(dissenting).
Although the majority is “not unmindful of decisions elsewhere at variance” with its own, apparently it ascribes little significance to the fact that without exception every court which has considered this issue has held that the government is not liable for the negligence of its prison officials under the Federal Tort Claims Act. See James v. U. S., 280 F.2d 428 (8th Cir.), cert. denied, 364 U.S. 845, 81 S.Ct. 88, 5 L.Ed.2d 69 (1960) following Lack v. U. S., 262 F.2d 167 (8th Cir. 1958); Jones v. U. S., 249 F.2d 864 (7th Cir. 1957); Muniz v. U. S., 60 Civ. 1624, S.D.N.Y., Nov. 4, 1960, rev’d 305 F.2d 285 (2nd Cir. 1962) (this day); Berman v. U. S., 170 F.Supp. 107 (E.D.N.Y.1959); Golub v. U. S., Civ. No. 148-117, S.D.N.Y. Oct. 5, 1959; Collins v. U. S., No. T-1509, D.Kan., Jan. 29, 1958; Trostle v. U. S., No. 1493, W.D.Mo., Feb. 20, 1958; Van Zuch v. U. S., 118 F.Supp. 468 (E.D.N.Y.1954); Shew v. U. S., 116 F.Supp. 1 (M.D.N.C.1953); Sigmon v. U. S., 110 F.Supp. 906 (W.D.Va.1953); Ellison v. U. S., No. 1003, W.D.N.C., July 26, 1951.1 However, what disturbs me is that not only does the majority opinion “interpret” the words of the Act in a manner which has been rejected by Circuit and District Courts repeatedly, but that it does this without the support of a shred of relevant legislative history. As a result, the Court has filled the vacuum created by Congressional silence with its own notions of public policy, but not a policy legitimately attributable to Congress. Not since Shadrach, Meshach, and Abednego has goodness triumphed with such ease. But I fear that the price of this triumph is too great, for with a sweep of the hand we disregard the traditional tools of adjudication.
Statutory construction of the nature indulged in by the Court in this case is hazardous business. The principal danger, realized in this case, is that courts will tread where Congress has not. Speaking of this very problem Justice Frankfurter perceptively notes:2
“In the realms where judges directly formulate law because the ehosen lawmakers have not acted, judges have the duty of adaptation and adjustment of old principles to new conditions. But where policy is expressed by the primary lawmaking agency in a democracy, that is by the legislature, judges must respect such expressions by adding to or subtracting from the explicit terms which the lawmakers used no more than is called for by the shorthand nature of language.”
Justice Frankfurter recognizes that “there are not wanting those who deem naive the notion that judges are expected to refrain from legislating in construing statutes,” cf. Clark, Federal Procedural Reform and States’ Rights; to a More Perfect Union, 40 Tex.L.Rev. 211, 223-229 (1961), and he is not unaware that “judges may differ as to the point at which the line [between adjudication and legislation] should be drawn.” 3 Nevertheless, this renowned jurist of undoubted experience in these matters warns that “the only sure safeguard against crossing the line * * * is an alert recognition of the necessity not to cross it and instinctive, as well as trained, reluctance to do so.” 4
*259In the instant ease the facts alleged in the complaint evoke great sympathy. But in the Court’s eagerness to afford relief I believe it has too easily overcome its usual considered reluctance to abandon notions of judicial restraint. I understand its position; I appreciate its generosity; and I agree that there are occasions in which, as Justice Holmes recognized, “judges do and must legislate * * * interstitially.”5 But this decision does not reflect the “molar to molecular” motion which Holmes envisioned; and I cannot join the Court in making its “judicial leap.”
“The judge, even when he is free, is still not wholly free. He is not to innovate at pleasure. He is not a knight-errant roaming at will in pursuit of his own ideal of beauty or of goodness * * * He is not to yield to spasmodic sentiment, to vague and unregulated benevolence * * * ” Cardozo, The Nature of the Judicial Process, p. 144 (1921).
It is proverbial that hard eases make bad law. Perhaps it is but another way of stating the same idea to suggest that hard cases also induce courts to “make” law where it is plain under the circumstances that such is not their constitutional function.
It is particularly unfortunate that in the present case the unwarranted judicial legislation has been accomplished, as if by sleight of hand, through the majority’s willingness to assume the very question presented for decision. Its opinion, which treats the question of government liability as if it were being considered for the first time by a federal court, assumes that the absence of any explicit provision excluding prisoner claims from the coverage of the Federal Tort Claims Act necessarily indicates a Congressional intent to include them. But, insofar as the scope of the waiver of immunity contemplated by the Act is concerned, the doctrine of expressio unius has been expressly rejected by the Supreme Court. Feres v. U. S., 340 U.S. 135, 138-139, 71 S.Ct. 153, 95 L.Ed. 152 (1950). Canons of construction cannot save us from “the anguish of judgment.”6 It is not enough that the statute “by its terms does not except prisoners”; our inquiry must be directed to the question whether Congress “intended” to include them. Or, stated more accurately under the circumstances of this case, in John Chipman Gray’s often quoted words, it is up to this Court “to guess what it would have intended on a point not present to its mind, if the point had been present.” 7
Would Congress have intended that a statute which waived sovereign immunity and subjected the government to varied liability to the general public for negligent operation of post office trucks, army airplanes, etc. also superimposed upon the closely regulated government-prisoner relationship 8 a liability to prisoners for negligent operation of our penal system ? A majority of this panel of the Court says that “it would seem” that it does; I respectfully disagree. Instead I am constrained to agree with the Court of Appeals for the Seventh Circuit that “it seems unlikely that it ever occurred to any of the members of Congress that the claim would be made that the remedies under that Act would be available to an inmate of a federal correctional institution.” Jones v. U. S., supra, 249 F.2d at pp. 865-866. This is not to say that Congress, if confronted with the partic*260ular issue of government liability for injuries sustained by prisoners through negligence of its agents might not devise a scheme of compensation.9 However, I maintain that it is not for the Court to speculate whether such a remedy will be provided in the wisdom of that august body. Rather, it is whether Congress meant that the Federal Tort Claims Act should accomplish that purpose.
A number of arguments have been advanced by the government, and accepted by other courts in considered opinions, which suggest that this statute of general and undefined application was not “intended” to apply in the prisoner situation. The first, and most significant, relates to problems inherent in judicial review of action taken by prison authorities to enforce prison discipline. The case of Muniz v. U. S., supra, decided this day, is an excellent illustration of this point. Muniz claimed that he was beaten by other inmates during a prison riot and that he sustained permanent injuries of a grave nature. He contended that the prison authorities were negligent both in the general manner, in which they ran the prison, and in the steps taken to control the riot. The guards, it was alleged, had locked the rioting prisoners in their dormitory; and this prevented Muniz from seeking assistance from the authorities or escaping from his tormentors.10 In holding that Muniz may sue under the Act, the trial judge will have to examine almost every facet of prison administration. In addition, I should suppose that he will be required to substitute his conception of “reasonable behavior” for that of the persons charged by statute with the responsibility of running the prisons. In Feres v. U. S., supra, the Supreme Court thought that analogous problems relating to review of military decisions and soldier discipline suggested that the Federal Tort Claims Act could not reasonably be construed to permit soldiers’ claims against the government. See U. S. v. Brown, 348 U.S. 110, 112, 75 S.Ct. 141, 99 L.Ed. 139 (1954); Jefferson v. U. S., 178 F.2d 518, 520 (4th Cir. 1959), aff’d sub nom., Feres v. U. S., supra; see also Healy v. *261U. S., 192 F.Supp. 325, 326-329 (S.D.N.Y.), aff’d, 295 F.2d 958 (2nd Cir. 1961). Whether the dire-consequences which the government claims will result from imposition of liability are overdrawn or not, they are certainly not wholly fanciful; and they suggest that there is reasonable cause for investigation of facts, and evaluation of professional expertise on the subject before liability is assumed by the government. Congress has ample facilities for such investigation; we do not. And we do not know what Pandora’s box we are opening by permitting government liability under these circumstances for the first time.
The majority argues that it is up to Congress to decide if imposition of tort liability will adversely affect prison discipline. I agree. But I would think that Congress, faced with these problems would take pains to discover whether the adverse circumstances prophesied by the government will result; and that only after weighing all of the information available to it would it decide whether it is more desirable to permit prisoner suits against the government or whether this is an area where in the public interest it is better to continue to retain the cloak of sovereign immunity.
The government has urged upon us other considerations which would suggest that Congress did not “intend” that the Act should extend to the prisoner situation. It directs our attention to the source of the liability contemplated by the Federal Tort Claims Act itself. Under that Act the law of the place where the injury is sustained determines the existence and measure of the government’s liability. Therefore, under its provisions, the right of a prisoner to recover damages for his injury will depend upon the law of the place of his confinement. The Court dismisses an objection based on this result as being insubstantial, since the lack of uniformity in the treatment of the public and in the nature of the federal obligation was clearly intended by the framers of the Act. But this legislative design is of no significance unless we assume that such lack of uniformity is of approximately equal desirability in all instances in which the government may be held liable for the acts of its agents, an assumption which is demonstrably false.
It is some justification for dissimilar treatment of injured persons that they should be permitted to recover damages from the government only to the same extent as they might recover from any other tortfeasor in the state where the tort victim chose to be present. This is a large country; for purposes of tort law it is divided, in effect, into fifty juris: dictions. If a person chooses to live, work, or travel in Montana, he cannot claim that he is being unjustly treated because under Montana law, he cannot recover for injuries sustained there — although he might have been able to recover under the laws of New York or California, if he had been injured in those places. He cannot be heard to complain unless he is willing to challenge the nature of our entire federal system. Congress thought it reasonable to subject the federal government to liability within that multi-jurisdictional framework rather than to create a federal tort law which might afford remedial relief to either a greater or lesser degree than would otherwise be available to the injured person. The “justice” of this approach, from the view of the injured person, was thought to outweigh the need for uniform rules of federal liability.
It is quite another thing, however, to say that Congress “intended” to make an injured prisoner’s right to recover damages depend on the wholly fortuitous circumstance of the location of his prison, chosen not by him but by the Bureau of Prisons. For example, a New York dope addict will be confined most likely in Lexington, Kentucky; or a dangerous criminal, convicted in Maine, may be imprisoned off the coast of California. There are 31 federal institutions in 24 states. Assignment of a prisoner to any one of them depends upon a multitude of factors, of which geographical proximity to his home is but one. It seems to me *262that it would be unfair to make a lottery-out of the prisoner’s right to recovery. Why should his recovery be dependent upon the chance that the Director of the Bureau of Prisons will choose the “right” state with the “right” law for the inmate’s incarceration? Is the Director now to make his assignments by a roulette wheel, with the “lucky” prisoner being assigned to the “right” prison in the “right” state?
It quite escapes me how the force of this argument is weakened by the fact that the Act has been held to permit prosecution of a claim under Mississippi law for negligent operation of a local lighthouse which caused a ship to run aground in that state’s waters, Indian Towing Co., Inc. v. U. S., 350 U.S. 61, 76 S.Ct. 122, 100 L.Ed. 48 (1955); or that negligent acts of the Forest Service in Washington create liability under Washington law for injury to property in that state. Rayonier, Inc. v. U. S., 352 U.S. 315, 77 S.Ct. 374, 1 L.Ed.2d 354 (1957). In those instances the only question is whether local injuries ought to be compensated under local law or whether the principle of uniformity of federal obligations requires otherwise. As I have said, more closely akin to the facts in this case are those found in Feres v. U. S., supra. There, the Supreme Court believed it made “no sense” to predicate liability for soldiers’ injuries “upon geographic considerations over which they have no control and to laws which fluctuate in existence and value,” 340 U.S. p. 143, 71 S.Ct. p. 158. Once more the analogy is approximate because of different facts. But it is plainly relevant, and other courts11 have considered it controlling, “for like reason doth make like law.” Coke, First Institute, 10. Not only is Feres similar by virtue of the fact that neither the soldier nor the prisoner has any choice concerning the place in which he must reside, but because in each of these instances the relationship of the parties is of a peculiarly federal nature. Lack v. U. S., supra, 262 F.2d at p. 169.
Finally, the government argues quite forcefully that even Congress has indicated that it did not “intend” to allow prisoner claims against the government under the Federal Tort Claims Act.12 Proceeding on the reasonable assumption that the significance of an enactment may be understood by examining its antecedents, its later history, and its relation to other enactments,13 we observe that years before the passage of the Tort Claims Act Congress had provided a limited compensation scheme for injuries sustained by prisoners while engaged in activities sponsored by the Federal Prison Industries Board. See 18 U.S.C. § 4126. And years after that Act was adopted, in 1961 to be precise, when the Attorney General proposed new legislation intended to provide “equal treatment to all prisoners who may be injured in the course of employment while con*263fined,”14 the House Committee on the Judiciary noted that:
“Presently there is no way under the general law to compensate prisoners injured while so engaged. Their only recourse has been to appeal to Congress, and this Committee has reported numbers of private relief bills for such prisoners.” 15
That proposed legislation became Public Law 87-317, 75 Stat. 681 on September 26, 1961. It seems to me that it is especially significant that although Congress was aware of the fact that prisoners could not recover damages for injuries “under the general law,” and saw fit to provide an extension of compensation-type relief to certain additional prisoners, it did not undertake to provide them with a comprehensive remedy under the “general law,” or to provide any remedy at all for other prisoners, such as Henry Winston, who are injured as a result of non-work activities. Is this not abundant indication that Congress is undertaking a gradual and selective program in dealing with the peculiar problem of compensation for injuries suffered by prisoners? Does this legislation not suggest that Congress prefers the use of administrative rather than judicial machinery for this purpose ? Is it reasonable to suppose that Congress would be cautiously extending this compensation-type remedy, limited in scope, and dependent wholly upon the discretion of the Attorney General,16 if it had already made available to all prisoners comprehensive relief under the Tort Claims Act? I am driven to a negative answer. In the clear absence of evidence of Congressional “intention”' concerning the coverage of the Tort Claims Act, it is apparent that this recent enactment by Congress casts doubt upon the Court’s interpretation of the Act. The majority opinion dismisses this argument by attempting to use another canon of construction, to wit, action taken by Congress several months ago “is not admissible on the intent of an earlier Congress.” But, Justice Frankfurter gives us the answer to this: “to illumine these dark places in legislative composition all the sources of light must be drawn upon.”17
The majority dismisses as of little significance the fact that Congress has repeatedly considered and passed private legislation to compensate prisoners for injuries which result from negligent acts of prison officials. These private bills reflect complete awareness by Congress of the unanimous judicial opinion until today, that the Federal Tort Claims Act afforded no alternative relief.18 It may be true that private bills such as these are regarded as the “traditional preserve of individual Congressmen.” But this must be viewed in the light of our knowledge that “Congress has adopted a policy of not passing private bills where relief is available under the Tort Claims Act.” Lack v. U. S. supra, 262 F.2d at p. 171.
I submit that before the Court takes such a bold step in this delicate area, concerning as it does the entire field of treatment of transgressors by the government and the methods to be employed in protecting society, the Court ought to be reasonably certain that its decision in fact reflects the policy of Congress. In the light of the past history of litigation in this area and subsequent Congressional action, I do not think the decision of the Court is founded upon Congressional policy or judicial precedent. The time may have come when it is *264deemed- politically, socially, and economically wise to permit prisoners to entertain suits against the government. And such considerations may outweigh the possibility of damage to discipline in <our penal system or other problems of administration. But that is for Congress to decide. This- it has not done. We must be careful to avoid giving the impression that when judges think Congress has been too slow in legislating, they will assume the duties of “knights-errant,” and will find the means (under the guise of “interpretation”) to show their impatience. In an instance where legislative “intent” and judicial precedent is so clearly the other way, this is dangerous dogma. I would follow the course taken by the Supreme Court in Feres v. U. S., supra, which in denying relief to servicemen under the statute for reasons similar to those espoused in this dissenting opinion said:
“There are few guiding materials for our task of statutory construction. No committee reports or floor debates disclose what effect the statute was designed to have on the problem before us, or that it -even was in mind. Under these circumstances, no conclusion can be above challenge, but if we misinterpret the Act, at least Congress possesses a ready remedy.” 340 U.S. 135, 138, 71 S.Ct. 153, 155.
How much more forceful is this admonition when we observe the long line of cases interpreting the Act as affording no relief to prisoners asserting claims such as this one,19 and that there has been no action by Congress over the years to “remedy” any “misinterpretation."
. But see Lawrence v. U. S., 193 F.Supp. 243 (N.D.Ala.1961).
. Westin, The Supreme Court: Views from Inside, p. 83 (1961).
. Id., p. 82.
. Id.
. Southern Pacific Co. v. Jensen, 244 U.S. 205, 221, 37 S.Ct. 524, 531, 61 L.Ed. 1086 (1917).
. Westin, op. cit. supra, at p. 91 (Frankfurter, J.).
. Gray, Nature and Sources of the Law: Statutes (2 ed. 1921). See also Learned Hand’s concurring opinion in Guiseppi v. Walling, 144 F.2d 608, 624, 155 A.L.R. 761 (1944):
“As nearly as we can, we must put ourselves in the place of those who uttered the words, and try to divine how they would have dealt with the unforeseen situation; and, although their words are by far the most decisive evidence of what they would have done, they are by no means final * * * ”
. See Lack v. U. S., supra, 262 F.2d at p. 169.
. As the majority opinion points out, Congress has seen fit to provide a limited compensation scheme for prisoners injured in any ■work activity. The Court of Appeals for the 8th Circuit suggests with much persuasiveness that “if Congress had intended to create a cause of action for negligence in favor of the * * federal prisoner, it would * * * likely * * * have placed a limitation on the amount of recovery * * Lack v. U. S., supra, at p. 171.
. Although it is difficult to conceive of a more vivid example of the extent of “outside” interference with prison operations which will result from the imposition of judicial scrutiny, other situations come to mind quite readily.
For instance, it regularly occurs that prisoners inflict severe injuries on “unpopular inmates.” This may result from the operation of kangaroo courts, or from personal grudges, racial hatreds, or problems related to sex deviation. Frequently these assaults occur when the prisoners are permitted to congregate in large groups during leisure hours. When this happens, under the decision of the Court in this ease, will the trial judge have to decide whether it was unreasonable to allow the prisoners to congregate without very extensive supervision? If this does constitute negligence, will there arise some per se liability for the consequences, even if it is shown that for lack of Congressional appropriation the prisons are not adequately staffed for such purposes? Or, are we inviting prison authorities to restrict such leisure periods lest assaults be attributed to an inadequate number of guards? How many guards would be adequate? One for each prisoner or one for ten prisoners? These questions are unanswerable without more information which Congress could obtain after a full hearing on proposed legislation. Is there not a danger that we may tempt the prison authorities, in an excess of caution, to revert to outmoded methods of strict disciplinary administration?
Moreover, we are dealing with all sorts of inmates who have been isolated for society’s own protection. Many of them resent society, the judge who sent them to prison, their families, etc. There is more or less continuing opportunity for violence in prisons, and psychiatrists look upon this as a type of “safety valve” for the release of these resentments.
. See Jones v. U. S., supra, 249 F.2d at p. 866: “It is hardly conceivable that Congress intended in the passage of the Tort Claims Act to give prisoners rights of action ‘in accordance with the law of the place where the act or omission occurred,’ ” quoting Sigmon v. U. S., supra, 110 F.Supp. at p. 908; Lack v. U. S., supra, 262 F.2d at p. 169.
. It also urges that it does not follow that because suits against individual prison employees have been maintained on occasions, suits against the government should therefore be permitted. I agree, for to apply this reasoning is to do little more than put an unwarranted gloss an the area of the law under discussion. Essentially, we are dealing with the sovereign’s traditional right to immunity from suit in the absence of express waiver. Surely the majority does not seriously urge that because suits against employees have been possible, and there has not been either a multiplicity of suits or impairment of prison discipline, it follows that a like situation will result if the government can be sued directly. Does not the ability of the defendant to make good a judgment against it play an important part in the consideration of whether to bring a suit? The question, as Chief Justice White used to say, answers itself.
. See Westin, op. cit., supra, pp. 83-86.
. H.Rep. No. 534, 87th Cong., 1st Sess. 3.
. H.Rep. No. 534, supra, at p. 2.
. See 63 Yale L.J. 423 (1954).
. Baltimore & Ohio R. R. v. Kepner, 314 U.S. 44, 60, 62 S.Ct. 6, 13, 86 L.Ed. 28 (1941).
. See e. g., Act of July 14, 1956, Private Law 773, Chapter 615, 70 Stat. A124, and the accompanying report of the Senate Committee on the Judiciary, S. Rep. No. 1976, 84th Cong., 2d Sess. 2, quoted in Lack v. U. S., supra, 262 F.2d at p. 171.
. See Klein v. U. S., 268 F.2d 63, 64 (2nd Cir. 1959), in which a panel of this Court, in another situation, thought that these cases provided “persuasive analogy” for denial of relief under the Act.