undertook to prove that the locomotive was defective in some particular by a somewhat indirect method. Evidence was introduced to show that the locomotive was coaled at Grand Junction when it went north and again when it returned going south. The distance traveled was about eighteen miles. The tender held nine or ten tons of coal. The evidence also tended to show that the amount of coal consumed on the trip from Grand Junction north was unusual. Witnesses for appellee were permitted to testify that the unusual consumption of coal indicated that defects existed in the engine from which the emission of cinders and other substances might reasonably be expected.
The burden was on appellee to make out a prima facie ease of negligence against the appellant. Proof that the fire was set by a burning ember from one of appellant’s engines was sufficient for this purpose. Stewart v. Iowa Cent. R. Co., 136 Iowa 182; Neal v. Chicago, R. I. & P. R. Co., 129 Iowa 5; Hemmi v. Chicago G. W. R. Co., 102 Iowa 25. There was no eyewitness to the setting of the fire. All of the evidence tending to fix liability upon the appellant.-at this point is circumstantial in character. The rule, well settled in this state, is that:
“A theory cannot be said to be established by circumstantial evidence, even in a civil action, unless the facts relied upon are of such a nature, and are so related to each other, that it is the only conclusion that can fairly or reasonably be drawn from them.” Asbach v. Chicago, B. & Q. R. Co., 74 Iowa 248.
See, also, Neal v. Chicago, R. I. & P. R. Co., supra; Dingmon v. Chicago & N. W. R. Co., 194 Iowa 721.
It is the contention of appellant that the evidence wholly fails to establish such circumstances so related to each other as to make out a prima facie case of negligence. First, it is argued by appellant that sparks and burning embers emitted by a locomotive could not possibly be carried even the shortest distance shown in the evidence of sufficient vitality to set a fire; second, that the circumstances disclose that the fire originated in the bam and not on top of the load of straw; third, that it is as *154probable that the spark, if any, which set. the fire was thrown off by a gasoline engine in operation at the pump house.
The testimony on behalf of appellant touching the condition of the locomotive was that it was equipped with a Mudge-Slater Spark arrester which was in good condition and that no substance of any substantial size could possibly have been thrown out of the smoke stack. The evidence as to the construction of the spark-arresting device goes fully into detail, and, if it were in perfect condition, it is difficult to see how a burning ember could have been thrown out by the engine and carried by the wind the necessary distance to set the fire.
It was permissible for appellee to prove that the engine in question emitted sparks and burning embers which set fire to other substances immediately preceding the fire in question. Lanning v. Chicago, B. & Q. R. Co., 68 Iowa 502; Slossen v. B., C. R. & N. R. Co., 60 Iowa 215; International Harvester Co. v. Chicago, M. & St. P. R. Co., 186 Iowa 86. The evidence of the witness Van Gundy on this point was clearly admissible. Appellee was not precluded by the testimony introduced by appellant that the spark arrester was one in general use and one of the best devices known. Appellant’s witnesses who testified in chief that it would be impossible for an engine equipped with a spark arrester of the type referred to to throw off a cinder or other burning ember of sufficient vitality to set a fire 650 feet from the right of way admitted on cross-examination that, if fires were set by embers thrown off by the locomotive in the manner claimed by appellee in this case, the device must have been defective. One witness testified that, in his opinion, if chunks varying in size from a hickory nut to a walnut were thrown out of the smoke stack, there must have been large holes in the front netting.
No other cause for the fire is shown unless it originated from a spark thrown off by the gasoline engine. The gasoline engine was more than ninety feet directly east of the load of straw. Assuming that the wind was blowing from the south, as claimed by appellee and as some of the evidence tends to show, it is difficult to perceive how a spark thrown off by the gasoline engine could have been carried thereby to the load of straw. Practically the same physical fact would exist if the wind were blowing from the southeast. In such case, the spark would most probably have *155been carried to the northwest a safe distance east of the load of straw. Of course, there is testimony that the wind shifted at times and the currents of air about the barn might possibly have been so variable as to carry a spark to the load of straw. The probability that this could have occurred is very remote.
Furthermore, subsequent tests made of the gasoline engine would seem almost conclusively to demonstrate that it could not have been the source of the fire. All of the witnesses who testified on this point either stated affirmatively or admitted that the wind on the day in question had a very great velocity. The jury may well have found that it came from the southeast. Likewise, the evidence is clearly sufficient to sustain a finding by the jury that the fire originated on the top of the load of straw and that it was communicated therefrom to the barn.
The contention of appellant that the fire originated inside the barn is bottomed upon the testimony of the two witnesses for appellee that immediately after the discovery of the fire on the top of the load of straw they went to the barn but were unable to release the stock therein confined because of the extreme heat at the northwest and southeast corners of the barn. It must be confessed that, if appellee’s theory is correct, the fire spread with great rapidity after the load of straw caught. Appellee traveled less than ninety feet from the house to the northwest corner of the barn which he found on fire and the heat so intense he could not reach the door. As stated, the evidence shows that the day -was hot and dry, the barn old, and the wind blowing at a high velocity. In such circumstances, the flames would naturally spread with great rapidity and power.
It seems to us that, in the absence of any other probable cause of the fire, when all of the facts and circumstances are considered, the jury could properly infer that the fire was caused by a burning ember thrown out of the smoke stack of the locomotive and carried by the wind to the premises. Babcock v. Chicago & N. W. R. Co., 62 Iowa 593, 594. No one was about the barn and no one had been on top of the load of straw. Of course, the mere discovery of the fire immediately after the passing of the locomotive is not alone sufficient. General Ins. Co. v. Northern Pac. R. Co., 280 U. S. 72 (74 L. Ed. 172). The evidence, as we have shown, goes further than that in this case.
II. Appellee was permitted to introduce evidence over *156the objections of appellant of other fires set by locomotives in the vicinity of the Friel premises on other occasions prior and subsequent to the fire in question. The locomotive from which the spark or burning ember was emitted, if any-, and which caused the fire, was fully identified by the evidence in this case, and therefore, evidence of fires set by other engines was inadmissible for the purpose of proving negligence. Bell v. Chicago, B. & Q. R. Co., 64 Iowa 321; Babcock v. C. N. W. R. Co., supra; International Harvester Co. v. Chicago, M. & St. P. R. Co., supra. This rule does not, however, exclude testimony of other fires for all purposes. It was admissible in a proper case, for instance, for the purpose of showing the accumulation of inflammable material upon the company’s right of way. International Harvester Co. v. Chicago, M. & St. P. R. Co., supra. Evidence of this character is also admissible for the purpose of showing material physical facts, such as the possible distance burning substances thrown off by a locomotive may be carried and a fire set thereby. Russell v. Chicago, M. & St. P. R. Co., 195 Iowa 993; Dumbarton Realty Co. v. Everist (Iowa), 174 N. W. 269 (not officially reported); Budd v. Ann Arbor R. Co., 200 Mich. 250 (166 N. W. 927); 3 Elliott on Railroads (3d Ed.), Section 1768.
The testimony to which objection was interposed tended to show that sparks or burning substances were thrown by engines of appellant a distance of forty rods. The court in its instructions to the jury limited this testimony strictly to proof of the physical facts, and the jury was not permitted to consider it on the issue of negligence.
III. At least three of the witnesses for appellee, prior to the commencement of this action, signed written statements prepared by and at the request of one Clifford, the claim agent of appellant. Appellee refused to sign a similar statement. The evidence on the part of Clifford on this point was that he prepared the statement from information given him by appellee; that he gave him the original copy to read over and to sign and return if the contents were found satisfactory. The statement was neither signed nor returned. Counsel for appellee on cross-examination interrogated Clifford at considerable length as to this statement. The purport of the examination was that Clifford had prepared a statement which did not represent the facts *157and that for this reason appellee had refused to sign it. Objections were interposed upon the ground that it was not cross-examination. The objections were overruled. Clifford had in his possession a typewritten document which he testified was a carbon duplicate of the original. This duplicate was offered by appellant but excluded by the court. The statement is incorporated in the abstract. Appellant should have been permitted to rebut any imputation or inference which the jury might find or draw from the cross-examination. While the statement was not, as stated, signed, it was quite conclusive as to its nature and contents. Appellant was entitled to the whole conversation and all that was said on the subject. We think the statement should have been received in evidence and we cannot say that its exclusion was without prejudice. The citation of authorities is not necessary, but see Glassman v. Chicago, R. I. & P. R. Co., 166 Iowa 254; State v. Hudson, 110 Iowa 663, 666; State v. Rutledge, 135 Iowa 581, 587.
IV. Van Gundy, also a witness for appellee, signed a statement prepared by Clifford, purporting to give his understanding of certain material phases of the case. He, with others, was working at a windmill on a farm north of the Friel farm when the freight train went by. The statement signed by him recited that the wind was blowing very hard from directly south, that the engine was emitting large volumes of smoke which did not rise very high above the train, and that he saw no sparks or burning substance coming from the engine. In his testimony upon the trial, he testified that he saw burning embers nearly as large as a walnut emitted from the smoke stack; that he saw a number of pieces of this character; and that the smoke from the engine was carried by the wind to the northwest. Counsel for appellant on cross-examination sought to interrogate the witness concerning the signed statement. Extracts from the statement were read to him and he was asked if the language to which his attention was called was contained therein at the time he signed it. Objections were sustained to this cross-examination, and of the rulings of the court at this point appellant complains.
We think it obvious that the cross-examination was improperly limited by the court’s rulings. The testimony of the witness as given on the trial was in direct conflict with some of the language contained in the statement and counsel for appellant *158had a right on cross-examination to go fully and thoroughly into the matters covered by the examination in chief and to test the credibility of the witness. The rules applicable at this point are elementary and we cannot say that no prejudice resulted from the court’s ruling.
V. Appellee introduced no direct evidence tending to show the condition of the engine in question. Appellant offered evidence to the effect that the engine was equipped with MudgeSlater spark arrester, a device much used, and that it was in good condition. Appellee did introduce testimony tending to show that appellant’s engine was coaled at Grand Junction going north and again when returning. This, it is claimed, was unusual. The tender held about nine tons and the distance the train went after leaving Grand Junction until its return was but a few miles. The record does not disclose either the amount of coal in the tender when the train arrived at Grand Junction going north or how much coal was put in at either coaling. Basing his answers upon this testimony, the witness Friel was permitted to express the opinion that the front end of the engine was burned out; that the screens were in bad condition or that the flues were dirty. This testimony was objected to on the ground that no proper foundation had been laid for the hypothetical question. It is elementary that hypothetical questions must be based upon and limited to evidence introduced upon the trial. Bennett v. City of Marion, 119 Iowa 473; Ney v. Eastern Iowa Tel. Co., 162 Iowa 525; Spiers v. Hendershott, 142 Iowa 446. It is equally well settled that the question as to whether a proper foundation has been laid for a particular hypothetical question is in a large degree a matter of discretion on the part of the court. State v. Donovan, 128 Iowa 44; Lanning v. Chicago, B. & Q. R. Co., 68 Iowa 502. The two rules do not conflict. If there is no evidence upon which the hypothetical question may be based, the proponent is not aided by the discretionary power of the court. It seems to us that no proper foundation was laid for this testimony. It is not based upon testimony in itself tending more than remotely, if at all, to show the condition of the engine and was really based upon a mere possible inference to be drawn from the single fact. The inference sought to be drawn was doubtful and precarious. It opened up a large field for the speculation of the jury. No such facts were shown as to *159constitute a proper foundation for expert testimony. The objections should have been sustained.
VI. Prejudicial misconduct on the part of the court and also of counsel for appellee is charged. In view of a reversal upon other grounds, we deem it unnecessary to consider the assignments at this point. Other questions are discussed by counsel. They are not likely to arise upon a retrial of the case. What we have said disposes of the principal propositions relied upon for reversal. For the reasons pointed out, the judgment of the court below is reversed. — Reversed.
All the justices concur.