This is an appeal from the dismissal of appellant’s habeas corpus petition attacking his confinement in Saint Elizabeths Hospital under the Sexual Psychopath Act, adopted in 1948. 22 D.C.Code §§ 3503-11 (1967).
Appellant was first committed to Saint Elizabeths in 1952, when he was 18 years old, on account of his tendency to indecently expose himself in public. Indecent exposure is ordinarily punishable by a jail sentence of not more than 90 days.1 22 D.C.Code § 1112(a) (1967). Appellant was confined in Saint Elizabeths for 15 years. Upon his release in 1967, he married a woman he had met as a fellow patient; they now have one child.2 Several months after his release, however, he was arrested on charges of six different acts of indecent exposure and was once again determined to be a “sexual psychopath” within the meaning of 22 D.C.Code § 3503(1) .3
At the commitment hearing in the Court of General Sessions, both the examining psychiatrists from D.C. General Hospital, Drs. Boschulte and Gordon, concluded that appellant- was a sexual psychopath under the Sexual Psychopath Act, but recommended outpatient care in lieu of confinement in Saint Elizabeths. They testified that renewed hospitalization would “increase his emotional tensions” and “remove him for many years from self-support or support of his family” without “benefit [to] anyone.” Instead, they prescribed a detailed program of outpatient therapy, coupled with medication to alleviate his tensions. Conceding that appellant’s recent offenses had been committed while he was receiving outpatient treatment, the doctors believed these aberrances were caused by the temporary absence of “two very important people.” Together with *1097appellant’s counsel, these psychiatrists offered to organize a treatment program which would offer some hope of preventing recurrences of such offensive conduct while appellant remained in the community.
While expressing a decided inclination to accept this offer, the court concluded that the terms of the Act precluded consideration of. less drastic alternatives to indefinite confinement even if they were clearly preferable.4 Accordingly, appellant went off to Saint Elizabeths again, whence he filed the instant unsuccessful petition for habeas corpus.
The burden of appellant’s argument is that proceedings under the Sexual Psychopath Act are essentially equivalent to ordinary civil commitments; that ordinary civil commitments, under the more recent Hospitalization of the Mentally 111 Act5 of 1964, require that possible dispositions less restrictive than total confinement must be explored and found to be inadequate before confinement can be ordered;6 and that since sexual offenders as a class are not more dangerous than the ordinarily commit-table mentally ill as a class,7 his confinement without consideration of possible less restrictive dispositions deprived him of the equal protection of the laws.8 Accordingly, he says the provision of the 1964 Act concerning alternative dispositions must be read into the Sexual Psychopath Act in order to save it from constitutional infirmity.9
Subsequent to the dismissal of appellant’s petition below and to the filing of his brief in this court, we announced our decision in Millard v. Harris,10 in which we undertook an examination of the Sexual Psychopath Act in light of its legislative history and of the later passage of the 1964 Act. We concluded that the scope of the Sexual Psychopath Act of 1948 must be significantly restricted in two ways in order to avoid serious constitutional problems that would be raised by a broad construction of the statutory term “sexual psychopath.” First, the words “not insane” must be read to mean “not ‘mentally ill.’ ” Second, a finding of “dangerousness” must be based on a high probability of substantial injury.11
In the proceedings below, appellant conceded that he was a “sexual psychopath” under the construction of the statute that prevailed prior to Millard. But *1098both appellant and appellee agree that, unless Millard is to be abandoned, this case must be remanded to determine whether appellant is in fact a sexual psychopath under the statute as now construed. In aid of the remand, we consider the import of Millard and its application to this case.
I.
Prior to the 1964 Act, commitment as a sexual psychopath and so-called civil commitment were mutually exclusive. By its terms, the Sexual Psychopath Act applies only to persons who are “not insane”;12 and the civil commitment law prior to 1964 authorized commitment only of those who were “insane.”13 When these statutes were enacted, “insanity” was largely equated with “psychosis.” 14 The Sexual Psychopath Act thus filled a gap in the commitment law.15 It was intended to be a humanitarian alternative to punishment for mentally disturbed potential sexual offenders who, some thought, could not be civilly committed and who, under the M’Naghten Rule then in effect, could not even plead “insanity” to criminal charges arising from their uncontrollable misconduct.
The term “insanity” and the attitudes it reflected were increasingly found artificial and inadequate16 Recognizing this trend, Congress at the instance of Senator Ervin of North Carolina undertook an extensive examination of the civil commitment law. This effort brought forth the 1964 Act, which its drafters hoped would serve as an enlightened model for state legislation. It authorizes compulsory treatment for one suffering from “mental illness” and “likely to injure himself or other persons.” 21 D.C.Code § 545(b) (1967). “Mental illness” was defined as any
psychosis or other disease which substantially impairs * * * mental health * * *
21 D.C.Code § 501 (1967) (emphasis added). As we noted in Millard, this statute immediately raises a question as to whether there remains any gap for the Sexual Psychopath statute to fill.17 That is, it may well be that anyone committable under the Sexual Psychopath Act is mentally ill and likely to injure others, and therefore also committable under .the 1964 Act.18 If so, commitment under the Sexual Psychopath Act would deprive a potential sexual offender of several important protections available to him under the 1964 Act. Among these are procedural safeguards 19 and the requirement of a judicial inquiry into less restrictive alternatives to confinement.20 Such discrimination against the mentally ill whose illness manifests itself in sexual conduct would raise serious constitutional problems.21 The sexual nature of anticipated “dangerous” conduct may be relevant to the type of treatment to be given, but it cannot justify arbitrary differences in procedures for commitment or conditions of confinement.22
*1099In Millard we declined to leap to the conclusion that there was no longer any permissible role for the Sexual Psychopath Act.23 Instead, we sought to save the statute by construing it to avoid the equal protection problems which would arise from an overlap with the new civil commitment law. We held that the statutory words “not insane” must now be read to mean “not ‘mentally ill’ ” within the meaning of the 1964 Act.24 Thus, the 1948 Sexual Psychopath Act now applies only to those who are not “mentally ill,” while compulsory treatment of those who are “mentally ill” is governed by the 1964 Act. This construction restores the original relationship of mutual exclusivity between sexual psychopath commitments and other civil commitments. Under Millard, it remains for future cases to show whether there are in fact any dangerous sexual recidivists who are not “mentally ill” within the broad meaning of the 1964 Act.
In the case at bar, the examining doctors concluded, and appellant conceded, that he is “not insane”; but the doctors had no occasion to consider whether he was nonetheless “mentally ill.” Since there is accordingly no record on this question, we must remand for a hearing and findings of fact necessary to a determination whether the statute was properly applied to appellant.
II.
In Millard we found it unnecessary to decide whether the appellant was mentally ill, because from the “plethora of evidence”25 in the record we could conclude that he was not “dangerous to other persons” within the meaning of the Sexual Psychopath Act, and thus was not committable under it. In the instant case, appellant’s dangerousness was not in issue below, so the record does not permit us to say whether he is dangerous. On remand, the court will have to decide that question if it determines that he is not mentally ill.26
We attempted in Millard to provide an analytical framework to guide lower courts in applying the conclusory term “dangerous to others.” 27 Without some such framework, “dangerous” could readily become a term of art describing anyone whom we would, all things considered, prefer not to encounter on the streets. We did not suppose that Congress had used “dangerous” in any such Pickwickian sense. Rather, we supposed that Congress intended the courts to refine the unavoidably vague concept of “dangerousness” on a case-by-case basis, in the traditional common-law fashion.
This does not mean, however, that the statutory language may be disregarded. To be “dangerous” for the purposes of the Sexual Psychopath Act, one must be
likely to attack or otherwise inflict injury, loss, pain, or other evil on the objects of his desire 28
The focus of the statute is not on expected conduct, but on the harm that may flow from that conduct. Commitment cannot be based simply on the determination that a person is likely to *1100engage in particular acts. The court must also determine the harm, if any, that is likely to flow from these acts. A mere possibility of injury is not enough; the statute requires that the harm be likely. For no matter how certain one can be that a person will engage in particular acts, it cannot be said that he is “likely to * * * inflict injury” unless it can also be said that the acts, if engaged in, are likely to result in injury.29
These determinations must be made on the basis of the record in the particular case before the court. The expert testimony will therefore be relevant to three questions of fact: (1) the likelihood of recurrence of sexual misconduct ; (2) the likely frequency of any such behavior; and (3) the magnitude of harm to other persons that is likely to result.
Having found the facts, the court must then determine as a matter of law — in this case, as a matter of statutory construction — whether those facts provide a legal basis for commitment. Two questions must be answered in making this determination. The first is what magnitude of harm will justify commitment. It is clear that Congress did not intend to authorize indefinite preventive detention for those who have a propensity to behave in a way that is merely offensive or obnoxious to others;30 the threatened harm must be substantial.31 Thus, commitment under the Sexual Psychopath Act requires that a person be found likely to engage in sexual misconduct in circumstances where that misconduct will inflict substantial injury upon others.
The second question is what likelihood of harm will justify commitment. It may well be impossible to provide a precise definition of “likely” as the term is used in the statute. The degree of likelihood necessary to support commitment may depend on many factors. Among the particularly relevant considerations are the seriousness of the expected harm, the availability of inpatient and outpatient treatment for the individual concerned, and the expected length of confinement required for inpatient treatment.32
It is particularly important that courts not allow this second question to devolve, by default, upon the expert witnesses.33 Psychiatrists should not be asked to tes*1101tify, without more, simply whether future behavior or threatened harm is “likely” to occur. For the psychiatrist “may — in his own mind — be defining Tikely’ to mean anything from virtual certainty to slightly above chance. And his definition will not be a reflection of any expertise, but * * * of his own. personal preference for safety or liberty.”34 Of course, psychiatrists may be unable or unwilling to provide a precise numerical estimate of probabilities, and we are not attempting to so limit their testimony. But questioning can and should bring out the expert witness’s meaning when he testifies that expected harm is or is not “likely.” Only when this has been done can the court properly separate the factual question — what degree of likelihood exists in a particular case — from the legal one — whether the degree of likelihood that has been found to exist provides a justification for commitment.
Millard had been committed because of exhibitionism and masturbation in public, but he had performed no such acts during six years of hospitalization. Extensive expert testimony was presented at his habeas corpus hearing to show that he was likely to engage in such conduct only on infrequent occasions. The experts testified that isolated exposures to acts of this nature were likely to cause psychological harm only to a small group of uniquely sensitive women.35 If Millard would but rarely expose himself at all, he was prima facie unlikely to do so in the presence of a member of this small group. Consequently, the record showed that he was not likely to inflict substantial harm on others — at least in the absence of evidence as to the vulnerability to injury of any likely viewers of his occasional self-exposures.
Appellant is, like Millard, a potential exhibitionist, though the record does not disclose that he is also a public masturbator. There is no indication in the record that he has ever been violent or assaultive. On the other hand, unlike Millard, he has apparently engaged in numerous recent acts of indecent exposure. Thus, one of the predicates of our holding that Millard was not dangerous under the Act is not present in this case.
Appellant may, however, be able to demonstrate that he will not be a frequent offender. His doctors offer an explanation for his recent spate of lapses which suggests that a repetition may be avoidable. The likelihood that appellant will voluntarily accept a course of outpatient treatment which will prevent fre- t quent lapses is a factor to be considered in determining whether he is likely to inflict substantial injury on others. And even if appellant cannot show that his future acts of exhibitionism will be infrequent, he may nevertheless be able to show that they will be harmless. In light of evidence as to the character and size of the likely viewing audience and the harm it may suffer, the District Court will have to consider whether appellant’s potential exhibitionism can be deemed a sufficiently grave danger to warrant an indeterminate commitment.
III.
We have not, either here or in Millard, ] decided any constitutional questions. But S when a determination of “dangerousness” will result in a deprivation of liberty, no ^ court can afford to ignore the very real constitutional problems surrounding in- { carceration predicated only upon a sup- ' posed propensity to commit criminal acts. ¿ Incarceration may not seem “punishment” to the jailors, but it is punishment / to the jailed.36 Incarceration for a mere *1102propensity is punishment not for acts, but for status,37 and punishment for status is hardly favored in our society.38 In essence, detention for status is preventive detention.39
Only a “blind court”40 could ignore the intense debate, in and out of Congress, over the extent to which the Constitution can tolerate preventive detention.41 Similar questions have been raised sporadically for years,42 but the problem has rarely been analyzed43 It may be that in some circumstances preventive detention is in fact permissible. If so, such detention would have to be based on a record that clearly documented a high probability of serious harm,44 and circumscribed by procedural protections as comprehensive as those afforded criminal suspects.45 Detention for any significant period of time would have to be attended by periodic review 46 as well as continuing assurance of bona fide efforts at treatment suited to the particular individual detained.47
Unquestionably, Congress may prohibit acts of exhibitionism even if such acts are unlikely to do serious harm; and Congress may punish willful violations of laws forbidding indecent behavior. But the test of what anticipated conduct may justify preventive detention cannot be simply whether the legislature has power to prohibit such conduct or to attack the evil it portends.48 Congress may legislate to protect many different interests — psychic and esthetic as well as physical and economic. But while it may prohibit ugly billboards because they give offense, it may not lock up ugly people for the same reason. The power to control an evil does not remove all restric*1103tions on the means that may be employed for that purpose.49 This principle is fundamental to the constitutional order.
On the present record, confinement of appellant under the Sexual Psychopath Act would deprive him of his liberty indefinitely — and perhaps permanently— vfor a propensity to commit acts punishable by a fixed jail sentence. Moreover, confinement would ignore, and apparently frustrate, his treatment needs. Confinement for a mere propensity is preventive detention. Particularly when the act in question is commonly punishable only by a short jail sentence, indefinite confinement, even though labeled “civil,” is preventive detention with a vengeance.50 If required by the Sexual Psychopath Act, it would raise not only one but many difficult constitutional issues:
(1) Is the harm threatened by a potential exhibitionist sufficiently serious to provide a constitutional justification for indefinite deprivation of liberty ? 51
(2) Is the possibility that no harm will in fact result from appellant’s future conduct sufficiently large to make incarceration based on possible harm arbitrary and capricious and therefore in violation of the due process clause?52
(3) Does the Sexual Psychopath Act provide adequate procedural due process to permit indefinite detention ? 53
(4) Is the absence in the Sexual Psychopath Act of the procedural protections accorded those sought to be committed under the 1964 Act a denial of equal protection? 54
(5) In view of 1964 Act, does the equal protection clause require consideration of the adequacy of less restrictive alternatives to hospital confinement for a “sexual psychopath” ? 55
(6) If less restrictive alternatives would in fact adequately protect the public while best promoting appellant’s rehabilitation, is confinement a deprivation of his liberty without the justification required by the due process clause? 56
(7) If appellant’s need for treatment requires that he not be confined, is indefinite confinement because of his condition cruel and unusual punishment ? 57
These hard questions may be avoided if appellant is not a “sexual psychopath.” 58 If he is not, he is either pot committable at all, or committable only under the 1964 Act, which requires the *1104consideration of alternatives he seeks in the instant petition.59
IV.
The dissenting opinion is in large part an attack on Millard v. Harris, which controls our decision today.60 But although it misreads the holdings and rationales of both decisions,61 it raises issues of such
importance that with all due respect, we are constrained to address ourselves to some of the dissent’s statements.
First. On remand, the District Court will have to determine whether appellant is “mentally ill” and therefore within the purview of the Hospitalization of the Mentally 111 Act. But if it determines that he is not “mentally ill,” it cannot avoid determining whether he is “dangerous” within the meaning of the Sexual Psychopath Act. In light of this, it is difficult to understand the objection to defining the statutory concept of “dangerous” for the guidance of the District Court on remand.62 There is ample precedent for our action here,63 and principles *1105of sound judicial administration require us to make our remand order intelligible to the court and parties below.64
Second. The dissent apparently would construe the statute so as either to read out of it the word “likely” or to read into it an unascertainable congressional intent to confine all exhibitionists as “dangerous.”65 Even if the statute were fairly susceptible of either construction, however, a long and consistent line of authority would counsel us not to adopt it “unless no choice is left.” United States v. Rumely, 345 U.S. 41, 46, 73 S. Ct. 543, 97 L.Ed. 770 (1953). For a statute “must be construed with an eye to possible constitutional limitations so as to avoid doubts as to its validity.” Lucas v. Alexander, 279 U.S. 573, 577, 49 S.Ct. 426, 428, 73 L.Ed. 851 (1929). This “principle of wisdom and duty” 66 has so often been repeated that it should no longer need discussion.67 But the pre*1106mature decision of constitutional questions against which the dissent so rightly inveighs could not be avoided if we were to follow its construction of the statute, since that construction would require us to confront the constitutional questions directly.
Third. Opinions are written for explication and instruction. Neither function would be served if, having concluded that the statute must be construed to avoid constitutional doubts, we then made no mention of the doubts we sought to avoid. Neither here nor in Millard have we decided any constitutional questions. But crucial factors affecting decisions are not to be kept as secrets known only to the court. Bench, bar, and litigants alike would be ill served if we were to conceal the bases for judgment.
Fourth. The dissent apparently considers the constitutional questions we have sought to avoid deciding so insubstantial that their very mention occasions rebuke. “This is not,” we are told, “a situation involving strictly preventive detention.” By our construction of the statute we have done our best to make this statement true. But the dissent would read the statute to bring this issue into the clearest possible focus. Appellant has been “treated” in Saint Elizabeths Hospital for fifteen years — his entire adult life. Both of the examining psychiatrists testified that further hospitalization will not improve — and may aggravate — his condition, while outpatient treatment offers some hope of a cure. Perhaps the consequences of his exhibitionism could be shown to present such a grave risk to the public that he may legally be confined indefinitely solely for their protection. But if this is to be done, we should at least acknowledge what we are doing. We are not confining an unfortunate man in order to provide him with treatment that will lead to his rehabilitation; we are locking him up for our own “protection,” in direct opposition to his treatment needs. Awareness of this fact might impel us to approach the problem with greater humility and caution.
Fifth. The dissent takes us to task for giving insufficient weight to “the possibility of serious psychological harm which might result to small children” (emphasis added). We are not unaware of this possibility.68 But the statute says that commitment is justified only if the person to be committed is “likely to * * inflict injury” on other persons.69 Whatever may be the requisite standard of likelihood,70 there is surely no warrant for reading “likely” as synonymous with “might possibly.”
Sixth. One premise lies at the core of the dissenting opinion: that exhibitionism is necessarily dangerous, no matter what the circumstances.71 This assumption so pervades the dissent that it can flatly state, with no citation of authority whatsoever, that Congress has decided that exhibitionism is dangerous; consequently, the dissent strenuously objects to what it conceives to be our disregard of Congressional intent.72 But the House *1107Report on the bill,73 the Senate Report,74 and the Congressional Record75 are empty of any indication that Congress even considered the question, let alone answered it. Determination of the legislative intent is often difficult, and even careful examination of the legislative history may leave grounds for disagreement. At the very least, however, courts should determine the intent of Congress from the legislative record, and not simply infer it from general notions of proper policy.
Seventh. The dissent, in successive paragraphs, first attempts to remove this case from the sphere of “preventive detention” by focusing on appellant’s past conduct, and next resurrects the tired cliche that “we are not here dealing with penal legislation.” The first of these paragraphs of course ignores both the language and operation of the Sexual Psychopath Act, which requires neither a criminal conviction nor any showing, attended by the procedural safeguards constitutionally required in criminal cases, that the person to be committed has engaged in specified proscribed conduct. And the two paragraphs taken together well illustrate the willingness with which many of us deny the reality of preventive detention. All too often courts justify a punitive disposition by looking to past conduct, while 'simultaneously ignoring the procedural requirements of criminal cases by invoking the false promise of “nonpenal” treatment and rehabilitation. “Non-criminal” commitments of so-called dangerous persons have long served as preventive detention, but this function has been either excused or obscured by the promise that, while detained, the potential offender will be rehabilitated by treatment. Notoriously, this promise of treatment has served only to bring an illusion of benevolence to what is essentially a warehousing operation for social misfits.
Predicting future behavior and evaluating its consequences is a uniquely difficult,- if not impossible task.76 It must be forthrightly confronted, not avoided by sugar-coating reality. There is no way on this record that we can escape the reality that a “penal” incarceration would set appellant free in 90 days, while the dissent’s “non-penal” solution would likely confine him for years, if not for the rest of his life. The record here contains u¿contradicted expert testimony that appellant’s continued confinement * would not “benefit anyone,” while outpatient treatment offers the best chance for improvement of his condition. Judicial speculation is no substitute for record evidence.
We remand to the District Court for proceedings not inconsistent with this opinion.
So ordered.
. If the exposure is knowingly made to a child under 16 years of age, the maximum jail term is one year. 22 D.C.Code § 1112(b) (1967).
. This information was supplied by appellant’s counsel on oral argument.
. 22 D.C.Code § 3508 (1967) provides: If tlie patient is determined to be a sexual psychopath, the court shall commit him to Saint Elizabeths Hospital to be confined there until released in accordance with section 22-3509. (Emphasis added.)
. 21 D.C.Code §§ 501-91 (1967) [hereinafter 1964 Act].
. 21 D.C.Code § 545(b) (1967); Lake v. Cameron, 124 U.S.App.D.C. 264, 364 F.2d 657 (1966).
. To be committed under the Sexual Psychopath Act, one must be
a person, not insane, who by a course of repeated misconduct in sexual matters has evidenced such lack of power to control his sexual impulses as to be dangerous to other persons because he is likely to attack or otherwise inflict injury, loss, pain, or other evil on the objects of his desire.
22 D.C.Code § 3503(1) (1967).
To be committed under the Hospitalization of the Mentally 111 Act, one must be mentally ill and, because of that illness, “likely to injure himself or other persons if allowed to remain at liberty.” 21 D.C. Code § 545(b) (1967).
. See Baxstrom v. Herold, 383 U.S. 107, 86 S.Ct. 760, 15 L.Ed.2d 620 (1966) ; Fuller v. United States, 129 U.S.App.D.C. 53, 55-56, 390 F.2d 468, 470-471 (1967) (opinion of Chief Judge Bazelon) ; Cameron v. Mullen, 128 U.S.App.D.C. 235, 241-244, 387 F.2d 193, 199-202 (1967).
. Fuller v. United States, supra note 8. Of. Bolton v. Harris, 130 U.S.App.D.C. 1, 395 F.2d 642 (1968) ; Millard v. Cameron, 125 U.S.App.D.C. 383, 373 F.2d 468 (1966).
. 132 U.S.App.D.C. 146, 406 F.2d 964 (1968).
. Id. at 151-153, 406 F.2d at 969-971 (not mentally ill) ; id. at 153-160, 406 F.2d at 971-978 (dangerousness). For the earlier construction, see, e. g., Carras v. District of Columbia, 183 A.2d 393 (D.C. Mun.App.1962).
. See note 7, supra.
. 53 Stat. 1299 (1939) (formerly 21 D.C. Code § 315; repealed 1964).
. Millard, supra note 10, 132 U.S.App. D.C. at 149, 406 F.2d at 967.
. This was explicitly recognized by Congress, which emphasized that treatment was to be afforded “sexual psychopaths” “in a manner similar to the treatment afforded insane persons.” H.R.Rep.No. 1787, 80th Cong., 2d Sess. 4 (1948), U.S. Code Cong.Serv. p. 1718.
. Id. at 149-150, 406 F.2d at 967-968.
. Id. at 151, 154, 406 F.2d at 969, 972. The 1964 Act’s definition of mental illness has been broadly construed. See In re Alexander, 125 U.S.App.D.C. 352, 854^ 355, 372 F.2d 925, 927-928 (1967).
. See Millard, supra note 10, 132 U.S. App.D.C. at 163-167, 406 F.2d at 981-985 (Wright, X, concurring).
. See Millard, supra note 10 at 152, 406 F.2d at 970.
. See Lake v. Cameron, 124 U.S.App.D.C. 264, 364 F.2d 657 (1966).
. See pp. 1101-1102, infra.
. Baxstrom v. Herold, 383 U.S. 107, 111, 86 S.Ct. 760 (1966) ; Cameron v. Mullen, 128 U.S.App.D.C. 235, 242-244, 387 F.2d 193, 200-202 (1967).
. Millard, supra note 10, 132 U.S.App. D.C. at 154, 406 F.2d at 972.
. Id. at 153, 406 F.2d at 971.
. Id. at 155, 406 F.2d at 973.
. Appellant conceded that he was a sexual psychopath (and therefore implicitly that he was “dangerous to others”) only as the statute had been construed prior to Millard. That concession of course will not preclude him from showing if he can that he is not dangerous under the statute as limited in Millard.
If appellant is mentally ill, his “dangerousness” will be in issue only if proceedings are instituted under the Hospitalization of the Mentally 111 Act. We note here only that the standard for compulsory treatment under that Act may well differ from the standard for compulsory commitment under the Sexual Psychopath Act. See note 7 supra for the different statutory formulations.
. Millard, supra note 10 at 155, 406 F.2d at 973 and Part VI.
. 22 D.C.Code § 3503(1) (1967) (emphasis added).
. See Millard, supra note 10 at 159-160, 406 F.2d at 977-978.
. See 94 Cong.Rec. 4887 (April 26, 1948) :
Mr. Chelf: When it is found that a person is a sexual pervert and he ought to be confined, is there a penalty, and, if so, what is the penalty? What is the final punishment?
.Mr. MacKinnon: Mr. Speaker, Title II of the bill [Title II became the Sexual Psychopath Act] does not deal with offenses of that character at all, except in extremely aggravated situations.
* * *
Title II of the bill is not aimed at the ordinary sexual pervert, nor is it limited to perverts. It is aimed at persons who are a dangerous menace to society and who by a repeated course of conduct have evidenced that they are a danger- . ous menace to society. * * *
(emphasis added). Mr. MacKinnon, who was prominent in the drafting of the Act, see id. at 4886, is of course now Judge MacKinnon of this court.
. Millard, supra note 10 at 155-160, 406 F.2d at 973-978.
. For an analysis of these and other relevant factors, see Livermore, Malmquist & Meehl, On the Justification for Civil Commitment, 117 U.Pa.L.Rev. 75 (1968). See also Leary v. United States, 395 U.S. 6, 32-36, 89 S.Ct. 1532, 23 L.Ed.2d 57 (1969), interpreting the “rational connection” test for statutory presumptions laid down in Mobile, J. & K. C. R. Co. v. Turnipseed, 219 U.S. 35, 31 S.Ct. 136, 55 L.Ed. 78 (1910), to require that “it can * * * be said with substantial assurance that the presumed fact is more likely than not to flow from the proved fact on which it is made to depend.” 395 U.S. at 36, 89 S.Ct. at 1548.
. For discussion of a similar problem in the context of the insanity defense in criminal cases, see Washington v. United . States, 129 U.S.App.D.C. 29, 36-42, 390 4. F.2d 444, 451-457 (1967).
. Dershowitz, “Psychiatry in the Legal Process: ‘A Knife that Cuts Both Ways,’ ” Address Delivered at the Harvard Law School Sesquicentennial Celebration, September 22, 1967.
. There was also testimony that small children might be harmed, but only by repeated exposure to abnormal sexual behavior. The record provided no grounds for finding any likelihood of repeated exposures. Millard, supra note 10, 132 U.S. App.D.C. at 160, 406 F.2d at 978.
. United States v. Brown, 381 U.S. 437, 458, 85 S.Ct. 1707, 14 L.Ed.2d 484 (1965).
. Robinson v. California, 370 U.S. 660, 666, 82 S.Ct. 1417, 8 L.Ed.2d 758 (1962) ; see id. at 678-679 (Harlan, X, concurring) .
. Robinson v. California, supra; Edwards v. California, 314 U.S. 160, 184-185, 62 S.Ct. 164, 86 L.Ed. 119 (1941) (Jackson, J., concurring).
. See Amsterdam, Federal Constitutional Restrictions on the Punishment of Crimes of Status, Crimes of General Obnoxiousness, Crimes of Displeasing Police Officers, and the Like, 3 Crim.L.Bull. 205, 225-227 (1967) ; cf. Ricks v. United States, 134 U.S.App.D.C. 215, 414 F.2d 1111, 1116 (Dec. 23, 1968).
. See United States v. Rumely, 345 U.S. 41, 44, 73 S.Ct. 543, 97 L.Ed. 770 (1953) ; Bailey v. Drexel Furniture Co. (Child Labor Tax Case), 259 U.S. 20, 37, 42 S.Ct. 449, 66 L.Ed. 817 (1922).
. See Hearings Before the Subcomm. on Constitutional Rights of the Senate Comm, on the Judiciary, “Amendments to the Bail Reform Act of 1966," 91st Cong., 1st sess. (1969).
. The classic statement is that of Mr. Justice Jackson, sitting as Circuit Justice in Williams v. United States, 184 F.2d 280, 282-283 (2d Cir. 1950) :
Imprisonment to protect society from predicted but unconsummated offenses is so unprecedented in this country and so fraught with danger of excesses and injustice that I am loath to resort' to it, even as a discretionary judicial technique to supplement conviction of such offenses as those of which defendants stand convicted.
. But see Hearings, supra note 41, and sources there cited; Livermore, Malmquist & Meehl, supra note 32.
. Cf. Whitney v. California, 274 U.S. 357, 378-379, 47 S.Ct. 641, 71 L.Ed. 1095 (1927) (Brandeis & Holmes, JJ., concurring) .
. In re Gault, 387 U.S. 1, 27-28, 87 S.Ct. 1428, 18 L.Ed.2d 527 (1967) ; Specht v. Patterson, 386 U.S. 605, 608-609, 87 S. Ct. 1209, 18 L.Ed.2d 326 (1967) ; cf. Minnesota ex rel. Pearson v. Probate Court, 309 U.S. 270, 275-276, 60 S.Ct. 523, 84 L.Ed. 744 (1940).
. Since any such detention would be based on present status rather than the commission of a past act, regular procedures for review would be essential to assure that a basis for the commitment continued to exist. Cf. Hirabayashi v. United States, 320 U.S. 81, 108, 63 S.Ct. 1375, 87 L.Ed. 1774 (1943) (Douglas, J., concurring) .
. Robinson v. California, supra note 37, 370 U.S. at 665, 82 S.Ct. 1417.
. Whitney v. California, supra note 44, 274 U.S. at 377-78, 47 S.Ct. 641.
. “[E]ven though the governmental purpose be legitimate and substantial, that purpose cannot be pursued by means that broadly stifle fundamental personal liberties when the end can be more narrowly achieved. * * * [by] less drastic means for achieving the same basic purpose.” Shelton v. Tucker, 364 U.S. 479, 488, 81 S.Ct. 247, 252, 5 L.Ed.2d 231 (1960).
. In Millard, we observed that since the sole justification for commitment under the sexual psychopath statute is * * * dangerousness to others * * * we must view the statute realistically ' as one which borders close upon preventive detention — detention which under our statute does not even require prior conviction of a criminal act. Millard, supra note 10, 132 U.S.App.D.C. at 155, 406 F.2d at 973.
. Whitney v. California, 274 U.S. 357, 377-378, 47 S.Ct. 641 (1927) (Brandeis & Holmes, JJ., concurring).
. See Leary v. United States, 395 U.S. 6, 36, 89 S.Ct. 1532 (1969).
. See Specht v. Patterson, 386 U.S. 605, 608, 87 S.Ct. 1209 (1967) ; of. Skinner v. Oklahoma, 316 U.S. 535, 545, 62 S.Ct. 1110, 86 L.Ed. 1655 (1942) (Stone, C. J., concurring).
. Baxstrom v. Herold, 383 U.S. 107, 111, 86 S.Ct. 760 (1966).
. Bolton v. Harris, 130 U.S.App.D.C. 1, 395 F.2d 642 (1968), and cases there cited.
. Shelton v. Tucker, 364 U.S. 479, 488, 81 S.Ct. 247, 5 L.Ed.2d 231 (1960) ; see Griswold v. Connecticut, 381 U.S. 479, 485, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965).
. Robinson v. California, 370 U.S. 660, 665, 82 S.Ct. 1417 (1962).
. Either because he is “mentally ill” or because he is not “dangerous to others.”
. See note 6 supra and accompanying text.
. The dissent suggests that courts ignore “the dicta of Millard” and “the wide-ranging dicta of today’s majority opinion.” Infra, p. 1108. Even if our construction of the statutory definition of “dangerousness” in Millard and here could fairly be considered dictum, the suggestion would be open to considerable doubt. See the opinion of Mr. Justice Cardozo, speaking for a unanimous Court in Hawks v. Hamill, 288 U.S. 52, 58-59, 53 S.Ct. 16, 77 L.Ed. 510 (1933) ; of. Brush v. Commissioner of Internal Revenue, 300 U.S. 352, 373, 57 S.Ct. 495, 81 L.Ed. 691 (1937). But in fact the dissent’s reference to “the dicta of Millard” and of today’s opinion can be read only as an invitation to trial courts to ignore the holdings of both cases. When there has been an “application of the judicial mind to the precise question,” Carroll v. Lessee of Carroll, 57 U.S. (16 How.) 275, 286, 14 L.Ed. 936 (1853), decision of the question is the holding of the case. Cohens v. Virginia, 19 U.S. (6 Wheat.) 264, 399, 5 L.Ed. 257 (1821) (Marshall, C. J.). We need not rest on so broad a ground, however. Since the result in Millard rests solely on the construction of “dangerousness,” that construction constitutes binding precedent. Any suggestion that trial courts disregard it violates principles of sound judicial administration. See also note 64 infra.
. To avoid confusion, we emphasize: (1) that Congress has never indicated any intention to detain all exhibitionists indefinitely as “dangerous”; (2) that we did not question congressional power to punish acts of exhibitionism; (3) that we did not in any way restrict the kinds of ends and interests that Congress may, by appropriate means, seek to promote; (4) that we did not say that harm to only a small segment of the population could not be sufficient to support a commitment under the Sexual Psychopath Act, so long as that harm was, in the language of of the Act, likely to occur; (5) that we did not exclude any lasting harm of any sort to children, women, or anyone else from the category of “substantial injury” which may justify commitment under the Act, provided there is evidence that such harm is likely to occur; and (6) that far from relying on any psychiatric theories of our own, we simply accepted in Millard the testimony of the psychiatrists concerning the likely consequences of Mr. Millard’s exhibitionism, as in this case we must rely on the expert testimony concerning Mr. Cross. Nor is this list exhaustive.
. The dissent argues that the question of dangerousness is foreclosed, because (1) appellant did not contest the question at the initial hearing, which was conducted without benefit of the construction of that term which we reached in Millard-, therefore (2) he “necessarily” must be taken to have conceded that he was “likely to injure himself or others” under the quite different phraseology of the Hospitalization of the Mentally 111 Act; and that therefore (3) he must be taken to have conceded that he is “dangerous” as that term was construed in Millard. In other words, by two simple non sequiturs, appellant’s concession of dangerousness made before Millard is converted into a concession of dangerousness under the limiting construction there reached. Compare Cole v. Arkansas, 353 U.S. 196, 201, 68 S.Ct. 514, 92 L.Ed. 644 (1948).
. E. g., King v. United States, 125 U.S. App.D.C, 318, 332-333, 372 F.2d 383, 397-398 (1967) ; Afro-American Pub. Co. v. Jaffe, 125 U.S.App.D.C. 70, 82-84, *1105366 F.2d 649, 661-663 (1966) (en banc) ; Coleman v. United States, 123 U.S.App. D.C. 103, 106-113, 357 F.2d 563, 566-573 (1965) (en banc) ; see Armstrong v. Board of Education, 323 F.2d 333, 338-339 (5th Cir. 1963); Davis v. Board of School Com’rs, 322 F.2d 356, 358-359 (5th Cir. 1963) (per curiam).
. This is a habeas corpus proceeding, and thus particularly inappropriate for any delay. See 28 U.S.C. § 2243 (1964). There is consequently no justification for the dissent’s suggestion that we express no opinion on the construction of the statute that must be applied if the court on remand determines that appellant is not “mentally ill.” Were we to follow the course suggested, no matter what construction was applied by the court on . remand the case would have to return here for final decision. Since appellant’s doctors have testified that continued confinement may aggravate his condition, delay should be minimized.
Nor is there any warrant for a suggestion that our resolution of the issue is mere dictum. The distinction between holding and dictum is not whether the point in question had to be decided in order that the court’s mandate could issue. The distinction turns on whether the court, in stating its opinion on the point, believed it necessary to decide the question or was simply using it by way of illustration of the case at hand. Cohens v. Virginia, 19 U.S. (6 Wheat.) 264, 399-400 (1821) ; see United States v. Poller, 43 E.2d 911, 912 (2d Cir. 1930) (L. Hand, A. Hand, & Swan, JJ.). See also Woods v. Interstate Realty Co., 337 U.S. 535, 537, 69 S.Ct. 1235, 93 L.Ed. 1524 (1949) ; Railroad Companies v. Schutte, 103 U.S. 118, 143, 26 L.Ed. 327 (1880). We believe it necessary in this case to provide a construction of the statute in order to cover issues within the scope of the remand hearing.
. Our decision in Millard turned upon the record, which showed that no harm was likely to flow from Mr. Millard’s expected future conduct. Millard, supra note 10, 132 U.S.App.D.C. at 155-160, 406 F.2d at 973-978. In° taking issue with Millard, the dissent speaks of “traumatic injury” and “psychic trauma” to bystanders without any discussion of the likelihood that such trauma will in fact occur. See p. 1108 infra. Although it may be true that in certain circumstances frequent exhibitionism is necessarily “dangerous” within the meaning of the Sexual Psychopath Act, there is certainly no evidence on the record either here or in Millard to support such a conclusion; we consequently left the issue open on remand. It would therefore appear that the dissent is either (1) relying upon unmentioned extra-record psychiatric theories in order to conclude the exhibitionism is dangerous per se, or else (2) giving no effect to the word “likely” in the statute. See p. 1106 infra.
. Mr. Justice Frankfurter, speaking for a unanimous Court in United States v. Rumely, 345 U.S. 41, 47, 73 S.Ct. 543, 97 L.Ed. 770 (1953).
. E. g., Grenada County Supervisors v. Brogden, 112 U.S. 261, 268-269, 5 S.Ct. 125, 28 L.Ed. 704 (1884) ; Knights Templars’ & Masons’ Life Indem. Co. v. Jarman, 187 U.S. 197, 204-205, 23 S.Ct. 108, 47 L.Ed. 139 (1902) ; Harriman v. ICC, 211 U.S. 407, 422, 29 S.Ct. 115, 53 L.Ed. 253 (1908); Crowell v. Benson, 285 U.S. 22, 62, 52 S.Ct. 285, 76 L.Ed. 598 (1932) ; Ashwander v. TVA, 297 U.S. 288, 348, 56 S.Ct. 466, 80 L.Ed. 688 (1936) (Brandeis, J., concurring); United States v. Rumely, 345 U.S. 41, 45-48, 73 S.Ct. 543, 97 L.Ed. 770 (1953) ; Watts v. United States, 394 U.S. 705, 707-708, 89 S.Ct. 1399, 22 L.Ed.2d 664 (1969). See also Blodgett v. Holden, 275 U.S. 142, 148, 48 S.Ct. 105, 72 L. Ed. 206 (1927) (Holmes, Brandeis, Sanford & Stone, JJ., concurring) ; Schneider v. Smith, 390 U.S. 17, 27, 88 *1106S.Ct. 682, 19 L.Ed.2d 799 (1968) ; Haynes v. United States, 390 U.S. 85, 91-92, 88 S.Ct. 722, 19 L.Ed.2d 923 (1968).
. See Millard, supra note 10, 132 U.S. App.D.C. at 160, 406 F.2d at 978; note 35 supra.
. 22 D.C.Code § 3503(1) (1967) (emphasis added). See notes 28-34 supra and accompanying text.
. See p. 1099 supra.
. See p. 1110 infra (“devastating impact” of “overt public misconduct”).
. See pp. 1109-1110 infra:
Likewise, in the present case, there is no assertion by Appellant, or determination by the majority, that it was irrational for Congress to conclude that exposure to Appellant’s conduct may cause harm.
(emphasis added). See also p. 1109 (“No reasons * * * advanced * * * which justify substitution by the majority of its own theories * * * for those of the legislative branch.”), 1109-1110 (courts should not “displace the policy judgments of the legislature”).
. H.R.Rep. No. 1787, 80th Cong., 2d Sess. (1948).
. S.Rep. No. 1377, 80th Cong., 2d Sess. (1948).
. The hill is mentioned at 94 Cong.Rec. 3884, 4802, 4885^887, 4927, 6267, 6310, 6593, 6747, 6851, 7114, 7981 (1948). The only discussion is at 4885-4887.
. See Dershowitz, On “Preventive Detention,” New York Review of Books, Mar. 13, 1969, at 22.