SANBORN, Circuit Judge,
after stating the case as above, delivered the opinion of the court.
After several witnesses had testified that the roadbed and the railroad track of the plaintiff in error from the top of the hill east of Sycamore to the water tank in that city, where the collision occurred, were in a poor condition; that some of the ties were rotten, and that many of the spikes were loose or lost; and after Charles A. Field had testified that he was a locomotive engineer, that he had been in the employment of the plaintiff in error for 11 years, that he was running a suburban train from Chicago to Sycamore at and before the time of the accident, that he knew the condition of the road from the top of the hill to the water tank, that it was in bad condition, and would sway a train as it passed over it, — -the trial court permitted him to testify, over the objection of the company, that the rough and uneven condition of the track was liable [426]*426to throw a pin out, and thus to part a train, and this ruling is as signed as error. This declaration was not the statement of any fact, but the communication of the opinion of the witness as to the effect of the roughness of the road. The rule undoubtedly is that a witness must state facts, and not opinions, but there is an exception which arises out of the necessity of the case that is as firmly established and as well known as the rule. It is that the opinions of witnesses possessing peculiar skill or knowledge of the subject-matter may be received in evidence whenever the facts are such that inexperienced persons are likely to prove incapable of forming a correct judgment without such assistance. The trial court was of the opinion that the testimony of this witness fell under the exception to the rule, and we are not convinced that there was any error in this view. The line of demarcation between competent and incompetent expert testimony is not always clear and definite, and judgments ought not to be reversed on account of the reception or rejection of such testimony unless there was a clear violation of the rule. It is not probable that the farmers, mechanics, and business men who composed the jury in this case were as capable of forming a judgment upon the effect of a rough railroad upon the links and pins with which the cars of a freight train are fastened together as a locomotive engineer who had been operating a railroad train for years. Motey v. Granite Co., 36 U. S. App. 682, 689, 20 C. C. A. 366, 370, 371, and 74 Fed. 155, 159; Railway Co. v. Edwards, 49 U. S. App. 52, 56, 24 C. C. A. 300, 302, and 78 Fed. 745, 747; Fireman’s Ins. Co. v. J. H. Mohlman Co., 62 U. S. App. 287, 291, 33 C. C. A. 347, 349, and 91 Fed. 85, 87; Clifford v. Richardson, 18 Vt. 620, 627. There were like objections to similar testimony of other witnesses, but for the reasons stated above we have reached the conclusion that they were properly overruled, and that the evidence of these witnesses was properly received.
The theory of the defendant in error at the trial was that the fire was communicated to the flowing liquid, and that the explosion was caused, by fire from the dining car which stood on the spur track about eight feet south from the tank from which the inflammable fluid escaped. The theory of the plaintiff in error was that the fire was set to the liquid by Price’s lighted lantern, and that it was his negligence in approaching the fluid with this light in his lantern that caused the fatal result. In order to prove that the theory of the defendant in error was unfounded, the railroad company introduced evidence to the effect that no fire and no coals were taken or escaped from the dining car. The plaintiff in error complains that the court subsequently permitted the administratrix to prove by one Hibbard, who was a locomotive engineer, that he had pulled tanks of gasoline with his engine, that brakemen with lighted lanterns had inspected leaking tanks, and that he never knew of a case in which the light from a brakeman’s lantern caused the liquid to take fire. The objection to this testimony was that it was not proper testimony in rebuttal. But the record shows that the testimony of another witness to the same effect had already been received in rebuttal, without challenge, before this objection was presented, and [427]*427that this testimony was not contradicted; so that no prejudice could hare resulted from the cumulative evidence of the subsequent witness. If the admission of his testimony was error, it was error without prejudice, and no ground for reversal.
It is assigned as error that witnesses were permitted to testify in rebuttal that they saw a danger signal — a red flag — between Sycamore and the top of the hill a month or more before the accident occurred. But the entire question of the condition of the railroad at this place, of the repairs that had been made upon it during many months prior to the accident, and of the daily work of the section men upon it was before the jury, and the testimony upon every phase of it was in conflict. This testimony would have been competent if it had been introduced in chief, and in this state of the case a trial court has much discretion in the admission of rebutting testimony. It was guilty of no such abuse of this discretion here as would warrant a reversal of this judgment.
There are other assignments of error regarding the admission and rejection of evidence. They have all been carefully examined,, and found to be untenable. They are either disposed of by the views which we have already expressed, or they are of insufficient importance to warrant their statement and discussion.
The chief reliance of counsel for the plaintiff in error is not upon their objections to the testimony. It is upon their contention that the court below should have instructed the jury to return a verdict in favor of the railway company. They insist that there were many questions presented by the evidence and submitted to the jury which it was error for the court to refuse to decide, and that, if it had decided any one of them, the logical and unavoidable result would have been a peremptory instruction in favor of the company. The assignments of error which refer to this matter are numerous and voluminous. They assail various portions of the charge of the court, its refusal to grant numerous requests for instructions, and its failure to peremptorily instruct the jury in favor of the company. But, when they are carefully analyzed, they all come to this: that for one reason or another the court erred because it did not direct a verdict for the railway company. Before entering upon a discussion of the questions which these assignments present, it is well to call to mind the established rules by which they must be determined. It is conceded that at the close of the evidence there is always a preliminary question for the judge before the case can he properly submitted to the jury, and that is whether or not there is any substantial evidence upon which the jury can properly render a verdict in favor of the party who produces it, and that, if there is no such evidence, it is the duty of the court to direct the jury to return a verdict against him. Commissioners v. Clark, 94 U. S. 278, 284; North Pennsylvania R. Co. v. Commercial Nat. Bank, 123 U. S. 727, 733, 8 Sup. Ct. 266; Railroad Co. v. Converse, 139 U. S. 469, 11 Sup. Ct. 569; Laclede Fire-Brick Mfg. Co. v. Hartford Steam-Boiler Inspection & Insurance Co., 19 U. S. App. 510, 515, 9 C. C. A. 1, 4, and 60 Fed. 351, 354; Gowen v. Harley, 12 U. S. App. 574, 585, 6 C. C. A. 190, and 56 Fed. 973; Motey v. Granite Co., 36 U. S. App. 682, 686, 20 [428]*428C. C. A. 366, 368, and 74 Fed. 155, 157; Railway Co. v. Belliwith, 55 U. S. App. 113, 121, 28 C. C. A. 358, 362, and 83 Fed. 437, 441. But it is equally well settled that it is only when the evidence leaves the material facts admitted or undisputed, and only when these facts are such that reasonable men, in the exercise of an honest and impartial judgment, can fairly draw but one conclusion from them, that the court may properly withdraw the case from the jury. If the evidence relative to the material facts is contradictory, or if, from the admitted or established facts, the unprejudiced minds of reasonable men may well draw different conclusions, it is the duty of the court to submit the issues to the jury. Railway Co. v. Jarvi, 10 U. S. App. 439, 451, 3 C. C. A. 433, 438, and 53 Fed. 65, 70; Fuel Co. v. Danielson, 12 U. S. App. 688, 696, 6 C. C. A. 636, 640, and 57 Fed. 915, 920; Drake v. Stewart, 40 U. S. App. 173, 178, 22 C. C. A. 104, 107, and 76 Fed. 140, 143; Railway Co. v. Ives, 144 U. S. 408, 417, 12 Sup. Ct. 679; Railroad Co. v. Converse, 139 U. S. 469, 11 Sup. Ct. 569; Railroad Co. v. Pollard, 22 Wall. 341. Bearing these familiar rules in mind, let us consider the claims of the plaintiff in error.
It is claimed that the court was wrong in submitting to the jury the questions whether or not the railroad track was in bad condition where the train parted, between the top of the hill east of Sycamore and the water tank in that city, whether or not the poor condition of the track caused the train to part, and whether or not the negligence of the company in the care of this track was the proximate cause of the fire, the explosion, and the death of Price. The charge of error in submitting the first two questions is refuted by the fact that five witnesses testified that in portions of this track many of the ties were rotten, and that many of the spikes which should have held the rails to them were loose, or entirely out, sb that, as some of them said, the rails would sink into the ties and roadbed as trains passed over them; and four witnesses testified that the roughness and unevenness of this track, produced by the rotten ties and loose rails, had a tendency either to throw pins or break couplings as heavy freight trains passed over it. All this testimony was contradicted, it is true, but, after all, it was ample to raise a dispute over these issues, and to make it the imperative duty of the court to submit them to the jury. The third question —the question whether or not the roughness of the road was the proximate cause of the fire, the explosion, and the death — must be considered in the light of the finding which the jury must have made that the company was negligent in the care of this portion of its railroad. The proximate cause of an injury is the primary, moving cause, without which it would not have been inflicted, but which, in the natural and probable sequence of events, and' without the interposition of any new or independent cause, produces the injury. The nature of this cause, its relation to its effect, and the difficulty and perplexity which often condition its discovery, have been so frequently and exhaustively considered and illustrated by this court that a mere reference to some of the opinions which treat of these subjects will amply elucidate the brief definition we have [429]*429given, and will render its further consideration here both unnecessary and unprofitable. Railway Co. v. Elliott, 12 U. S. App. 381, 386, 5 C. C. A. 347, 349, and 55 Fed. 949; Insurance Co. v. Melick, 27 U. S. App. 547, 552, 553, 12 C. C. A. 544, 547. and 65 Fed. 178, 181; Railway Co. v. Callaghan, 12 U. S. App. 541, 547, 6 C. C. A. 205, 208, and 56 Fed. 988, 991. The question here is not whether or not, in our opinion, the roughness of the track was the proximate cause of the explosion and the death, but whether or not all reasonable men of unprejudiced minds would draw that conclusion from the facts of this case; for, if they would not, the question was properly submitted to the jury. In Railway Co. v. Kellogg, 94 U. S. 469, 474, 476, Mr. Justice Strong, who delivered the opinion of the court, said:
“The true rule is that wliat is the proximate cause of an injury is ordinarily a question for the jury. It is not a question of science or of legal knowledge. * * * In the nature of tilings, iliere is in every transaction a. succession of events more or less dependent upon those preceding-, and it is the province of a jury to look at this succession of events or facts, and ascertain whether they are naturally and probably connected with each other by a continuous sequence, or are dissevered by new and independent agencies: and this must be determined in view of the circumstances existing at the time.”
To the same effect are Insurance Co. v. Melick and Railway Co. v. Callaghan, supra. Now, it is common knowledge that the burning of ears, the ignition and explosion of inflammable liquids carried upon them, and the death of men who ride in or on them are sometimes caused by the collision of engines or of the ears which compose the trains which they draw. In the case at bar Hie railroad was found by the jury to have been rough and uneven. There vas evidence that this roughness tended to break a coupling or throw a pin in a heavy train as if passed over it. Such a train parted as it was running over this rough track. The separated sections collided, and tore open a tank of inflammable fluid which the train carried, and the liquid flowed out. There was a rule of the company which required the rear brakeman to be on the top of his train as it came down the hill to the place of this collision, and another rule which made it the duty of the conductor to enforce the former rule. The rear brakeman testified once that he was in his place on the top of the train, and he afterwards testified that he was not there, and that these rules were violated by the conductor’s direction. There was testimony that the conductor was warned of the danger of approaching the flowing liquid with his lantern after the collision, and that the witness who gave this testimony uad stated that lie gave him no such warning. It was night. Engines carry fire to draw their loads, and trainmen carry lights in their lanterns to see their trains. The conductor walked along the side of Hie train with his lighted lantern in his hand, as is customary in the night, for the purpose of discharging his duty of watching and attending to the property in his charge. The inflammable liquid which had spread over the ground took fire, exploded, and killed him. There was testimony that fire could not have been communicated to it from the dining car. There was testimony that it could not have been communicated from the lantern. [430]*430In this state of the case, can any one say that all reasonable men would draw the conclusion that the roughness of the railroad was not the primary moving cause which produced this fatal result; that the final explosion and death would have occurred if the railroad had been sound, solid and smooth; or that the fire, explosion, and death were produced by some new and independent cause, which was not itself brought about by the roughness of the road, but which turned aside the natural sequence of events, and produced the terrible disaster? A studious and careful examination of this question in the light of the evidence in this record has forced us to the conclusion, whatever might have been our opinion upon the original issue of cause and effect, that it cannot be truthfully answered in the affirmative. The proximity of the uneven road, the rotten ties, and loose rails, and the negligence which permitted their existence, to the explosion and the death, is too close, their probable connection in the relation of cause and effect is too great, and there is too much doubt whether there was any act or negligence of the conductor which affected the natural sequence of events between them, and constituted a new and independent cause of the fatal result, to warrant any other conclusion. The question whether or not the fiegligence of the company was the proximate cause of the death of Price was rightfully submitted to the jury.
It is said that this is a transitory action, that its cause arose in the state of Illinois, that the rule in that state is that the complainant must allege and prove that the party injured was without fault on his part, and that the court below was in error because it did not charge the jury in this case to return a verdict for the company on the ground that the defendant in error did not plead or prove that the deceased was not guilty of negligence which contributed to his death. But the rule of the national courts is settled and uniform that contributory negligence is an affirmative defense, which must be established by a preponderance of evidence. It may appear from the testimony introduced by the plaintiff, or from that presented by the defendant, but, in the absence of all evidence on the subject, it is no fault or defect of the plaintiff’s case that he fails to plead or prove that the defense of contributory negligence does not exist. In the jurisprudence of the national courts he is not called upon to establish the negative. Railroad Co. v. Gladmon, 15 Wall. 401, 406; Railroad Co. v. Horst, 93 U. S. 291, 299; Eddy v. Wallace, 4 U. S. App. 264, 273, 1 C. C. A. 435, 440, and 49 Fed. 801, 804.
Another claim is that, if the ties were rotten, if the spikes were loose or lost, and if the railroad was rough and uneven, and therefore dangerous, Price must have known it, and must have assumed the risk and danger of its condition, and that the court erred because it did not so instruct the jury, and because it left the question of his assumption of the risk for their determination. It was the duty of the railroad company to exercise ordinary care to furnish a reasonably safe railroad, and to use ordinary care to inspect and to keep it in a reasonably safe condition for the operation of [431]*431trains upon it. When the conductor, Price, entered upon or continued in the employment of the company lie assumed all the risks and dangers of Ms occupation which were known to him, and all which a reasonably prudent man in Ms situation would have known. But the degrees of care on account of a defective railroad which che law required of the railway company and of the conductor were dilieiont. It was the duty of the former to inspect and care for it. It was the duty of the latter to operate his train over it carefully. In the conduct of his train, and in the absence of knowledge, or of reasonable means of knowledge or notice, to the contrary, he had the right to assume that the company had discharged, and would continue to discharge, its duty in the care of the track. Moreover, the opportunities of the two parties to know the decayed condition of the ties and the looseness of the rails were not the same. The company, through its section men, passed slowly over these ties and rails daily for the express purpose of examining and repairing this railroad. The conductor rode rapidly over the road, with his mind and care directed to the operation of his train. His lips are closed. What he knew and what he ought to have known of the defects of this railroad must be interred from the facts and circumstances which are established by the testimony of others. Again, there was evidence tending to show that repairs were being made upon the road just east of Sycamore shortly before the accident, and, even if the conductor knew that the road was rough and dangerous, he may have assumed that it was being repaired, and that it would be in a safe condition before he passed over it again. There were other conductors operating trains over this defective portion of the railroad, and they continued in their employment. The rule is that the danger from the defects of the railroad or machinery furnished the employé must have been so obvious and threatening that a reasonably prudent man in his situation would have avoided them, in order to charge the injured servant with contributory negligence because he continued in the discharge of his duty, and thereby assumed the risks. Railway Co. v. Jarvi, 10 U. S. App. 439, 450, 3 C. C. A. 433, 437, and 53 Fed. 65, 69; Kane v. Railway Co., 128 U. S. 91, 94, 9 Sup. Ct. 16; Railroad Co. v. McDade, 135 U. S. 554, 570, 573, 10 Sup. Ct. 1044; Cook v. Railway Co., 34 Minn. 45, 24 N. W. 311; Myers v. Iron Co., 150 Mass. 125, 22 N. E. 631. In the light of the legal principles to which we have referred, the evidence in this case left the extent of the knowledge of the decayed ties, the loose rails, and the defective road, and of the risks and dangers therefrom, with which Price was chargeable under the law, in a state too uncertain and indefinite to warrant the trial court in withdrawing the question of his assumption of these risks from the jury.
There was a rule of the company which required the brakemen to he on the top of their train as it came down the hill into Sycamore, and which directed the conductors to see that their brakemen complied with this rule. There were on this train an engineer, a head brakeman, a rear brakeman, and the conductor, Price. The rear brakeman testified that Price told him not to go out on the top of the train as it came down into Sycamore, and that he and [432]*432Price remained on the rear platform of the caboose until the collision occurred. On cross-examination he did not deny that he had testified before the coroner’s jury that he was on the top of the train about two miles east of Sycamore, and that he was out there all the time until he arrived at Sycamore; nor that he had also testified that he was on the platform of the caboose during this time. There was uncontradicted testimony that, if he had been in his proper place on the top of the train, and if he had discovered that the train had parted, he could have set the brakes on the rear section, and could have prevented the collision. None of the trainmen discovered the parting of the train until the collision occurred. The court instructed the jury that, if Price told the rear brakeman not to go on the top of the train, if he did not go, and if he would have been able to stop the rear section of the train, and prevent the collision, if he had been there, then they would be justified in finding negligence on the part of Price. He charged them that, if the accident was caused or aided by the failure of the rear brakeman to be on the top of the train, and if his failure to be there was due to the instructions of Price, those facts would be a defense to this action, and their verdict must be for the railway company. He added:
“If, however, the evidence fails to satisfy you that Mr. Price told Mr. Stewart [who was the rear brakeman] not to go on top of the train, even if Stewart did not go on the top of the train, if the fact was not due to the negligence of Price, but of Stewart acting without the direction or contrary to the direction of Price, that would be negligence of Mr. Stewart, and not chargeable to Mr. Price, as contributory to the accident.”
This portion of the charge is assigned as error, and it is insisted that the evidence here conclusively shows that Price was guilty of contributory negligence, and that the court should have instructed the jury to that effect. There are two reasons why we are unable to sustain this assignment. One is that the testimony of the rear brakeman as to his position was so contradictory, and the motive for him to excuse his own delinquency, by laying the fault on one whose lips were closed forever, so strong, that we are by no means convinced that all reasonable men would conclude from this evidence that Price told him not to obey the rules of the company. He testified at one time that he was on the top of the train all the way from a point two miles east of Sycamore until he reached the city, and at another time that he was on the rear platform of the caboose during all this time. Both of these statements were not true, and, if he testified falsely on one material issue, the jury and the court were permitted to disbelieve his testimony upon another. The credibility of such a witness, and the effect that should be given to his testimony, were questions peculiarly within the province of the jury. Another reason why the assignment cannot be sustained is that, if it were conceded that it was Price’s negligence which prevented the rear brakeman from occupying his place on the top of the train, it is so doubtful, under the evidence, whether or not he would have discovered the parting of the train before the collision, if be had been on the top of it, so that he would have stopped the rear section; in other words, it is so doubtful whether or not the neg[433]*433ligence of Price contributed to the accident and death that the court could not properly have taken that question from the jury. The evidence does not disclose where the head brakeman was as the train came down the hill into Sycamore; but bis place was on the top of the train, and on the forward end of the train, and the presumption is that he was in that place, and was there discharging his duty. The train contained 22 cars, and it parted 7 cars from the engine. Yet neither the head brakeman nor the engineer discovered the break until the collision occurred. In this state of the case it is at least as probable that the rear brakeman would not as that he would have discovered the break in season to have prevented the collision if he; had been on the top of the train. In other-words, if Price was guilty of negligence here, there is as much ground for reasonable men to conclude that his negligence did not as that it did contribute to the injury, and the question of his contributory negligence was, therefore, for the jury, and not for the court.
After the collision occurred, Price took his lantern, and went forward on the south side of the train, to learn- what had happened, and to care for the property in his charge. The engineer of the waterworks at Sycamore testified that, as the conductor was walking along the side of the train towards the escaping liquid, whose flowing was audible, be said to him, “You had better not take your lantern if you are going np there; you might set it on fire;” but that the conductor went on with his lighted lantern, and was killed. This witness had testified to the same facts before the coroner's jury. Two witnesses came, who said that between the time when this engineer gave his testimony before the coroner’s jury and the time of the trial he had told them that he was mistaken in his testimony before the coroner’s jury, and that he did not warn Price not to approach the flowing liquid with his lighted lantern. There were also two witnesses who testified to the effect that a fire in a lighted lantern such as the conductor carried would not ignite gasoline or a like inflammable fluid which liad leaked out of a tank upon the ground if the lantern was supplied with a globe, and was in good order. The court submitted the question to the jury whether or not Price received the warning, and whether or not, under all the circumstances of the case, he was guilty of contributory negligence in going forward with his lighted lantern to the place where the liquid had (lowed out over the ground. It. is insisted that this was error, and that the court should have instructed the jury that his act in approaching the flowing liquid with his lantern was negligence which contributed to bis death, and which entitled the railway company to a verdict. But when the collision occurred it was the duty of Price Lo go forward along his train, to find out what damage had been done, to prevent more injury, and to protect and preserve the contents of the tank, and the other property on his train. It was night. He could not see clearly without his lighted lantern. Whether or not he was warned of the danger of approaching the escaping fluid was a disputed question under the evidence. There was testimony that the gasoline could not have been ignited from the dining car, and that it could not [434]*434have been fired from the lantern, and yet it burned. The disputed questions of fact were for the jury. In this medley of contradictions and uncertainties, how could the court say that all reasonable men must conclude that this conductor, who walked forward by the side of the train with his lighted lantern in the usual way, in the discharge of his duty, exposed himself without ordinary care to dangers which a reasonably prudent man would have perceived, and that he thereby contributed to his own death? It could not rightfully say so, and it did not say so. There was too much doubt what facts reasonable men would find existed under this conflicting testimony, and what deductions they would draw from the facts which they did find, to allow a court to pursue any other course.
The discussion and decision of the questions which were presented in this case are now concluded. Their consideration has led to repeated readings of the evidence and of the charge of the court, and our conclusion is that the trial was full and fair, that there was no error in the admission or rejection of evidence, and that the charge of the court was not only a correct declaration of the principles of law involved, but a clear, terse, and logical presentation and application of them to the salient issues of the case. If, however, we had been led to doubt the correctness of any of those portions of the charge of the court which have been challenged, or of its refusal to direct a verdict for the defendant, or to give any of the refused requests for instructions presented to it on account of any of the reasons alleged by the counsel for the plaintiff in error, the judgment could not have been reversed upon the record before us. The correctness of all these rulings depends upon the evidence before the trial court. The record discloses the fact that only a portion of this evidence has been returned to us. Thát which is found in the bill of exceptions and in the record is in narrative form, and the certificate is limited to the statement that “the foregoing is substantially all the evidence.” A certificate that the substance of the evidence is returned is not sufficient to warrant the appellate court in reversing a judgment on the ground that the trial court refused to direct a verdict. Railroad Co. v. Washington, 4 U. S. App. 121, 131, 1 C. C. A. 286, 292, and 49 Fed. 347, 353; Railway Co. v. Harris, 27 U. S. App. 450, 457, 12 C. C. A. 598, 603, and 63 Fed. 800, 805; Taylor-Craig Corp. v. Hage, 32 U. S. App. 548, 16 C. C. A. 339, 340, and 69 Fed. 581, 582; Association v. Shryock, 36 U. S. App. 658, 664, 20 C. C. A. 3, 6, and 73 Fed. 774, 777. “The burden of proof to show that there was no evidence to warrant a charge is on him who asserts an error of that character; and, if he would maintain his claim, he must either present all the evidence to the appellate court, so that the reviewing court can see for itself what the evidence was, or he must present a bill of exceptions which has the certificate of the trial court that no evidence of the character in question was presented to it.” U. S. v. Patrick, 36 U. S. App. 645, 656, 20 C. C. A. 11, 17, and 73 Fed. 800, 806.
The judgment below is affirmed.