concurring:

The phrase “assumption of risk” is an excellent illustration of the extent to which uncritical use of words bedevils the law. A phrase begins life as a literary expression; its felicity leads to its lazy repetition; and repetition soon establishes it as a legal formula, undiscriminatingly used to express different and sometimes contradictory ideas. Thus, in the setting of one set of circumstances, “assump*69tion of risk” has been used as a shorthand way of saying that although an employer may have violated the duty of care which he owed his employee, he could nevertheless escape liability for damages resulting from his negligence if the employee, by accepting or continuing in the employment with “notice” of such negligence, “assumed the risk.” In such situations “assumption of risk” is a defense which enables a negligent employer to defeat recovery against him. In the setting of a totally different set of circumstances, “assumption of risk” has a totally different meaning. Industrial enterprise entails, for all those engaged in it, certain hazards to life and limb which no amount of care on the part of the employer can avoid. In denying recovery to an employee injured as a result of exposure to such a hazard, where the employer has in no sense been negligent or derelict in the duty owed to his employees, courts have often said that the employee “assumed the risk.” Here the phrase “assumption of risk” is used simply to convey the idea that the employer was not at fault and therefore not liable.

Plainly enough only mischief could result from using a single phrase to express two such different ideas. Such ambiguity necessarily does harm to the desirability of clarity and coherence in any civilized system of law. But the greater mischief was that in one of its aspects the phrase “assumption of risk” gave judicial expression to a social policy that entailed much human misery. The notion of “assumption of risk” as a defense — that is, where the employer concededly failed in his duty of care and nevertheless escaped liability because the employee had “agreed” to “assume the risk” of the employer’s fault — rested, in the context of our industrial society, upon a pure fiction. And in all English-speaking countries legislation was necessary to correct this injustice. In enforcing such legislation the courts should not lose sight of the ambiguous nature of the doctrine with which the *70legislation dealt. In giving effect to the legislative policy, care must be taken lest such ambiguity perpetuate the old mischief against which the new legislation was directed.

Our present concern is with the Federal Employers’ Liability Act. Prior to 1939, the only inroad made by the Act upon the doctrine of “assumption of risk” as a defense to liability arising from negligence was that in any action brought by an employee, he “shall not be held to have assumed the risks of his employment in any case where the violation by said common carrier of any statute enacted for the safety of employees contributed to the injury or death of such employee.” Section 4 of the Act as amended April 22, 1908, c. 149, 35 Stat. 65. The provision was construed, naturally enough, to mean that “the assumption of risk as a defense is abolished only where the negligence of the carrier is in violation of some statute enacted for the safety of employees. In other cases, therefore, it is retained.” Jacobs v. Southern Ry. Co., 241 U. S. 229, 235. By only partially withdrawing the defense of “assumption of risk,” Congress enabled the railroads to avoid liability in many situations where the employee’s injury resulted from the negligence of the carrier in the only way in which an employer can be negligent, namely, through the negligence of its servants. In other words, Congress continued to sanction the fiction of attributing to employees a willingness to bear the consequences of the carrier’s negligence, other than that arising from its violation of a statute enacted for the safety of employees.

This was the unfortunate situation which the 1939 amendment, the Act of August 11, 1939, c. 685, 53 Stat. 1404, sought to remedy. To § 4 was added the provision that in any action brought by an employee he “shall not be held to have assumed the risks of his employment in any case where such injury or death resulted in whole or in part from the negligence of any of the officers, agents, *71or employees of such carrier. . . .” The effect of this provision is to make it clear that, whatever other risks an employee may assume, he does not “assume the risk” of the negligence of the carrier or its other employees. Once the negligence of the carrier is established, it cannot be relieved of liability by pleading that the employee “assumed the risk.”

But the 1939 amendment left intact the foundation of the carrier’s liability — negligence. Unlike the English enactment which, nearly fifty years ago, recognized that the common law concept of liability for negligence is archaic and unjust as a means of compensation for injuries sustained by employees under modern industrial conditions, the federal legislation has retained negligence as the basis of a carrier’s liability. For reasons that are its concern and not ours, Congress chose not to follow the example of most states in establishing systems of workmen’s compensation not based upon negligence. Congress has to some extent alleviated the doctrines of the law of negligence as applied to railroad employees. By specific provisions in the Federal Employers’ Liability Act, it has swept away “assumption of risk” as a defense once negligence is established. But it has left undisturbed the other meaning of “assumption of risk,” namely, that an employee injured as a consequence of being exposed to a risk which the employer in the exercise of due care could not avoid is not entitled to recover, since the employer was not negligent.

The point is illustrated by two opinions of Mr. Justice Holmes. In Schlemmer v. Buffalo, R. & P. Ry. Co., 205 U. S. 1, 12-13, he called attention to the danger of relieving from liability for negligence by talking about “assumption of risk” — a danger resulting from the ambiguity of the phrase. “Assumption of risk” by an employee may be a way of expressing the conclusion that he has been guilty of contributory negligence. But an employee can*72not be charged with contributory negligence simply because he “assumed the risk”; the inquiry is, did his conduct depart from that of a reasonably prudent employee in his situation? As Mr. Justice Holmes admonished us in the Schlemmer case, “unless great care be taken, the servant’s rights will be sacrificed by simply charging him with assumption of the risk under another name.” Ibid. That case was decided before the Federal Employers’ Liability Act was in force. In a later case arising under the Act, Chesapeake & Ohio Ry. Co. v. Nixon, 271 U. S. 218, Mr. Justice Holmes for a unanimous Court reversed a judgment for the plaintiff on the ground that the employee’s death was caused by a failure to keep a lookout which was one of the “usual risks” of his employment. To be sure, this decision was made prior to the 1939 amendment, but in this respect that enactment makes no change in the law. The basis of an action under the Act remains the carrier’s negligence. The carrier is not to be relieved from the consequences of its negligence by any claim that the employee “assumed the risk” of its negligence. But neither is the carrier to be charged with those injuries which result from the “usual risks” incident to employment on railroads — risks which cannot be eliminated through the carrier’s exercise of reasonable care.

“Assumption of risk” as a defense where there is negligence has been written out of the Act. But “assumption of risk,” in the sense that the employer is not liable for those risks which it could not avoid in the observance of its duty of care, has not been written out of the law. Because of its ambiguity the phrase “assumption of risk” is a hazardous legal tool. As a means of instructing a jury, it is bound to create confusion. It should therefore be discarded. But until Congress chooses to abandon the concept of negligence, upon which the Act now rests, in favor of a system of workmen’s compensation not dependent upon negligence, the courts cannot discard the *73principle expressed, in one of its senses, by the phrase “assumption of risk,” namely, that a carrier is not liable unless it was negligent.

Perhaps no field of the law comes closer to the lives of so many families in this country than does the law of negligence, imbedded as it is in the Federal Employers’ Liability Act. It is most desirable, therefore, that the law should not be cloudy and confused. I am not at all certain that the Circuit Court of Appeals misconceived the nature and extent of the carrier’s liability after the 1939 amendment, rather than merely obscured its understanding by beclouding talk about “assumption of risk.” But since I agree that the District Court should have allowed the case to go to the jury on the issue of negligence, I concur in the decision.

delivered the opinion of the Court.

The petitioner’s husband and intestate, John Lewis Tiller, was a policeman for the respondent railroad. Among his duties was that of inspecting the seals on cars in railroad yards to make sure that no one had tampered with them. He had held this position for some years, was familiar with the yard, and was aware, in the words of the court below, that respondent’s employees “are instructed that they must watch out for the movement of the trains as no employee watches out for them and no lights are used at night on the head end of back-up movements except when an employee is placed at the back end with a lantern to protect a road crossing.” The Circuit Court of Appeals found that there was evidence sufficient to sustain the following account of the tragedy:

On the night of March 20, 1940, Tiller was standing between two tracks in the respondent’s switch yards, tracks which allowed him three feet, seven and one-half inches of standing space when trains were moving on both sides. *56The night was dark1 and the yard was unlighted. Tiller, using a flashlight for the purpose, was inspecting the seals of the train moving slowly on one track when suddenly he was hit and killed by the rear car of a train backing in the opposite direction on the other track. The rear of the train which killed Tiller was unlighted although a brakeman with a lantern was riding on the back step on the side away from Tiller. The bell was ringing on the engine but both trains were moving, and the Circuit Court found that it was “probable that Tiller did not hear cars approaching” from behind him. No special signal of warning was given.

Petitioner brought this suit to recover damages under the Federal Employers’ Liability Act, 45 U. S. C. § 51 et seq. The complaint alleged negligent operation of the car which struck defendant and failure to provide a reasonably safe place to work. Respondent denied negligence, pleaded contributory negligence on the part of the defendant, and set up as a separate defense that the deceased had assumed all the risks “normally and necessarily incident to his employment.” After the plaintiff’s evidence had been heard the defendant moved for a directed verdict on the grounds (a) that the evidence disclosed no actionable negligence and (b) that the cause of the death was speculative and conjectural. The motion was granted, judgment was accordingly entered for the defendant and the Circuit Court of Appeals, interpreting the decision of the district court as resting on a conclusion that the evidence showed no negligence, affirmed. 128 F. 2d 420. This result was based on a holding that the deceased had assumed the risk of his position and that therefore there was no duty owing to him by respondent. We granted certiorari because of the important question in*57volved in the Circuit Court of Appeals’ interpretation of the scope and effect of the 1939 amendment to the Federal Employers’ Liability Act, 53 Stat. 1404, 45 U. S. C. 54. The amendment provides that an “employee shall not be held to have assumed the risks of his employment in any case where such injury or death resulted in whole or in part from the negligence of any of the officers, agents, or employees of such carrier.”

The Circuit Court distinguished between assumption of risk as a defense by employers against the consequence of their own negligence, and assumption of risk as negating any conclusion that negligence existed at all. The court reasoned that if, for example, the respondent had negligently failed to provide a workman with a sound tool, and he was thereby injured, it could not under the amendment claim that he had assumed the risk of using the defective implement; but that if a workman were injured in the ordinary course of his work, as in such a switching operation as this, the assumption of risk might still be relied upon to prove that the respondent had no duty to protect him from accustomed danger. The court rejected petitioner’s argument that since the doctrine of assumption of risk had been abolished “the carrier can no longer interpose it as a shield against the consequences of its neglect and hence is liable for injuries to its employees in its railroad yards or elsewhere, unless it takes precautions for their safety commensurate with the danger that they are likely to encounter.” In rejecting this argument the court below put the core of its decision in these words: “The conclusion is inescapable that Congress did not intend to enlarge the obligation of carriers to look out for the safety of their men when exposed to the ordinary risks of the business, and that in circumstances other than those provided for in the amended section of the statute, the doctrine of the assumption of the risk must be given its accustomed weight.” [Italics added.]

*58We find it unnecessary to consider whether there is any merit in such a conceptual distinction between aspects of assumption of risk which seem functionally so identical, and hence we need not pause over the cases cited by the court below, all decided before the 1939 amendment, which treat assumption of risk sometimes as a defense to negligence, sometimes as the equivalent of non-negligence.2 We hold that every vestige of the doctrine of assumption of risk was obliterated from the law by the 1939 amendment, and that Congress, by abolishing the defense of assumption of risk in that statute, did not mean to leave open the identical defense for the master by changing its name to “non-negligence.” As this Court said in facing the hazy margin between negligence and assumption of risk as involved in the Safety Appliance Act of 1893, “Unless great care be taken, the servant’s rights will be sacrificed by simply charging him with assumption of the risk under another name;”3 and no such result can be permitted here.

Perhaps the nature of the present problem can best be seen against the background of one hundred years of master-servant tort doctrine. Assumption of risk is a judicially created rule which was developed in response to the general impulse of common law courts at the begin*59ning of this period to insulate the employer as much as possible from bearing the “human overhead” which is an inevitable part of the cost — to someone — of the doing of industrialized business.4 The general purpose behind this development in the common law seems to have been to give maximum freedom to expanding industry.5 The assumption of risk doctrine for example was attributed by this Court to “a rule of public policy, inasmuch as an opposite doctrine would not only subject employers to unreasonable and often ruinous responsibilities, thereby embarrassing all branches of business,” but would also encourage carelessness on the part of the employee.6 In the *60pursuit of its general objective the common law took many forms and developed many doctrines. One of the first was the fellow servant-assumption of risk rule which originated in Priestley v. Fowler the in Priestley v. Fowler, the Court said, “The servant is not bound to risk his safety in the service of his master, and may, if he thinks fit, decline any service in which he reasonably apprehends injury to himself: and in most of the cases in which danger may be incurred, if not in all, he is just as likely to be acquainted with the probability and extent of it as the master.”

As English courts lived with the assumption of risk doctrine they discovered that the theory they had created had become morally unacceptable but of such legal force that it could not be repudiated.7 8 The English sought to eliminate the fellow servant rule, which placed the burden of an employee’s negligence as it affected another employee on the injured person rather than on the business enterprise, by the Employers’ Liability Act of 18809 and found that the assumption of risk doctrine still left the employee in a hopelessly unprotected position. In the leading case *61of Thomas v. Quartermaine, 18 Q. B. D. 685 (1887), the court held that an employee standing on a three foot runway between two unfenced vats who was attempting to dislodge a piece of wood from, one of the vats and who by accident fell into the other and was scalded was barred from recovery. Since he had long known of the possible dangers of the narrow passage he was held to have assumed the risk of his position. In 1897 the English finally abandoned the common law remedy altogether as a protection for injured employees and adopted a workmen’s compensation law. 60 & 61 Vict. c. 37.

This Court accepted the assumption of risk doctrine as applied to railroad employees, at least in part, in 1879.10 That decision placed the employee’s assumption of risk upon the theory that an agreement to assume the risk was implied from the terms of the employment contract.

Prior to the passage of the Federal Employers’ Liability Act of 1906 the assumption of risk doctrine, except for a considerable vagueness as to its relation with contributory negligence, was fairly well known. 11 It had already been applied generally at the time of the adoption of the Act because of acceptance of the theory that the employee’s compensation was based upon the added risk to his position and that he could quit when he pleased. Tuttle v. Detroit, G. H. & M. Ry., supra; and compare for a restatement of this view after the passage of the Employers’ Liability Act, Seaboard Air Line v. Horton, 233 U. S. 492, 504.12 Federal and state courts, with some notable excep*62tions, accepted and applied the rule with all of its implications and consequences except when expressly prohibited from doing so by statute.13

Congress took a major step toward modification of the common law barrier against employee recovery in accident suits in the Federal Employers’ Liability Act of 1906, 34 Stat. 232, repassed with alterations not material in 1908, 35 Stat. 65. This Act, in its principal features, abolished the fellow servant rule, substituted comparative negligence for the strict rule of contributory negligence, and allowed survivors’ actions for tort liability. Section 4 of that Act, as interpreted by this Court in Seaboard Air Line v. Horton, supra, perpetuated the defense of assumption of risk.14 Unfortunately, from the standpoint of legal clarity, the Act as interpreted required careful distinction between assumption of risk and contributory negligence, since assumption of risk was an absolute bar to recovery *63while contributory negligence merely reduced the amount of recovery. The great uncertainty existing prior to the Act as to what the margin between these doctrines was15 thus became of real significance. The language of the statute itself seemed to impel the courts to practice “the niceties, if not casuistries, of distinguishing between assumption of risk and contributory negligence, conceptions which never originated in clearly distinguishable categories, but were loosely interchangeable until the statute attached such vital differences to them.” Pacheco v. N. Y., N. H. & H. R. Co., 15 F. 2d 467. For an attempt to distinguish between the doctrines, see Schlemmer v. Buffalo, R. & P. Ry. Co., supra, 12, and the same case at 220 U. S. 590, 596.

The assumption of risk clause in the statute became the subject of endless litigation. The Federal Code Annotated and the United States Code Annotated devote over thirty pages each of fine type merely to the citation and brief summary of the reported decisions; and the number of unreported and settled cases in which the defense was involved must run into the thousands.16 Aside from the difficulty of distinguishing between contributory negligence and assumption of risk many other problems arose. One of these was the application of the “primary duty rule” in which contributory negligence through violation of a company rule became assumption of risk. Unadilla Valley Ry. Co. v. Caldine, 278 U. S. 139; Davis v. Kennedy, 266 U. S. 147. Other complications arose from the introduction of “promise to repair,” “simple tool,” and “peremptory order” concepts into the assumption doc*64trine.17 In the disposition of cases the question of a plaintiff’s assumption of risk has frequently been treated simply as another way of appraising defendant’s negligence,18 as was done by the court below in the instant case.

It was this maze of law which Congress swept into discard with the adoption of the 1939 amendment to the Employers’ Liability Act, releasing the employee from the burden of assumption of risk by whatever name it was called. The result is an Act which requires cases tried under the Federal Act to be handled as though no doctrine of assumption of risk had ever existed.

If this were not sufficiently clear from the language of the amendment, any doubt would be dissipated by its legislative history. The 1939 bill19 was introduced by Senator Neely and was supported at the hearings by the railway labor unions. It was accepted both by the unions and the railroads that the bill would utterly and completely abolish the defense of assumption of risk.20 The report of the Senate Judiciary Committee struck at the *65basic reasons advanced by common law courts for the existence of the doctrine, declared it unsuited to present day activities, and described them as out of harmony with the equitable principles which should govern determinations of employer-employee responsibilities.21 The bill, as described in the report, was clearly aimed at making the principles of comparative negligence the guiding rules of decision in accident cases: “The adoption of this proposed amendment will, in cases in which no recovery is now allowed, establish the principle of comparative negligence, which permits the jury to weigh the fault of the injured employee and compare it with the negligence of the employer, and, in the light of the comparison, do justice to all concerned.”22

*66The purpose of the Act is made clearer upon analysis of the House bill which was rejected by the conference committee in favor of the Senate bill which is now the law. The House bill23 was intended to preserve some part of the doctrine of assumption of risk, preserving that defense except “where said employee has not had actual notice of any negligently maintained condition or practice.” The bill, unlike the Senate bill as the Representative reporting it explained, left untouched the rule of Toledo, St. L. & W. R. Co. v. Allen, 276 U. S. 165, “namely, that in the absence of special custom or unusual circumstances, a man who is run over by a switching movement cannot recover.” 24 It was the Allen opinion on which the court below in the instant case particularly relied. But the House bill, which the chief railroad counsel appearing before the Senate committee conceded would make no change in the existing law,25 was rejected in conference. The Allen case was specifically and caustically discussed at the Senate hearings, and the Senate bill was clearly aimed at ending its rule.26

The doctrine of assumption risk can not be “abolished in toto”27 and still remain in partial existence as the court below suggests. The theory that a servant is completely barred from recovery for injury resulting from his master’s negligence, which legislatures have sought to eliminate in *67all its various forms of contributory negligence, the fellow servant rule, and assumption of risk, must not, contrary to the will of Congress, be allowed recrudescence under any other label in the common law lexicon. The Act of 1908 and the amendment of 1939 abolish the post-Priestley v. Fowler defenses and authorize comparison of negligence instead of barring the employee from all recovery because of contributory negligence. They leave for practical purposes only the question of whether the carrier was negligent and whether that negligence was the proximate cause of the injury.

In this situation the employer’s liability is to be determined under the general rule which defines negligence as the lack of due care under the circumstances; or the failure to do what a reasonable and prudent man would ordinarily have done under the circumstances of the situation; or doing what such a person under the existing circumstances would not have done.28 A fair generalization of the rule is given in the Senate Committee report on the 1939 amendment: “In justice, the master ought to be held liable for injuries attributable to conditions under his control when they are not such as a reasonable man ought to maintain in the circumstances.”29 Of course in any case the standard of care must be commensurate to the dangers of the business. Hough v. Railway Co., 100 U. S. 213, 218; cf. Northern Pacific R. Co. v. Herbert, 116 U. S. 642, 652.

No case is to be withheld from a jury on any theory of assumption of risk; and questions of negligence should under proper charge from the court be submitted to the jury for their determination. Many years ago this Court said of the problems of negligence, “We see no reason, so *68long as the jury system is the law of the land, and the jury is made the tribunal to decide disputed questions of fact, why it should not decide such questions as these as well as others.” Jones v. East Tennessee, V. & G. R. Co., 128 U. S. 443, 445. Or as we have put it on another occasion, “Where the facts are in dispute, and the evidence in relation to them is that from which fair-minded men may draw different inferences,” the case should go to the jury.30

We think that the question of negligence on the part of the railroad and on the part of the employee should have been submitted to the jury. The decision below is reversed and the case is remanded for further proceedings in conformity with this opinion.

Reversed.