TIMBERS, Circuit Judge:

Appellant Warren, who is confined involuntarily at Whiting Forensic Institute — a Connecticut state mental institution — after having been tried by a jury for first degree murder and having been found not guilty by reason of insanity, filed a petition for a writ of habeas corpus in the District of Connecticut challenging the constitutionality of his confinement. The instant appeal is from a judgment entered on an opinion by M. Joseph Blumenfeld, District Judge, denying that petition. 472 F.Supp. 1061. For reasons other than those relied upon by the district court, we affirm.

On July 5, 1971 appellant was arrested for fatally shooting his neighbor. In November of that year he was indicted for first degree murder. Following lengthy pretrial proceedings, which included a psychiatric examination and competency hearing, appellant was brought to trial in early 1974 before Judge Saden and a jury in the Superior Court for New Haven County. On *927 February 27, 1974 the jury returned a verdict of not guilty by reason of insanity. 1

After his acquittal, appellant was ordered confined temporarily at the Connecticut Valley Hospital, in accordance with Conn. Gen.Stat. § 53a — 47(a) (Rev.1958, Supp.1975), so that a determination could be made whether he was “mentally ill to the extent that his release would constitute a danger to himself or others.” 2 An initial commitment hearing was held in the Superior Court before Judge Saden in May 1974. At this hearing, evidence was presented that appellant was being administered a major anti-psychotic drug, mellaril. The effect of this drug was to reduce the patient’s anger and hostility and give him “a more realistic grasp on reality”. According to testimony by Dr. Patrick Lee — a unit chief at the Connecticut Valley Hospital and the only witness at the hearing — appellant had improved significantly as a result of taking this medication and showed a marked decrease in his tendency for violence. It was Dr. Lee’s opinion that, while taking this medication, appellant was “not a danger to himself or others in or outside the hospital”.

Dr. Lee, however, emphasized that appellant’s continuing use of the medication was of critical importance. He testified that appellant’s condition would revert back to its previous state — marked by hostility, failure to get along with others, and outbursts of rage — if he were to stop taking the medication. Citing recent studies by the Veterans Administration which showed that almost half of the readmissions to its hospitals were due to the failure of patients to continue taking such medication after their release, Dr. Lee concluded that appellant’s “release at this time would constitute a danger to himself or others.” Dr. Lee further testified that in his view appellant would benefit from a further period of hospitalization.

At the conclusion of this hearing, Judge Saden found that appellant was mentally ill to the extent that his release would constitute a danger to himself or others. In support of this finding, the court stated that appellant’s “condition is not eliminated by medication; it is only controlled .... [MJedication does not affect the cause, only the symptoms.” The court held that “[tjhere is a strong possibility that the defendant, if released, might not continue use of the anti-psychotic medication without which he is a danger to himself or others.” The court on May 24, 1974 therefore ordered that appellant be confined at the Connecticut Valley Hospital. 3

*928 Appellant appealed to the Connecticut Supreme Court from this commitment order, challenging both the sufficiency of the evidence presented at the commitment hearing and the constitutionality of the “preponderance of the evidence” standard applied at the hearing pursuant to § 53a-47(a)(4). The Connecticut Supreme Court rejected appellant’s contention that the state should have been required to prove his dangerousness by more than a mere preponderance of the evidence, as the state is required to do in order to confine persons in ordinary civil commitment proceedings. State v. Warren, supra, 169 Conn, at 215, 363 A.2d at 96. Citing “significant differences between the civil commitment situation and the commitment of an accused found not guilty because of insanity,” id., the court upheld the constitutionality of the less stringent standard applied in this case. The court further held that the evidence at appellant’s hearing was sufficient to satisfy this standard of proof. Id. at 213, 363 A.2d at 94-95.

In December 1974 a second hearing — this time regarding appellant’s right to be released from confinement pursuant to § 53a-47(c) 4 — was held in the Superior Court before Judge Saden. Dr. Lee and Dr. Burk-hard Voelkening, a psychiatrist at the Connecticut Valley Hospital, testified on the basis of reports submitted to the court stating that appellant was taking anti-psychotic tranquilizing medication and exhibited no symptoms of overt psychosis or thought disorder. Both reports concluded that, if appellant continued to take his medication and received outpatient psychiatric care, his release would not constitute a danger.

At this hearing, however, Dr. Lee testified that, because appellant’s medication treated only his symptoms and not his underlying mental illness, he would become psychotic in about six weeks if he were to cease taking the medication. Dr. Lee further testified that one of the reasons appellant presently was not dangerous was because he was taking medication in the structured, supervised setting of the mental hospital. According to Dr. Lee, appellant needed “the control setting of a staff handling him”; outside the hospital and without such controls, in Dr. Lee’s opinion, the chances were forty to fifty percent that appellant would stop taking the medication and would revert back to his psychotic, potentially violent state.

Dr. Voelkening was in substantial accord with Dr. Lee’s analysis. He attributed appellant’s improved condition to the medication and the structured hospital setting where appellant was living; outside the hospital — even if he were getting psychiatric treatment and medication on an outpatient basis — appellant might still become dangerous if faced with stress and frustration; and appellant’s homosexuality or future involvement with alcohol were cited as possible sources of such stress. Dr. Voelkening concluded that “[gjiven the right combination of circumstances of stress and frustration, and not taking medication and *929 having no structure,” appellant could become dangerous again.

Judge Saden held on December 19, 1974, after the second hearing, that appellant was not entitled to be released from confinement. He found that the state had proven by a preponderance of the evidence that appellant remained mentally ill to the extent that his release would constitute a danger to himself or others.

Ten months later, a second release hearing was held in the Superior Court before Judge Irving Levine. Dr. Mehadin Arafeh, Superintendent of the Connecticut Valley Hospital, Dr. Michael Sheard, Associate Professor of Psychiatry at Yale University, Dr. Voelkening, and Dr. Lee each submitted reports and testified at the hearing. The substance of their reports and testimony was that appellant was then taking anti-psychotic medication and at the time was free from overt psychosis. Appellant’s improvement was attributed both to the medication and the stable, secure hospital environment. Outside the controlled hospital setting, there was an increased risk that stress and anxiety would cause appellant to stop taking his medication. 5 In such event, appellant was likely to exhibit violent behavior as he had in the past.

After this second release hearing, Judge Levine on November 4, 1975 held that the state again had established by a preponderance of the evidence that appellant was mentally ill to the extent that his release would constitute a danger to himself or others. The court accordingly ordered that appellant’s confinement be continued.

In December 1975 appellant was transferred from the Connecticut Valley Hospital to Whiting Forensic Institute, where he is presently confined.

In May 1977 he petitioned the district court for a writ of habeas corpus, claiming that his confinement violated due process because an unconstitutional standard of proof was adhered to by the state courts at each of the three hearings. The district court held 6 that the Connecticut statute *930 governing appellant’s commitment, § 53a-47, was constitutional on its face, concluding that the more stringent standard of proof required for involuntary civil commitment proceedings by the Supreme Court in Addington v. Texas, 441 U.S. 418 (1979), was not required for the commitment of insanity acquittees.

The district court, however, also held that the evidence presented at the hearings was not sufficient to support a finding that appellant constituted a danger to himself or others. The district court found that the state courts, by relying on a presumption of continuing dangerousness stemming from appellant’s criminal act, in effect had shifted the burden of proof on the issue of dangerousness to appellant. Such a shift, according to the district court, violated the Connecticut statutory requirement that the state bear the burden of proving appellant’s dangerousness by a preponderance of the evidence. The district court nevertheless concluded that no constitutional right had been violated, since a state may require an insanity acquittee to bear the burden of proving that he no longer is dangerous due to a mental illness. The district court therefore denied the petition for a writ of habeas corpus.

II.

Two issues of constitutional proportion are before us on this appeal: (1) whether the Connecticut statute which establishes the burden of proof in proceedings for the commitment of persons found not guilty by reason of insanity is constitutional on its face; and (2) whether that statute is constitutional as applied to appellant.

We turn first to the issue of the facial constitutionality of the commitment statute. The statute in question, § 53a-47, provides that in order to confine an insanity acquittee to a mental institution, or to prevent his release after confinement, the state must prove by a preponderance of the evidence that he is mentally ill to the extent that his release would constitute a danger to himself or others.

Our determination of the constitutionality of this statute begins with a consideration of Addington v. Texas, 441 U.S. 418 (1979). In Addington, the Court addressed the question of what standard of proof is required by the due process clause in civil proceedings to commit individuals involuntarily to a state mental institution. The Court concluded that the standard of proof in such proceedings must be greater than the preponderance of the evidence standard applicable to most civil cases. Rather, the Court held due process requires a clear and convincing standard of proof in such proceedings. Id. at 431-33.

Arguing that an insanity acquittee is no different than any other person who faces involuntary commitment, appellant contends that the holding in Addington controls the instant case and that the preponderance of the evidence standard provided by § 53a-47 to commit insanity acquittees should be held unconstitutional on its face. We disagree.

Courts repeatedly have recognized that relevant differences between persons acquitted by reason of insanity and other persons facing civil commitment can justify different treatment of these two groups in commitment proceedings. Baxstrom v. Herold, 383 U.S. 107, 111 (1966); Powell v. Florida, 579 F.2d 324, 332-34 (5 Cir. 1978); United States v. Brown, 478 F.2d 606 (D.C. Cir. 1973); Bolton v. Harris, 395 F.2d 642, 650-52 (D.C.Cir. 1968). The precise question before us, therefore, is whether such differences can justify Connecticut’s commitment of insanity acquittees on a lesser standard of proof than that applied in other civil commitment proceedings.

The function of a standard of proof is to “instruct the factfinder concerning the degree of confidence our society thinks he should have in the correctness of factual conclusions for a particular type of adjudication.” Addington v. Texas, supra, 441 U.S. at 423; accord, In re Winship, 397 U.S. 358, 370 (1970) (Harlan, J., concurring). The standard serves to allocate the risk of error between society and the individual facing commitment. As a result, in con *931 sidering what standard should govern in a commitment proceeding, courts must weigh “the individual’s interest in not being involuntarily confined” against “the state’s interest in committing the emotionally disturbed under a particular standard of proof.” Addington v. Texas, supra, 441 U.S. at 425.

The obvious difference between insanity acquittees and other persons facing commitment is the fact that the former have been found, beyond a reasonable doubt, to have committed a criminal act. 7 Insanity acquit-tees thus have “proved” themselves a danger to society at one time. Non-acquittees, in contrast, have not been found by any factfinder to have harmed society as a result of their mental illness. 8 This difference, we believe, gives rise to considerations which justify a lesser standard of proof to commit insanity acquittees than to commit other persons.

One of the primary objects of any standard of proof is to provide the individual with appropriate protection from the harm that may result from an erroneous decision. The more harm which a defendant will suffer as the result of an erroneous decision, the more stringent the burden of proof which society should shoulder to prove its case. Ordinarily, if a person is wrongly committed to a mental institution, he may suffer great harm at the hands of society. He is wrongly deprived of his liberty when he poses no danger to society. Furthermore, he is greatly stigmatized. The cost of an erroneous decision to commit a person therefore is great indeed. Unfortunately, since inherently there is so much uncertainty involved in predicting the dangerousness of a mentally ill person, there is a real possibility that “a factfinder might decide to commit an individual based solely on a few isolated instances of unusual conduct.” Addington v. Texas, supra, 441 U.S. at 427. Therefore, by imposing a more stringent burden of proof, society has sought to “impress the factfinder with the importance of the decision”, and thus reduce the chance that a harmless individual will be confined erroneously to an institution. Id.

The same considerations do not come into play when the person facing commitment is an insanity acquittee. He need not be overly protected against the possibility that the factfinder will commit him based on “a few isolated instances of unusual conduct.” The insanity acquittee already has been found to have committed a criminal act. Furthermore, if an insanity acquittee is committed because of an erroneous determination that he is mentally ill, then the odds are high that he may have been found not guilty on insanity grounds because of a similar erroneous determination that he is not sane. While the acquittee therefore may be deprived erroneously of his liberty in the commitment process, the liberty he loses is likely to be liberty which society mistakenly had permitted him to retain in the criminal process. Concomitantly, while society derives no benefit from erroneously confining ordinary persons who are not in fact mentally ill and dangerous, the erroneous confinement of an insanity acquittee who in fact was not mentally ill at the time of his crime indirectly benefits society by keeping a “sane” criminal off the streets.

Next to the actual deprivation of liberty, the greatest harm to a person erroneously committed to a mental institution is the stigma attached to the commitment. Addington v. Texas, supra, 441 U.S. at 426. The harm that such stigma causes an individual provides another cogent reason for “weighting” the standard of proof scales in favor of the ordinary individual faced with civil commitment. It is hard to view the insanity acquittee with the same solicitude. He already has been found to have committed an unlawful act — with all the attendant stigmatization — and has escaped punishment solely because he convinced the fact-finder in the criminal proceeding that there was merit to his claim of insanity. Any *932 stigma resulting from the label “mentally ill and dangerous” certainly attached at the time the accused was found not guilty by reason of insanity. Additional stigma which might result from subsequent commitment to a mental hospital must be regarded as minimal, if any.

Finally, we cannot ignore the danger of “calculated abuse of the insanity defense”. United States v. Brown, supra, 478 F.2d at 611. We are reluctant to provide criminal defendants with a loophole at society’s expense by enabling those who have committed criminal acts first to escape criminal punishment by pleading insanity and thén to escape confinement completely if the government fails to prove by clear and convincing evidence that the defendant will continue to be prone to the very same abnormalities that he sought to establish in his past behavior. While imposing the less stringent burden of proof on the state at commitment proceedings for insanity acquittees may not foreclose strategic abuse of insanity pleas by defendants, the heightened risk to a sane defendant that he subsequently will be confined to a mental hospital may reduce his willingness to chance a disingenuous insanity defense. Even if it may be said that such deterrence is slight, we do not believe that the Constitution forecloses the states from making use of it.

We therefore conclude that relevant distinctions can be made between insanity acquittees and other persons facing involuntary commitment. While the due process clause ordinarily requires the strict “clear and convincing” standard of proof in civil commitment proceedings, it does not prohibit the states from providing for the lesser “preponderance of the evidence” standard in commitment hearings for that select group of persons who previously have been tried for criminal acts and found not guilty by reason of insanity. Accord, United States v. Brown, supra. Not only is use of the preponderance standard not precluded in this case by the Court’s holding in Addington; it is fully supported by the Court’s logic in Lynch v. Overholser, 369 U.S. 705, 715 (1962):

“The criminal defendant who chooses to claim that he was mentally irresponsible when his offense was committed is in quite a different position. It is true that he may avoid the ordinary criminal penalty merely by submitting enough evidence of an abnormal mental condition to raise a reasonable doubt of his responsibility at the time of committing the offense. Congress might have thought, however, that having successfully claimed insanity to avoid punishment, the accused should then bear the burden of proving that he is no longer subject to the same mental abnormality which produced his criminal acts. Alternatively, Congress might have considered it appropriate to provide compulsory commitment for those who successfully invoke an insanity defense in order to discourage false pleas of insanity. We need go no further here than to say that such differentiating considerations are pertinent to ascertaining the intended reach of this statutory provision.”

Appellant in the instant case was tried and found by a jury to have killed another human being. He escaped punishment solely because he successfully pled that he was insane. He cannot fairly be said to occupy the same position as, for example, a man who has never harmed anyone in his life but who faces civil commitment because he may have dangerous propensities due to a mental disease. We do not believe the Constitution compels the states to shoulder the same burden of proof in order to commit persons in such different situations. 9

*933 III.

Having upheld the facial constitutionality of the Connecticut statute providing for commitment of an insanity acquit-tee when the state proves its case by a preponderance of the evidence, we turn now to the question whether the statute was constitutionally applied to appellant.

Appellant contends that there was not sufficient evidence at either his commitment or release hearings to satisfy the state’s burden of proving him mentally ill and dangerous. Specifically, he argues that there was no evidence that he was presently dangerous and therefore that the state courts unconstitutionally had shifted the burden to him of proving that he was not dangerous. The district court agreed that the state courts had shifted the burden of proof to appellant — in violation of state law — but went on to hold that such a shift in burden of proof did not violate the Constitution.

We agree with the result reached by the district court that the Connecticut statute in question is not unconstitutional as applied to appellant, but we do so for reasons other than those relied upon by the district court. We hold that the state did satisfy its burden of proving appellant’s dangerousness by a preponderance of the evidence and therefore that it violated neither state law nor the Constitution. 10

Under Connecticut law, appellant initially could be committed to a mental hospital, *934 and then subsequently could be prevented from being released from such confinement, only if the state proved by a preponderance of the evidence that he was “mentally ill to the extent that his release would constitute a danger to himself or others.” §§ 53a-47(a) and (c). Here the essence of appellant’s claim is not that the state failed to prove he was mentally ill. As the district court noted, the evidence before the state courts “unequivocally established that [appellant] had an underlying mental illness at the time of the commitment hearings.” 472 F.Supp. at 1070 n.16. Appellant argues, however, that the state did not sustain its constitutional burden of proving him dangerous. We disagree.

Foremost among the evidence before the state courts was the fact that appellant already had committed the most extreme form of violence: the killing of another human being. His acquittal on grounds of insanity constituted proof beyond a reasonable doubt that he once had been a danger to society. While the insanity acquittal did not constitute dispositive proof of present dangerousness, it was substantial evidence that appellant remained a danger to society. State v. Krol, 68 N.J. 236, 261, 344 A.2d 289, 302 (1975); see Powell v. Florida, supra, 579 F.2d at 333; United States v. Ecker, 543 F.2d 178, 196-99 (D.C.Cir. 1976), cert. denied, 429 U.S. 1063 (1977); Cameron v. Mullen, 387 F.2d 193, 201 (D.C.Cir. 1967). One means of demonstrating that an individual poses a “danger to himself or others” is to show that he has “a propensity to commit criminal acts”. E. g., Reynolds v. Sheldon, 404 F.Supp. 1004, 1009 (N.D.Tex.1975). Certainly past commission of a criminal act must be accorded considerable weight in determining the existence of a “propensity to commit criminal acts”.

Appellant argues that, since he killed his neighbor approximately three years before his initial commitment, that act of violence should be discounted. Instead, appellant contends, the state should have been required to prove his dangerousness with evidence of more recent acts of violence. We are unpersuaded by this argument. We do not consider appellant’s act of homicide so remote in time to be devoid of evidentiary value. Furthermore, the lack of evidence that appellant has engaged in more recent violent acts or threats must be viewed in light of the fact that he has been in custody ever since he killed his neighbor. It obviously is more difficult for a man who is in jail or confined to a mental hospital to translate his violent propensities into actual conduct than it is for a man who is free to act as he pleases.

Beyond this, at the time of the initial commitment proceedings and the subsequent release hearings, evidence was adduced that appellant was suffering from the same mental illness — paranoid schizophrenia — from which he suffered at the time he killed his neighbor. The symptoms were being controlled by an anti-psychotic drug, but the underlying disease remained unaltered. Even with the medication, appellant was guarded and suspicious. If he were to stop taking the medication, he would revert back to his uncontrolled, psychotic state — in which he once had killed a man. A medical expert, Dr. Lee, testified at the initial commitment proceedings that if appellant were released from the hospital and did not take his medication, he would be a danger to himself or others. This conclusion was supported not only by appellant’s single criminal act, but by his long history of impulsivity and poor control of aggressive behavior — a history which dated back to his childhood. Far from being of the opinion that appellant’s condition was cured, or even in remission, Dr. Lee testified that appellant “would benefit from a further period of hospitalization.” This opinion was based on the expert’s “examination of all available records and personal interviews with” appellant.

Since appellant’s mental illness was uncured, it was of critical importance that he regularly take his medication, without fail. A lapse could have deadly consequences. Yet there was considerable evidence at the hearing that appellant could not be trusted to take his medication indefinitely, as was required. Many mentally ill patients released on medication fail to take required *935 medication once they have left confinement. The court heard testimony that recent Veterans Administration studies revealed that “close to fifty percent of the readmissions to those hospitals were due to failure of patients to realize the need to continue taking medication.” While Dr. Lee acknowledged that it was possible for some patients to treat themselves adequately on an outpatient basis, he did not believe that such a possibility eliminated the danger that appellant would pose to society if released.

On the basis of this evidence, we hold that the Connecticut court was justified in concluding that the state had shown by a preponderance of the evidence that appellant would be dangerous if released. Contrary to the conclusion reached by the district court, we do not. believe that the state court shifted to appellant the burden of proving that he was not dangerous. The burden remained where it originated — on the state. This burden was sustained by the evidence which the state adduced at the hearing, namely, that appellant still was mentally ill, that he had acted violently in the past due to this illness, and that he posed a substantial problem of lapsing into his dangerous psychotic state in the future if released.

We further conclude that the state sustained its burden of proof at the subsequent release hearings as well. If anything, the evidence at the release hearings was even stronger in support of the finding that appellant would be dangerous if released.

At the first release hearing, the court heard testimony that without medication appellant would become psychotic in as little as six weeks. He had demonstrated that he was dangerous when psychotic. Furthermore, his improvement in confinement was attributed not only to the medication, but also to the highly structured environment at the mental hospitals. Without this controlled setting, there was a forty to fifty percent chance that he would fail to take his medication. Stress and anxiety could cause him to lapse back into his psychotic condition. There would be greatly increased exposure to such stress outside the shelter of the Connecticut Valley Hospital or Whiting Forensic Institute.

In addition, evidence was adduced at the first release hearing which indicated that, even while on medication, appellant’s judgment and attitude remained impaired. He was described as emotionally cold and lacking in remorse. He also remained suspicious. These remaining mental impairments cast doubt on the likelihood that he would realize the need to take his medication indefinitely. As one expert testified, “[gjiven the right combination of circumstances of stress and frustration, and not taking his medication and having no structure,” appellant would pose a danger to himself or others.

Similarly, at the second release hearing evidence showed that, while medication controlled appellant’s overt psychosis, it did not eliminate his anxiety and tenseness. Stress caused by his homosexual feelings and exposure to the trials of life outside the institution were cited as factors which could lead him to stop taking the needed medication if released. Without such medication, it was likely that appellant would quickly become psychotic and potentially violent. Again, appellant’s recent improvement was linked not only to his medication but to the structure, staff treatment, and security at the mental institution.

We conclude that the evidence at both release hearings was sufficient to sustain the state’s burden of proof that appellant’s release would constitute a danger to himself or others. The evidence established that appellant was mentally ill and that he could act violently if his psychosis were not controlled. In a secure, structured environment with regularly administered medication, his overt symptoms and violent propensities were controlled; but the illness itself remained, along with impaired judgment and suspiciousness. Appellant had a history of aggression leading up to his killing of a neighbor. If the symptoms of this illness were not kept under strict control, he could be dangerous. Without the structure and supervision at Whiting Forensic Insti *936 tute, there was a substantial risk that these symptoms would not be properly controlled. Regardless of whether the term “danger to himself or others” is construed to mean a propensity to commit criminal acts, Reynolds v. Sheldon, supra, or simply a reasonable foreseeability of danger to the patient or the community, e. g., Rosenfield v. Overholser, 262 F.2d 34, 35 (D.C.Cir. 1958), we conclude that at the three hearings the state proved by a preponderance of the evidence that appellant’s release would constitute such a danger.

Accordingly, the Connecticut statute governing these hearings, which we have held to be constitutional on its face, we also hold is constitutional as applied to appellant.

IV.

To summarize, we hold that:

(1) Since there are relevant differences between insanity acquittees and other persons facing civil commitment, the “clear and convincing” standard of proof ordinarily required by the Fourteenth Amendment in involuntary commitment proceedings is not required in proceedings to commit insanity acquittees. Conn.Gen. Stat. § 53a-47, which authorizes commitment and non-release of insanity acquit-tees if the state proves their mental illness and dangerousness by a “preponderance of the evidence”, is constitutional on its face.

(2) In the instant case, § 53a-47 was not unconstitutionally applied in appellant’s commitment and release proceedings. The Connecticut courts correctly held that the state had sustained the burden of proof imposed on it by § 53a-47. The Connecticut courts did not require appellant to bear'the burden of proving that he was not dangerous.

Therefore, while we do not reach our decision by the same route as the district court did, we do affirm its judgment denying appellant’s petition for a writ of habeas corpus.

Affirmed.

Notes

. Under Connecticut law, a jury verdict of not guilty by reason of insanity means that “the accused has been found beyond a reasonable doubt to have committed a criminal act, and, in addition, it has been found that at the time of the commission of that act there was at least reasonable doubt as to his sanity.” State v. Warren, 169 Conn. 207, 215, 363 A.2d 91, 96 (1975).

. Conn.Gen.Stat. § 53a-47(a) (Rev.1958, Supp. 1975) in relevant part provides:

“(1) When any person charged with an offense is acquitted on the grounds of mental disease or defect, the court shall order such person to be temporarily confined in any of the state hospitals for mental illness for a reasonable time, not to exceed ninety days, for an examination to determine his mental condition .... (2) The person to be examined shall be informed that, in addition to the examination provided for in subdivision (1), he has a right to be examined during such confinement by a psychiatrist of his own choice. (3) Within sixty days of the confinement pursuant to subdivision (1), the superintendent of such hospital and the retained psychiatrist, if any, shall file reports with the court setting forth their findings and conclusions as to whether such person is mentally ill to the extent that his release would constitute a danger to himself or others. Copies of such reports shall be delivered to the state’s attorney or prosecutor and to counsel for such person. (4) Upon receipt of such reports, the court shall promptly schedule a hearing. If the court determines that the preponderance of the evidence at the hearing establishes that such person is mentally ill to the extent that his release would constitute a danger to himself or others, the court shall confine such person in a suitable hospital or other treatment facility.”

. The term of confinement was to continue until appellant ceased to be mentally ill to the extent that his release would constitute a danger to himself or others. This term, however, was not to last longer than twenty-five years — a term of imprisonment which could have been imposed on appellant for the crime he committed.

. Conn.Gen.Stat. § 53a-47(c) (Rev. 1958, Supp. 1975) provides:

“(1) Upon certification by the superintendent of the hospital or institution that, in his opinion, such person is no longer mentally ill to the extent that his release would constitute a danger to himself or others, the court may order the release of the person confined at the expiration of thirty days from the time such certificate is filed. (2) At the time such certificate is filed with the court, a copy shall be furnished to the state’s attorney or prosecutor who may request a hearing as to whether such person should be released. At such hearing, evidence of mental condition may be submitted. The confined person shall be released unless the state establishes by a preponderance of the evidence that such person is, at the time of hearing, mentally ill to the extent that his release would constitute a danger to himself or others. (3) The superintendent shall, during such confinement, submit to the court at least every six months a written report with respect to the mental condition of such person. Copies of such report shall be furnished to the state’s attorney or prosecutor and counsel for the confined person. The court, upon its own motion or at the request of the parties, may at any time hold a hearing to determine whether such person should be released prior to the expiration of the maximum period, in accordance with the standards set forth in subdivision (1), provided such a hearing shall be held at least every five years.”

. At the time of the second release hearing, appellant was taking tofranil to combat depression; tremin to prevent a physical reaction known as Parkinsonian syndrome; and proloxin to control his paranoid schizophrenic syndromes.

. Before reaching the merits, the district court rejected the state’s argument that appellant’s petition should have been dismissed for failure to exhaust state remedies. The exhaustion claim was two-fold. First, since appellant appealed only from the Superior Court order regarding his initial commitment proceeding and did not appeal from the separate orders involving his release hearings, the state argued that appellant was precluded from raising in the federal court any issues arising from the later proceedings. Second, the state argued that appellant did not fairly present to the state courts the claim that he was denied due process by the state courts’ shift to him of the burden of proving non-dangerousness.

With regard to the first claim, the district court concluded that the issues at each of the three hearings were the same and that the facts underlying appellant’s constitutional claims also were substantially the same at each hearing. The court therefore held that, since the Connecticut Supreme Court had addressed these issues in the context of the first hearing, “to require petitioner to relitigate on appeal the same claim after each periodic hearing would not serve the purposes of the exhaustion requirement.” 472 F.Supp. 1061, 1068. The district court also rejected the state’s claim that appellant never had raised his constitutional claims at the state level. The court held that, while “the due process claims petitioner raises in this court are more elaborately expressed than those presented to the Connecticut Supreme Court, that court was nonetheless presented with an opportunity to rule that the Superior Court’s application of Conn.Gen.Stat. § 53a^l7 violated petitioner’s due process rights.” Id. at 1067.

We agree with the district court that appellant had sufficiently exhausted his state remedies to allow him to raise his constitutional claims in the federal courts. We can discern no possible benefit in relegating appellant to the state courts when substantially the same legal issues before us on appeal already have been ruled upon at the state level and the record of the state court proceedings present those issues with clarity. To hold otherwise, in view of the fact that the state proceedings here involved began some six years ago, “might well invite the reproach that it is the [petitioner] rather than the state remedy that is being exhausted.” United States v. La Vallee, 306 F.2d 199, 203 (2 Cir. 1962) (Friendly, J., concurring). See Lunz v. Henderson, 533 F.2d 1322, 1328 n.8 (2 Cir.), cert. denied, 429 U.S. 849 (1976).

. Note 1, supra.

. Indeed many persons facing involuntary commitment are not even accused of having acted violently. They may only stand accused of having violent propensities which might result in future harm to society.

. Appellant argues that the “clear and convincing” standard is compelled in this case not only by the logic of the Supreme Court’s decision in Addington, but also by our own decision in Hollis v. Smith, 571 F.2d 685 (2 Cir. 1978).

In Hollis, the defendant, who had pled guilty to assault with intent to commit rape, was sentenced under a special state sex offender provision to an indeterminate term of one day to life as a result of the determination that he was a “danger to society or [was] capable of being benefited by the confinement envisaged under the statutory scheme.” Without this special sentencing provision, the defendant would have been subject to a maximum five year term for assault.

*933 While we rejected the defendant’s contention that he could not receive the stiffer sentence unless his dangerousness was proven “beyond a reasonable doubt”, we did hold that the clear and convincing standard of proof was required. “Nothing less”, we held, “would accord due process to a man, otherwise sentenceable for a maximum of five years, who instead was being sentenced to an indeterminate period of one day to life.” Id. at 695-96.

We do not think that Hollis compels a holding that the clear and convincing standard is required in this case. First, the deprivation imposed upon the defendant in Hollis was far greater than in the instant case. The sentence in Hollis was tantamount to greater punishment than that which society had authorized as punishment for the defendant’s criminal act. In this regard, the defendant in Hollis was comparable to a criminal who had been committed to a mental institution at the end of his prison sentence. Cf. Baxstrom v. Herold, 383 U.S. 107 (1966). In both cases, society must look for protection to the penal term which can be imposed for the crime committed. Any additional protection to society by means of increased detention must be sought on the same basis as commitment of non-criminals with dangerous propensities. See United States v. Brown, supra, 478 F.2d at 612.

In the instant case the court specified, in accordance with § 53a--47(b), that appellant’s confinement in the mental institution could not exceed twenty-five years, a term of confinement which could have been imposed if appellant had been convicted of murder rather than acquitted on insanity grounds. The commitment in this case, therefore, cannot be regarded as subjecting an individual to a deprivation of freedom beyond that ordinarily authorized for his criminal conduct. This is in sharp contrast to the commitment in Hollis.

Furthermore, most of our reasons for distinguishing Addington also are applicable to our determination that Hollis is not controlling here. The risk of error, for example, is shouldered more fairly by an insanity acquittee than by the convicted offender in Hollis. While an erroneously committed insanity acquittee may have benefited from a similar error which led to his criminal acquittal, the defendant in Hollis did not accrue any past “benefit” to offset an erroneous determination that he was a dangerous sex offender. Society also has a greater stake in committing an insanity acquittee than it does in winning lengthier sentences for Hollis defendants. If an insane person is acquitted of criminal charges and then wrongly escapes confinement in a mental institution, he is free to wreak havoc on society again. The Hollis defendant, however, still must serve a prison sentence for his crime even if the state fails to sustain its burden of proving that he is dangerous or can benefit from lengthier confinement.

While the stigma arising from erroneous court decisions is roughly equivalent for both the insanity acquittee and the Hollis defendant, we hold that the differences between these two potential subjects of commitment proceedings far outweigh the similarities and therefore justify differing standards of proof in commitment proceedings. ■

. We do not reach the question whether a state constitutionally may place the burden of proof on the insanity acquittee in initial commitment hearings. In this regard, however, we note that some courts have upheld state laws which place the burden of proof on an inmate in release, as opposed to commitment, proceedings. E . g., Dorsey v. Solomon, 604 F.2d 271, 274 (4 Cir. 1979); Allen v. Radack, 426 F.Supp. 1052 (D.S.D.1977).