Jones recovered a judgment for personal injuries received by him while he was working as a brakeman employed in interstate commerce on the defendant railroad. The fault alleged was the employment of a coupling device which could not be operated in uncoupling without the necessity of going between the cars, and recovery thus depended upon the Safety Appliance Act (section 2, Act of March 2, 1893; section 8606, C. S.).
The first question involved is whether there was any evidence to support the jury’s conclusion that the coupler was of the forbidden character. In a switching operation in the yards, a cut of cars was being pushed by the engine. The two leading cars farthest from the engine were to be, at the right moment and while still in motion, uncoupled, so that they might be kicked over a switch onto another track. Jones was riding at the right-hand forward corner of the third car. He had his left foot in the stirrup, which is on the side of the car at the corner. He was facing in, leaning forward, and holding onto the brake wheel with his left hand. The brake wheel staff was about two feet from the corner of the car. In his right hand he held a lantern, and his right foot rested upon the outer end of the lever, which was designed by its depression to lift the coupling pin upon the inner and short arm of the *527same lever. The operating motion would be a direct downward pressure. The apparatus was unquestionably, speaking in a general way and without reference to specific defects, of a standard and approved type. At what he judged the proper moment, Jones pushed down with his right foot. The pin did not lift.- He tried again, moving his foot a little further in, and the second effort was unsuccessful, or perhaps his foot slipped off without any effective effort. He then decided that he would get a better result if he put his body and his weight more directly over the lever, and accordingly he swung his right arm in, grasped the wheel with both hands, and then, excepting for his left foot, which remained in the stirrup and his left leg reaching to it, swung his body around between the ends of the cars and opposite the brake staff, and put his whole weight upon the lever near that point. This time the pin lifted and the cars uncoupled, but the inward swinging of his lantern, as he reached for the brake wheel, had been mistaken for a stop signal, the car he was upon suddenly slowed, and, as some of the evidence tended to show, he was' thrown down upon the track and thus hurt.
[1] The question chiefly argued has been whether the mere fact that the device did not operate upon the first or 'perhaps the second trial was any substantial evidence that there was a statutory defect, or was inconsistent with the thought of the phrase “can be operated.” It is pointed out that the statement in Chicago Co. v. Brown (an uncoupling case) 229 U. S. 317, 321, 322, 33 Sup. Ct. 840, 841 (57 L. Ed. 1204), “the failure of a coupler to work at any time, sustains a charge of negligence in this respect,” was merely a recital of counsel’s concession, and in a later coupling case, San Antonio Co. v. Wagner, 241 U. S. 476, 484, 36 Sup. Ct. 626, 60 L. Ed. 1110, the very point was carefully reserved. In Atlantic Co. v. Parker, 242 U. S. 56, 37 Sup. Ct. 69, 61 L. Ed. 150, the only holding is that the jury may infer a statutory defect if there is evidence that the required automatic coupling did not occur because there was too much lateral play in the drawheads, and the fact that some lateral play is necessary in order to permit coupling on a curve did not necessarily show that there was not an unnecessary amount in these drawheads. These cases are closest in point, but they fall short of any definite holding; and, in addition, it must be noted that the problem and the inference from nonaction are not the same in a matter of uncoupling as in a matter of coupling. Apparently cou'pling by impact will always occur automatically, unless there is some defect or some abnormal attendant condition. Hence a failure to couple usually tends to show a coupler defect. As to the uncoupling devices, it seems to be conceded by counsel, and we understand it to be true, that, however perfect they may be, they will not work at all while there is a binding friction upon the lifting pin, such as undoubtedly exists while there is a pulling strain. Hence a failure of the pin to lift may indicate that the device is defective, or may indicate that the effort was made at the wrong time. We are convinced that evidence that the lifting pin did not work upon the first effort may be sufficient evidence of an existing defect, if it also appears that the effort was of a character and made at a time when it would have naturally been ef*528fective save for a defect; but whether the present record is sufficient in this 'particular we need not decide, as the case must be tried again, and upon the new trial this point will be more carefully developed.1
[2] For another reason, the plaintiff was entitled to go to the jury. This outer and longer lever arm, instead of being horizontal, was bent down so that its outer end was considerably below its pivot in the car sill. This obviously made its operation more difficult by a brakeman standing in the stirrup, and yet operation by a brakeman in this position seems to have been contemplated. There is fair room for inference that this bent condition of the lever was a cause contributing to Jones’ injury, and it is not clear from the record that this bend was not a defect, either in original construction or from later distortion. True, the proof to this effect is rather vague and unsatisfactory; but, in view of Jones’ entire testimony, he cannot be taken as conclusively conceding that in this precise particular the apparatus was of standard and satisfactory form. It may well be, as defendant’s counsel say in the argument, that the lever was thus bent down so that it could be operated either by a brakeman on the car or by a man on the ground, and that the bend was not enough, to interfere with easy and efficient working from the car. It may be, also, that if defendant had shown this, and that this precise construction was approved, and this had not been disputed, except by Jones’ testimony in its present form, it would have cleared up any ambiguity and left no room for doubt; but the defendant took no evidence on this point, and rested its case on the supposed insufficiency of 'plaintiff’s proof. We think it misjudged the force of inferences which may properly be drawn from such proof, and that there was in this respect a case for the jury.
[3] Defendant assigns error because the trial judge refused to give instructions defining the statutory phrase “going between the ends of the cars,” and thus left it open for the jury to say that even the putting in of an arm or leg to the point within the corner of the car, where the outer end of the operating lever is compelled to be, would indicate a violation of the statute. We have no doubt an instruction should have been given to the effect that the statute did not reach that slight degree of “going between the ends of the cars”; but no harm was done by not giving it. The evidence was undisputed that Jones did 'put the greater part of his body between the ends of the cars within the rail line, and the verdict of the jury, under the charge as given, must be based upon the theory that this complete movement by Jones underlay the injury.
[4] By the fourth request to charge, the court was asked to tell the jury that, if Jones’ inability to make the cutting lever work the first time he tried was because of the friction between the coupler and the pin or the position of these two cars in relation to each other, and not because of any defect in the coupling apparatus, such failure to operate from that cause would not tend to show violation of the Safety Appli*529anee Act. In thus stating it, we somewhat paraphrase the request; but we think this meaning is clear, and that it is not essentially modified by some additional matters to which it refers. This request was refused, an exception taken, and nothing equivalent or of an approximate character was said in the general charge. In the opinion of a majority of this court, this was error, and of vital importance. The case for the plaintiff, on the theory that the lever was defective because of its bend, was at the best barely strong enough to justify submission. It is at least equally probable that the verdict for the plaintiff was based on the theory that the fact of inoperativeness at the first effort sufficiently indicated some defect in the Apparatus, and this, though the failure to operate was in fact due to the effort to make it work when the conditions of slack were wrong, and when it could not have worked, even if in no way defective.
The only reason .which suggests itself for not giving this fourth request is that there was no evidence that the failure to work was because of the condition of the slack. We cannot accept the suggestion. We can be no more certain of that than we can be that the bent condition of the lever was not a defect. Plaintiff’s own testimony tends to show that, during the pushing operation, as it was here conducted and before the engine began to slow, there would be a binding friction on the coupler pin; even if there would not be, there could be no certainty that, before this first pull, the slack between these two cars had not run out so as to make a pulling strain upon the pin. Perhaps neither of these things is*true; but we cannot assume something not conclusively shown by the record, in order to conclude that there was no prejudice from an erroneous instruction.
The judgment is reversed, and a new trial ordered.