Sergeant Perry Watkins v. United States Army

NORRIS, Circuit Judge,

concurring in the judgment:

I

I concur in the judgment requiring the Army to reconsider Sgt. Watkins’ reenlistment application without regard to his homosexuality. I cannot join the majority’s opinion, however, because I agree with the dissent that the judgment cannot rest on the doctrine of equitable estoppel. The Supreme Court has declined to approve the invocation of equitable estoppel against the government even in cases where the facts are no less sympathetic than the facts in Sgt. Watkins’ case. See, e.g., INS v. Miranda, 459 U.S. 14, 17-19 , 103 S.Ct. 281, 282-84 , 74 L.Ed.2d 12 (1982) (per curiam) (reversing Ninth Circuit decision equitably estopping INS from denying resident status to alien spouse of citizen when petitioner became ineligible during INS delay in processing application); INS v. Hibi, 414 U.S. 5 , 94 S.Ct. 19 , 38 L.Ed.2d 7 (1973) (per curiam) (reversing Ninth Circuit decision equitably estopping INS from denying citizenship to Filipino war veteran); Montana v. Kennedy, 366 U.S. 308, 314-15 , 81 S.Ct. 1336, 1340-41 , 6 L.Ed.2d 313 (1961) (government not estopped to deny citizenship to child of U.S. citizen born while his mother was living abroad, even though government official advised her that she could not return to the U.S. to have her baby). Indeed, the Supreme Court has expressed uncertainty as to whether equitable estoppel can ever be invoked against the government. See Heckler v. Community Health Servs., 467 U.S. 51, 60-61 , 104 S.Ct. 2218, 2224 , 81 L.Ed.2d 42 (1984). In any event, I see no justification for invoking the doctrine on the facts of this ease.

In my view, Watkins is entitled to relief because the Army denied him the equal protection of the laws by discharging and refusing to reenlist him solely on the basis of his homosexuality. Before addressing Watkins’ claim that the Army’s regulations *712 on homosexuality violate equal protection, however, I must address Watkins’ non-constitutional claim — that the Army’s discharge and reenlistment regulations are arbitrary and capricious under the Administrative Procedure Act, 5 U.S.C. § 706 (2)(A). 1 I reject this claim because Watkins does not argue that the Army’s regulations on homosexuality themselves violate the Administrative Procedure Act; rather he argues only that the regulations are arbitrary as applied to the facts of his case. Because he does not argue that the regulations on their face are arbitrary or capricious, Watkins’ APA claim must fail. See Watkins I, 721 F.2d at 690-91.

I now turn to Watkins’ claim that the Army’s regulations deny him equal protection of the laws in violation of the Fifth Amendment. 2 Watkins argues that the Army’s regulations constitute an invidious discrimination based on sexual orientation. 3 To evaluate this claim I must engage in a three-stage inquiry. First, I must decide whether the regulations in fact discriminate on the basis of sexual orientation. Second, I must decide which level of judicial scrutiny applies by asking whether discrimination based on sexual orientation burdens a suspect or quasi-suspect class, 4 which would make it subject, respectively, to strict or intermediate scrutiny. See City of Cleburne v. Cleburne Living Center, 473 U.S. 432, 439-41 , 105 S.Ct. 3249, 3253-55 , 87 L.Ed.2d 313 (1985). If the discrimination burdens no such class, it is subject to ordinary rationality review. Id. Finally, I must decide whether the challenged regulations survive the applicable level of scrutiny by deciding whether, under strict scrutiny, the legal classification is necessary to serve a compelling governmental interest; whether, under intermediate scrutiny, the classification is substantially related to an important governmental interest; or whether, under rationality review, the classification is rationally related to a legitimate governmental interest. See id.

II

I turn first to the threshold question raised by Watkins’ equal protection claim: Do the Army’s regulations discriminate on the basis of sexual orientation? The portion of the Army’s reenlistment regulation that bars homosexuals from reenlisting states in full:

Applicants to whom the disqualifications below apply are ineligible for RA [Regular Army] reenlistment at any time and *713 requests for waiver or exception to policy-will not be submitted....

c. Persons of questionable moral character and a history of antisocial behavior, sexual perversion or homosexuality. A person who has committed homosexual acts or is an admitted homosexual but as to whom there is no evidence that they have engaged in homosexual acts either before or during military services is included. (See note 1)_

k. Persons being discharged' under AR 635-200 for homosexuality_

Note: Homosexual acts consist of bodily contact between persons of the same sex, actively undertaken or passively permitted, with the intent of obtaining or giving sexual satisfaction, or any proposal, solicitation, or attempt to perform such an act. Persons who have been involved in homosexual acts in an apparently isolated episode, stemming solely from immaturity, curiousity [sic], or intoxication, and in the absence of other evidence that the person is a homosexual, normally will not be excluded from reenlistment. A homosexual is a person, regardless of sex, who desires bodily contact between persons of the same sex, actively undertaken or passively permitted, with the intent to obtain or give sexual gratification. Any official, private, or public profession of homosexuality, may be considered in determining whether a person is an admitted homosexual.

AR 601-280, U 2-21. Although worded in somewhat greater detail, the Army’s regulation mandating the separation of homosexual soldiers from service (discharge), AR 635-200, is essentially the same in substance. 5

*714 On their face, these regulations discriminate against homosexuals on the basis of their sexual orientation. Under the regulations any homosexual act or statement of homosexuality gives rise to a presumption of homosexual orientation, and anyone who fails to rebut that presumption is conclusively barred from Army service. In other words, the regulations target homosexual orientation itself. The homosexual acts and statements are merely relevant, and rebuttable, indicators of that orientation.

In spite of these facial appearances, the Army argues that its regulations target homosexual conduct rather than orientation. I cannot agree. A close reading of the complex regulations leaves no room for doubt that the regulations target orientation rather than conduct.

Under the Army’s regulations, “homosexuality,” not sexual conduct, is clearly the operative trait for disqualification. AR 601-280, K 2-21(c); see also AR 635-200, ¶ 15-l(a) (articulating the same goal). For example, the regulations ban homosexuals who have done nothing more than acknowledge their homosexual orientation even in the absence of evidence that the persons ever engaged in any form of sexual conduct. The reenlistment regulation disqualifies any “admitted homosexual” — a status that can be proved by “[a]ny official, private, or public profession of homosexuality” even if “there is no evidence that they have engaged in homosexual acts either before or during military service.” AR 601-280, ¶ 2-21(c) & note; see also AR 635-200, K 15-3(b). Since the regulations define a “homosexual” as “a person, regardless of sex, who desires bodily contact between persons of the same sex, actively undertaken or passively permitted, with the intent to obtain or give sexual gratification,” a person can be deemed homosexual under the regulations without ever engaging in a homosexual act. 601-280, K 2-21(c) & note (emphasis added); see also A.R. 635-200, 15-2(a) (same desire sufficient to make one homosexual). Thus, no matter what statements a person has made, and what conduct he or she has engaged in, the ultimate evidentiary issue is whether he or she has a homosexual orientation. Under the reenlistment regulation, persons are disqualified from reenlisting only if, based on any “profession of homosexuality” they have made, they are found to have a homosexual orientation. AR 601-280, K 2-21(c) & note. Similarly, under the discharge regulation a soldier must be discharged if “[t]he soldier has stated that he or she is a homosexual or bisexual, unless there is a further finding that the soldier is not a homosexual or bisexual.” AR 635-200, it 15-3(b) (emphasis added). In short, the regulations do not penalize all statements of sexual desire, or even only statements of homosexual desire; they penalize only homosexuals who declare their homosexual orientation.

True, a “person who has committed homosexual acts” is also presumptively “included” under the reenlistment regulation as a person excludable for “homosexuality.” AR 601-280, K 2-21(c); see also AR 635-200, K 15-3(a). But it is clear that this provision is merely designed to round out the possible evidentiary grounds for inferring a homosexual orientation. The regulations define “homosexual acts” to encompass any “bodily contact between persons of the same sex, actively undertaken or passively permitted, with the intent of obtaining or giving sexual satisfaction, or any proposal, solicitation, or attempt to perform such an act.” AR 601-280, K 2-21(c) & note; see also AR 635-200, UK 15-2(c) & 15-3(a) (stating the same in slightly different order). Thus, the regulations barring homosexuals from the Army cover any form of bodily contact between persons of the same sex that gives sexual satisfaction — from oral and anal intercourse to holding hands, kissing, caressing and any number of other sexual acts. Indeed, in this case the Army tried to prove at Wat *715 kins’ discharge proceedings that he had committed a homosexual act described as squeezing the knee of a male soldier, but failed to prove it was Watkins who did the alleged knee-squeezing. Moreover, even non-sexual conduct can trigger a presumption of homosexuality: The regulations provide for the discharge of soldiers who have “married or attempted to marry a person known to be of the same sex ... unless there are further findings that the soldier is not a homosexual or bisexual.” AR 635-200, II 15-3(c) (emphasis added). With all the acts and statements that can serve as presumptive evidence of homosexuality under the regulations, it is hard to think of any grounds for inferring homosexual orientation that are not included. 6 The fact remains, however, that homosexual orientation, not homosexual conduct, is plainly the object of the Army’s regulations.

Moreover, under the regulations a person is not automatically disqualified from Army service just because he or she committed a homosexual act. Persons may still qualify for the Army despite their homosexual conduct if they prove to the satisfaction of Army officials that their orientation is heterosexual rather than homosexual. To illustrate, the discharge regulation provides that a soldier who engages in homosexual acts can escape discharge if he can show that the conduct was “a departure from the soldier’s usual and custom-

ary behavior” that “is unlikely to recur because it is shown, for example, that the act occurred because of immaturity, intoxication, coercion, or a desire to avoid military service” and that the “soldier does not desire to engage in or intend to engage in homosexual acts.” AR 635-200, ¶ 15-3(a). The regulation expressly states, “The intent of this policy is to permit retention only of nonhomosexual soldiers who, because of extenuating circumstances engaged in, attempted to engage in, or solicited a homosexual act.” Id. at note (emphasis in original). Similarly, the Army’s ban on reenlisting persons who have committed homosexual acts does not apply to “[pjersons who have been involved in homosexual acts in an apparently isolated episode, stemming solely from immaturity, curiousity [sic], or intoxication, and in the absence of other evidence that the person is a homosexual.” AR 601-280, 112-21 note. If a straight soldier and a gay soldier of the same sex engage in homosexual acts because they are drunk, immature or curious, the straight soldier may remain in the Army while the gay soldier is automatically terminated. In short, the regulations do not penalize soldiers for engaging in homosexual acts; they penalize soldiers who have engaged in homosexual acts only when the Army decides that those soldiers are actually gay. 7

*716 In sum, the discrimination against homosexual orientation under these regulations is about as complete as one could imagine. The regulations make any act or statement that might conceivably indicate a homosexual orientation evidence of homosexuality; that evidence is in turn weighed against any evidence of a heterosexual orientation. It is thus clear in answer to my threshold equal protection inquiry that the regulations directly burden the class consisting of persons of homosexual orientation.

Ill

A

Before reaching the question of the level of scrutiny applicable to discrimination based on sexual orientation and the question whether the Army’s regulations survive the applicable level of scrutiny, I first address the Army’s argument that Bowers v. Hardwick, 478 U.S. 186 , 106 S.Ct. 2841 , 92 L.Ed.2d 140 (1986), forecloses Watkins’ equal protection claim. In Hardwick , the Court rejected a claim by a homosexual that a Georgia statute criminalizing sodomy deprived him of his liberty without due process of law in violation of the Fourteenth Amendment. More specifically, the Court held that the constitutionally protected right to privacy — recognized in cases such as Griswold v. Connecticut, 381 U.S. 479 , 85 S.Ct. 1678 , 14 L.Ed.2d 510 (1965), and Eisenstadt v. Baird, 405 U.S. 438 , 92 S.Ct. 1029 , 31 L.Ed.2d 349 (1972) — does not extend to acts of consensual homosexual sodomy. 8 See id. 478 U.S. at 190-96 , 106 S.Ct. at 2843-46 . The Court’s holding was limited to this due process question. The parties did not argue and the Court explicitly did not decide the question whether the Georgia sodomy statute might violate the equal protection clause. See id. at 196, n. 8 , 9 106 S.Ct. at 2846 n. 8.

The Army nonetheless argues that it would be “incongruous” to hold that its regulations deprive gays of equal protection of the laws when Hardwick holds that there is no constitutionally protected privacy right to engage in homosexual sodomy. Army’s Second Supp. Brief at 19. I could not disagree more. First, while Hardwick does indeed hold that the due process clause provides no substantive privacy protection for acts of private homosexual sodomy, nothing in Hardwick suggests that the state may penalize gays merely for their sexual orientation. Cf. Robinson v. California, 370 U.S. 660 , 82 S.Ct. 1417 , 8 L.Ed.2d 758 (1962) (holding that state violated due process by criminalizing the status of narcotics addiction, even though the state could criminalize the use of the narcotics — conduct in which narcotics addicts by definition are prone to engage). In other words, the class of persons involved in Hardwick — those who engage in homosexual sodomy — is not congruous with the class of persons targeted by the Army’s regulations — those with a homosexual orientation. Hardwick was a “conduct” *717 case; Watkins’ is an “orientation” case. 10

Second, and more importantly, Hardwick does not foreclose Watkins’ claim because Hardwick was a due process, not an equal protection case. 11 Although the Army acknowledges, as it must, that Hardwick does not discuss equal protection explicitly, the Army nonetheless argues that Hard-wick’s discussion of due process has equal protection implications. Specifically, the Army argues that the Hardwick Court, in holding that the criminalization of homosexual sodomy does not violate due process, decided sub silentio that the criminalization of heterosexual sodomy would violate due process. The Army concludes from this that Hardwick is controlling precedent that the government may discriminate against homosexuals without violating equal protection.

Both the premise and the conclusion of the Army’s argument are mistaken. In the first place, Hardwick did not decide sub silentio that heterosexual sodomy is constitutionally protected. Indeed, the Court expressly refused to take a position on whether heterosexual sodomy was protected by the due process clause. 12 Second, even if we accept, arguendo, the Army’s premise that the Hardwick Court drew a distinction between homosexual sodomy and heterosexual sodomy for due process purposes, such a distinction under the due process clause would have no bearing on whether the equal protection clause nonetheless prohibits official discrimination against homosexuals. I discuss these points in turn.

Implicit in the Army’s position is the proposition that the Court in Hardwick somehow did decide that the due process clause prohibits a state from criminalizing heterosexual sodomy. That is, the Army reads Justice White’s opinion in Hardwick as extending the zone of privacy first recognized in Griswold to heterosexual sodomy, thus drawing a due process line between heterosexual and homosexual sodomy. That reading of Hardwick flies directly in the face of footnote 2, which expressly reserves the question of the constitutionality of the Georgia statute as applied to heterosexual sodomy. See 478 U.S. at 188 n. 2, 106 S.Ct. at 2842 n. 2. 13

Even apart from the Court’s express reservation of this question, the Army’s reading of Hardwick is untenable. I see no basis for reading Hardwick as holding sub silentio that a right to engage in heterosexual sodomy is “deeply rooted in this Nation’s history and tradition” or “implicit in the concept of ordered liberty” — which would be necessary for heterosexual sodomy to qualify for due process protection *718 under Hardwick’s analysis. 14 Note that when the Court found the suggestion that homosexual sodomy qualified for due process protection to be “at best, facetious,” 478 U.S. at 194 , 106 S.Ct. at 2846 , it relied upon the historical fact that sodomy was a criminal offense at common law, under the laws of all 13 colonies, and, until 1961, under the laws of all 50 states. 478 U.S. at 192-94 , 106 S.Ct. at 2844-46 . Note further that the Court did not find it significant that these laws, as Justice Stevens pointed out in his dissent, drew no distinction between homosexual and heterosexual sodomy. See 478 U.S. at 214-15 , 106 S.Ct. at 2856-57 . 15 They outlawed all acts of sodomy, both homosexual and heterosexual.

In light of the historical record relied upon by the Court, there is no way to read Hardwick as establishing that heterosexual sodomy is “deeply rooted in this Nation’s history and tradition” while homosexual sodomy is not. I find it untenable, then, to interpret Hardwick as extending due process protection to heterosexual conduct while denying such protection to homosexual conduct. It is hard to imagine that the Court in Hardwick intended to suggest that acts of heterosexual sodomy implicate higher constitutional values than acts of homosexual sodomy.

Even if, as the Army implicitly argues, Hardwick did in fact extend constitutional protection to heterosexual sodomy while denying it to homosexual sodomy, such a differentiation between heterosexual and homosexual sodomy for due process purposes would have no bearing — none—on the entirely separate question whether official discrimination against homosexuals violates the equal protection clause. The relevant inquiry in equal protection jurisprudence is fundamentally different from the relevant due process inquiry. The due process clause, as the Court recognized in Hardwick , protects practices which are “deeply rooted in this Nation’s history and tradition.” The equal protection clause, in contrast, protects minorities from discriminatory treatment at the hands of the majority. Its purpose is not to protect traditional values and practices, but to call into question such values and practices when they operate to burden disadvantaged minorities. As Professor Sunstein puts it:

From its inception, the Due Process Clause has been interpreted largely (though not exclusively) to protect traditional practices against short-run departures. The clause has therefore been associated with a particular conception of judicial review, one that sees the courts as safeguards against novel developments brought about by temporary majorities who are insufficiently sensitive to the claims of history.

The Equal Protection Clause, by contrast, has been understood as an attempt to protect disadvantaged groups from discriminatory practices, however deeply engrained and longstanding. The Due Process Clause often looks backward; it is highly relevant to the Due Process issue whether an existing or time-honored convention, described at the appropriate level of generality, is violated by the practice under attack. By contrast, the Equal Protection Clause looks forward, serving to invalidate practices that were widespread at the time of its ratification and that were expected to endure. The two clauses therefore operate along different tracks.

Sunstein, supra note 10, at 1163.

The Supreme Court did not decide in Hardwick — and indeed has never decided in any case — whether discrimination against homosexuals violates equal protection. All Hardwick decided is that homo *719 sexual sodomy is not a practice so “deeply rooted in this Nation’s history and tradition” that it falls within the zone of personal privacy protected by the due process clause. It is perfectly consistent to say that homosexual sodomy is not a practice so deeply rooted in our traditions as to merit due process protection, and at the same time to say, for example, that because homosexuals have historically been subject to invidious discrimination, laws which burden homosexuals as a class should be subjected to heightened scrutiny under the equal protection clause. Indeed, the two propositions may be complementary: In all probability, homosexuality is not considered a deeply-rooted part of our traditions precisely because homosexuals have historically been subjected to invidious discrimination. In any case, homosexuals do not become “fair game” for discrimination simply because their sexual practices are not considered part of our mainstream traditions.

A hypothetical may help make the point. Suppose a city passed a “single family occupancy” housing ordinance allowing only members of the immediate, nuclear family to live in the same house. 16 Suppose further that a disproportionate number of black families in the community lived together in extended families that included, for example, cousins and grandparents. 17 Finally, suppose the ordinance was motivated by a racially discriminatory purpose. 18 A black family challenging the ordinance could raise a due process claim, arguing that the ordinance impermissibly intruded on “deeply rooted” family traditions. In real life, the Court found such a due process claim persuasive. 19 But suppose the Court had rejected the due process claim. Suppose the Court had instead agreed with the city of East Cleveland that the privacy interests protected by the Constitution do not include extended family relationships— that the due process clause does not “give grandmothers any fundamental rights with respect to grandsons.” 20 In that event, the black family could still challenge the ordinance on equal protection grounds, arguing that the ordinance discriminated against blacks. Could anyone seriously maintain that the Court’s hypothetical refusal to give due process protection to “extended family” living would have any bearing on the black family’s equal protection claim? Of course not. And the black family’s equal protection claim would be no less viable even if the Court in the hypothetical had ruled that due process does protect the nuclear family (in the hypothetical, the form disproportionately favored by the whites in the community) but does not protect the extended family (disproportionately favored by blacks).

The relationship between Hardwick and Watkins’ case is exactly the same as the relationship between the due process and equal protection claims in this hypothetical. Whether homosexual conduct is protected by the due process clause is an entirely separate question from whether the equal protection clause prohibits discrimination against homosexuals. And in answering this latter question, it makes no difference whether the Hardwick Court intended to extend due process protection to heterosexual conduct, but not homosexual conduct. In sum, the equal protection question presented by Sgt. Watkins simply is not answered — not in the slightest — by Hard-wick.

The Army also argues that Hardwick’s concern “about the limits of the Court’s role in carrying out its constitutional mandate,” 478 U.S. at 190 , 106 S.Ct. at 2843 , should prevent courts from holding that equal protection doctrine protects homosexuals from discrimination. To be sure, the *720 Court in Hardwick justified its decision to cabin the right to privacy largely by pointing to the problems allegedly created when judges recognize constitutional “rights not readily identifiable in the Constitution’s text” and “having little or no cognizable roots in the language or design of the Constitution.” 478 U.S. at 191, 194 , 106 S.Ct. at 2844, 2846 . The Court stressed its concern that such rights might be perceived as involving “the imposition of the Justices’ own choice of values on the States and the Federal Government” and that this antidemocratic perception might undermine the legitimacy of the Court. Id. Finally, the Court expressed the more specific concern about potential difficulties in defining the contours of the right to privacy. See id. at 195-96 , 106 S.Ct. at 2846-47 .

Whatever one might think about the Hardwick Court’s concerns about substantive due process in general and the right of privacy in particular, these concerns have little if any relevance to equal protection doctrine. 21 The right to equal protection of the laws has a clear basis in the text of the Constitution. This principle of equal treatment, when imposed against majoritarian rule, arises from the Constitution itself, not from judicial fiat. Moreover, equal protection doctrine does not prevent the majority from enacting laws based on its substantive value choices. Equal protection simply requires that the majority apply its values evenhandedly. Indeed, equal protection doctrine plays an important role in perfecting, rather than frustrating, the democratic process. The constitutional requirement of evenhandedness advances the political legitimacy of majority rule by safeguarding minorities from majoritarian oppression. The requirement of evenhandedness also facilitates a representation of minorities in government that advances the operation of representative democracy. 22 Finally, the practical difficulties of defining the requirements imposed by equal protection, while not insignificant, do not involve the judiciary in the same degree of value-based line-drawing that the Supreme Court in Hardwick found so troublesome in defining the contours of substantive due process. In sum, the driving force behind Hardwick is the Court’s ongoing concern with the expansion of rights under substantive due process, not an unbounded antipathy toward a disfavored group.

B

The Army also relies upon Better v. Middendorf 632 F.2d 788 (9th Cir.1980), cert. denied, 452 U.S. 905 , 101 S.Ct. 3030 , 69 L.Ed.2d 405 (1981), Hatheway v. Secretary of the Army, 641 F.2d 1376 (9th Cir.), cert. denied, 454 U.S. 864 , 102 S.Ct. 324 , 70 L.Ed.2d 164 (1981), and DeSantis v. Pacific Tel. & Tel. Co., 608 F.2d 327 (9th Cir.1979), to argue that the Ninth Circuit has already rejected the kind of equal protection attack Watkins makes. In my view, the equal protection question Watkins raises — whether the Army’s regulations should be subjected to strict scrutiny because homosexuals constitute a suspect class — was not addressed in any of these Ninth Circuit cases.

The Army’s reliance on Better is misplaced because Better, like Hardwick , is a substantive due process case, not an equal protection case. In rejecting a substantive due process challenge to Navy regulations providing for the discharge of personnel who engaged in homosexual acts, our court held in Better that substantive due process *721 required only that courts balance the governmental and individual interests at stake in a fashion similar to intermediate scrutiny. Be ller, 632 F.2d at 805-12. As now-Justice Kennedy’s carefully tailored opinion makes clear, Beller’s appeal did “not require us to address the question whether consensual private homosexual conduct is a fundamental right as that term is used in equal protection ... [and was] not presented to us as implicating a suspect or quasi-suspect classification.... Substantive due process, not equal protection, was the basis of the constitutional claim, and we address the case in those terms.” Id. at 807. Thus, Better, like Hardwick , has no relevance to Watkins’ claim that the challenged governmental regulations discriminate against a suspect class in violation of equal protection doctrine. See Sethy v. Alameda County Water Dist., 545 F.2d 1157 , 1159-60 (9th Cir.1976) (en banc) (a prior decision is not precedent on issues that were neither raised by counsel nor discussed in the opinion of the court); Sakamoto v. Duty Free Shoppers, 764 F.2d 1285, 1288 (9th Cir.1985) (same).

The Army’s reliance on Hatheway v. Secretary of the Army, 641 F.2d 1376 (9th Cir.), cert. denied, 454 U.S. 864 , 102 S.Ct. 324 , 70 L.Ed.2d 164 (1981), is also misplaced. In Better, our court reserved two distinct equal protection questions: first, whether the challenged regulations penalizing homosexual conduct burdened the exercise of a fundamental or important substantive right to engage in certain conduct; second, whether the challenged regulations discriminated against a suspect or quasi-suspect class. As explained below, in Hatheway we clearly answered the first of these discrete equal protection questions. The Army argues, however, that Hatheway also decided the second question reserved in Better — the question raised in Watkins’ claim — whether homosexuals constitute a suspect or quasi-suspect class. 23

Hatheway, a soldier convicted of committing sodomy in violation of 10 U.S.C. § 925 , claimed that the Army was prosecuting cases involving homosexual sodomy while refusing to prosecute cases involving heterosexual sodomy. Our court “understood Hatheway’s claim (that the commission of a homosexual act is an impermissible basis for prosecution) to be an equal protection argument,” Hatheway, 641 F.2d at 1382 , which we treated as resting on the branch of equal protection doctrine concerned with whether a governmental classification burdens a fundamental or important substantive right to engage in certain conduct. Thus, we explicitly characterized Hathe-way’s claim “that the commission of a homosexual act is an impermissible basis for prosecution” to be the sort of equal protection claim that “implicate[d] the ‘right to be free ... from unwarranted intrusions into one’s privacy.’ ” 641 F.2d at 1382 (quoting Stanley v. Georgia, 394 U.S. 557, 564 , 89 S.Ct. 1243, 1247-48 , 22 L.Ed.2d 542 (1969)). We then reasoned that the interest at stake in Hatheway was similar to the substantive interest at stake in Better. 641 F.2d at 1382 . Because in Better we decided that under the due process clause the right to engage in homosexual conduct merited “heightened solicitude,” but not strict scrutiny, in Hatheway we adopted this assessment for the purposes of our fundamental rights equal protection analysis. Accordingly, we applied intermediate scrutiny to the Army’s actions and held that “the selection of cases involving homosexual acts for Article 125 prosecutions” was permissible because such prosecutions bore “a substantial relationship to an important government interest.” Id. Thus, we rejected Hatheway’s claim based on an analysis of the fundamental rights branch of equal protection doctrine, the branch of equal protection doctrine upon which Watkins does not rely.

*722 The Army argues that Hatheway should nonetheless be read as having decided the suspect class question. In support of this argument, the Army relies upon a single sentence in a footnote — the opinion’s only reference to suspect class analysis. In footnote 6 we wrote: “Though ‘[t]he courts have not designated homosexuals a “suspect” or “quasi-suspect” classification so as to require more exacting scrutiny/ DeSantis v. Pacific Tel. & Tel. Co., 608 F.2d 327 , 333 (9th Cir.1979), heightened scrutiny is independently required where a classification penalizes the exercise of a fundamental right. See Shapiro v. Thompson, 394 U.S. 618, 634 , 89 S.Ct. 1322, 1331 , 22 L.Ed. 2d 600 (1969).” 641 F.2d at 1382 n. 6. Although I recognize that the intended purpose of this footnote is not entirely clear, I cannot fairly read this passing reference as an adjudication of the important and unresolved constitutional question whether homosexuals constitute a suspect or quasi-suspect class for the purpose of equal protection analysis. Rather, I read footnote 6 as simply clarifying the distinction between the suspect class and fundamental rights branches of equal protection doctrine while acknowledging that at the time of the Hatheway decision courts had not yet decided whether homosexuals constitute a suspect or quasi-suspect class. That the critical language in footnote 6 is taken directly from our opinion in DeSantis, 608 F.2d at 327, informs our reading. In De-Santis, we acknowledged that our court had not yet designated homosexuals as a suspect or quasi-suspect class, but we did not decide that homosexuals should not be so designated. See infra at 722-23. Similarly, in footnote 6 of Hatheway , we remarked on the existing state of the law with respect to homosexuals without deciding the open question whether homosexuals constitute a suspect or quasi-suspect class. In other words, I read Hatheway as interpreting the equal protection claim presented as resting solely on the fundamental rights branch of equal protection analysis. Hatheway is also distinguishable from this case because, like both Hardwick and Beller, Hatheway involved a classification based on homosexual conduct, not homosexual orientation. As I note throughout my opinion, this distinction is relevant to an analysis of Watkins’ particular equal protection claim.

Because I read Hatheway as not deciding the suspect class issue, and because the suspect class and fundamental rights branches of equal protection doctrine involve very separate inquiries, see e.g., San Antonio School Indep. District v. Rodriguez, 411 U.S. 1, 18-39 , 93 S.Ct. 1278, 1288-1300 , 36 L.Ed.2d 16 (1973); Perry, Modem Equal Protection, 79 Colum.L. Rev. 1023, 1074-83 (1979); Developments in the Law — Equal Protection, 82 Harv.L. Rev. 1065, 1087-1131 (1969), Hatheway does not stand in the way of Watkins’ equal protection claim. 24

Finally, I must reject the Army’s contention that in DeSantis v. Pacific Tel. & Tel. Co., 608 F.2d 327 (9th Cir.1979), our court held that homosexuals do not constitute a suspect or quasi-suspect class. In DeSan-tis, we considered whether homosexuals were a protected class within the meaning of 42 U.S.C. § 1985 (3), which secures a right of action against private parties who conspire to deprive “any person or class of persons of the equal protection of the laws.” We held that section 1985(3) protects only those groups that have been previously determined by Congress or the courts to need special Federal assistance in protecting their civil rights. 608 F.2d at 333. 25 Applying this standard, we concluded that homosexuals could not receive the protection of section 1985(3), in part because “[t]he courts have not designated homosexuals a ‘suspect’ or ‘quasi-suspect’ classification,” 608 F.2d at 333 (emphasis *723 added). We did not, and did not need to, consider whether homosexuals should be considered a suspect class. Thus, our decision that section 1985(3) did not protect homosexuals turned simply on the point that courts had not yet designated homosexuals a suspect class. Although DeSan-tis does not articulate the reasons that section 1985(3) requires a prior governmental determination, it seems likely— since section 1985(3) authorizes suits against private individuals and requires no state action — that our court’s interpretation of the statute was animated by concerns about providing potential defendants with sufficient notice of the statute’s scope. Cf. Marks v. United States, 430 U.S. 188, 192 , 97 S.Ct. 990, 993 , 51 L.Ed.2d 260 (1977) (judicial enlargement of the scope of criminal statute without fair notice violates due process).

C

While neither the Supreme Court nor the Ninth Circuit has decided the question presented in Watkins’ appeal — whether persons of homosexual orientation constitute a suspect class under equal protection doctrine — several other circuits have considered the different but related question whether laws burdening the class of individuals engaging in homosexual conduct trigger heightened scrutiny under the equal protection clause. Only one circuit, however, has given the issue more than cursory treatment. 26 In Padula v. Webster, 822 F.2d 97 (D.C.Cir.1987), the District of Columbia Circuit rejected an equal protection challenge to the FBI’s policy of discriminating against “practicing homosexuals” in its hiring decisions. The D.C. Circuit did not analyze whether the class of persons engaging in homosexual conduct satisfies the traditional indicia of suspectness, see infra at 723-728, but rather concluded summarily (as the Army urges us to .do here) that “[i]t would be quite anomalous, on its face, to declare status defined by conduct that states may constitutionally criminalize as deserving of strict scrutiny under the equal protection clause.” Id. at 103 . The D.C. Circuit reasoned that “[i]f the [Supreme] Court [in Hardwick ] was unwilling to object to state laws that criminalize the behavior that defines the class, it is hardly open to a lower court to conclude that state sponsored discrimination against the class is invidious. After all, there can hardly be more palpable discrimination against a class than making the conduct that defines the class criminal.” Id.

Padula’s reasoning rests on the false premise that Hardwick issues a blanket approval for discrimination against homosexuals. To repeat what I said above, Hardwick held only that the constitutionally protected right to privacy does not extend to homosexual sodomy. The case had nothing to do with equal protection. I see no principled way to transmogrify the Court’s holding that the due process clause permits states to criminalize specific sexual conduct commonly engaged in by homosexuals into a holding that the equal protection clause gives states a license to pass “homosexual laws” — laws imposing special restrictions on gays because they are gay. *724 Thus, I find Padula unpersuasive. Moreover, as I have reiterated thoughout this opinion, the regulations at issue here target orientation, not conduct — the trait at issue in Padula .

In sum, no federal appellate court 27 has decided the critical issue raised by Watkins’ claim: whether persons of homosexual orientation constitute a suspect class under equal protection doctrine. To be sure, Hardwick forecloses Watkins from making a due process claim that the Army’s regulations impinge on an asserted fundamental right to engage in homosexual sodomy. But Watkins makes no such claim. Rather, he claims only that the Army’s regulations discriminate against him because of his membership in a disfavored group — homosexuals. This claim is not barred by precedent.

IV

I now address the merits of Watkins’ argument that the Army’s regulations must be subjected to strict scrutiny because homosexuals constitute a suspect class under equal protection jurisprudence. The Supreme Court has identified several factors that guide our suspect class inquiry. I now turn to each of these factors.

The first factor the Supreme Court generally considers is whether the group at issue has suffered a history of purposeful discrimination. See, e.g., Cleburne, 473 U.S. at 441 , 105 S.Ct. at 3254-55 ; Massachusetts Bd. of Retirement v. Murgia, 427 U.S. 307, 313 , 96 S.Ct. 2562, 2566-67 , 49 L.Ed.2d 520 (1976); Rodriguez, 411 U.S. at 28 , 93 S.Ct. at 1293-94 ; Frontiero, 411 U.S. at 684-85 , 93 S.Ct. at 1769-70 (plurality). As the Army concedes, 28 it is indisputable that “homosexuals have historically been the object of pernicious and sustained hostility.” Rowland v. Mad River Local School Dist., 470 U.S. 1009, 1014 , 105 S.Ct. 1373 , 1376-77, 84 L.Ed.2d 392 (1985) (Brennan, J., dissenting from denial of cert.). Recently courts have echoed the same harsh truth: “Lesbians and gays have been the object of some of the deepest prejudice and hatred in American society.” High Tech Gays v. Defense Industrial Security Clearance Office, 668 F.Supp 1361, 1369 (1987) (invalidating Defense Department practice of subjecting gay security clearance applicants to more exacting scrutiny than heterosexual applicants); see also BenShalom v. Secretary of the Army, 703 F.Supp. 1372 (1989) (homosexuals historically subject to discrimination).

Discrimination against homosexuals has been pervasive in both the public and private sectors. Legislative bodies have excluded homosexuals from certain jobs and schools, and have prevented homosexuals marriage. In the private sphere, homosexuals continue to face discrimination in jobs, housing and churches. See generally Note, An Argument for the Application of Equal Protection Heightened Scrutiny to Classifications Based on Homosexuality, 57 S.Cal.L.Rev. 797, 824-25 (1984) (documenting the history of discrimination). Moreover, reports of violence against homosexuals have become commonplace in our society. In sum, the discrimination faced by homosexuals is plainly no less pernicious or intense than the discrimination faced by other groups already treated as suspect classes, such as aliens or people of a particular national origin. See, e.g., Cleburne, 473 U.S. at 440 , 105 S.Ct. at 3254 (identifying suspect groups).

The second factor that the Supreme Court considers in suspect class analysis is difficult to capsulize and may in fact represent a cluster of factors grouped around a central idea — whether the discrimination embodies a gross unfairness that is sufficiently inconsistent with the ideals of equal protection to term it “invidious.” Consideration of this additional factor makes sense. After all, discrimination exists against some groups because the animus is warranted — no one could seriously argue that burglars form a suspect class. See Tribe, The Puzzling Persistence of Process-Based Constitutional Theories, 89 Yale L.J. 1063 , 1075 (1980); Note, supra, at *725 814-815 & nn. 115-116. In giving content to this concept of gross unfairness, the Court has considered (1) whether the disadvantaged class is defined by a trait that “frequently bears no relation to ability to perform or contribute to society,” Frontiero, 411 U.S. at 686 , 93 S.Ct. at 1770 (plurality); (2) whether the class has been saddled with unique disabilities because of prejudice or inaccurate stereotypes; and (3) whether the trait defining the class is immutable. See Cleburne, 473 U.S. at 440-44 , 105 S.Ct. at 3254-57 ; Plyler, 457 U.S. at 216 n. 14, 219 n. 19, 220, 223, 102 S.Ct. at 2394 n. 14, 2395-96 n. 19, 2396, 2397-98; Murgia, 427 U.S. at 313 , 96 S.Ct. at 2566-67 ; Frontiero, 411 U.S. at 685-87 , 93 S.Ct. at 1769-71 (plurality). I consider these questions in turn.

Sexual orientation plainly has no relevance to a person’s “ability to perform or contribute to society.” Sergeant Watkins’ exemplary record of military service stands as a testament to quite the opposite. Moreover, as the Army itself concluded, there is not a scintilla of evidence that Watkins’ avowed homosexuality “had either a degrading effect upon unit performance, morale or discipline, or upon his own job performance.” ER at 26c.

This irrelevance of sexual orientation to the quality of a person’s contribution to society also suggests that classifications based on sexual orientation reflect prejudice and inaccurate stereotypes — the second indicium of a classification’s gross unfairness. See Cleburne, 473 U.S. at 440-441 , 105 S.Ct. at 3254-55 . I agree with Justice Brennan that “discrimination against homosexuals is ‘likely ... to reflect deep-seated prejudice rather than ... rationality.’ ” Rowland, 470 U.S. at 1014 , 105 S.Ct. at 1376-77 (Brennan, J., dissenting from denial of cert.) (quoting Plyler, 457 U.S. at 216 n. 14, 102 S.Ct. at 2394 n. 14).

The Army suggests that the opprobrium directed towards gays does not constitute prejudice in the pejorative sense of the word, but rather is simply appropriate public disapproval of persons who engage in immoral behavior. The Army equates homosexuals with sodomists and justifies its regulations as simply reflecting a rational bias against a class of persons who engage in criminal acts of sodomy. In essence, the Army argues that homosexuals, like burglars, cannot form a suspect class because they are criminals.

The Army’s argument rests on two false premises. First, as I have noted throughout this opinion, the class burdened by the regulations at issue in this case is defined by the sexual orientation of its members, not by their sexual conduct. See supra at 712-716. To my knowledge, homosexual orientation itself has never been criminalized in this country. Moreover, any attempt to criminalize the status of an individual’s sexual orientation would present grave constitutional problems. See generally Robinson v. California, 370 U.S. 660 , 82 S.Ct. 1417 , 8 L.Ed.2d 758 (1962).

Second, little of the homosexual conduct covered by the regulations is criminal. The regulations reach many forms of homosexual conduct other than sodomy such as kissing, hand-holding, caressing, and hand-genital contact. Yet, sodomy is the only consensual adult sexual conduct that Congress has criminalized, 10 U.S.C. § 925 . Indeed, the Army points to no law, federal or state, which criminalizes any form of private consensual homosexual behavior other than sodomy. The Army’s argument that its regulations merely ban a class of criminals might be relevant, although not necessarily persuasive, if the class at issue were limited to sodomists. But the class banned from Army service is not comprised of sodomists, or even of homosexual so-domists; the class is comprised of persons of homosexual orientation whether or not they have engaged in sodomy.

Finally, I turn to immutability as an indicator of gross unfairness. The Supreme Court has never held that only classes with immutable traits can be deemed suspect. Cf., e.g., Cleburne, 473 U.S. at 442 n. 10, 105 S.Ct. at 3255 -56 n. 10 (casting doubt on immutability theory); id. at 440-441 (stating the defining characteristics of suspect classes without mentioning immutability); Murgia, 427 U.S. at 313 , 96 S.Ct. at 2566 - *726 67 (same); Rodriguez, 411 U.S. at 28 , 93 S.Ct. at 1293-94 (same). I nonetheless consider immutability because the Supreme Court has often focused on immutability, see, e.g., Plyler, 457 U.S. at 220, 102 S.Ct. at 2396; Frontiero, 411 U.S. at 686 , 93 S.Ct. at 1770 (plurality), and has sometimes described the recognized suspect classes as having immutable traits, see, e.g., Parham v. Hughes, 441 U.S. 347, 351 , 99 S.Ct. 1742, 1745 , 60 L.Ed.2d 269 (1979) (plurality opinion) (describing race, national origin, alien-age, illegitimacy, and gender as immutable).

It is clear that by “immutability” the Court has never meant strict immutability in the sense that members of the class must be physically unable to change or mask the trait defining their class. People can have operations to change their sex. Aliens can ordinarily become naturalized citizens. The status of illegitimate children can be changed. People can frequently hide their national origin by changing their customs, their names, or their associations. Lighter skinned blacks can sometimes “pass” for white, as can Latinos for Ang-los, and some people can even change their racial appearance with pigment injections. See J. Griffin, Black Like Me (1977). At a minimum, then, the Supreme Court is willing to treat a trait as effectively immutable if changing it would involve great difficulty, such as requiring a major physical change or a traumatic change of identity. Reading the case law in a more capacious manner, “immutability” may describe those traits that are so central to a person’s identity that it would be abhorrent for government to penalize a person for refusing to change them, regardless of how easy that change might be physically. Racial discrimination, for example, would not suddenly become constitutional if medical science developed an easy, cheap, and painless method of changing one’s skin pigment. See Tribe, supra, at 1073-74 n. 52. See generally Note, The Constitutional Status of Sexual Orientation: Homosexuality as a Suspect Classification, 98 Harv. L.Rev. 1285, 1303 (arguing that the ability to change a trait is not as important as whether the trait is a “determinative feature of personality”).

With these principles in mind, I have no trouble concluding that sexual orientation is immutable for the purposes of equal protection doctrine. Although the causes of homosexuality are not fully understood, scientific research indicates that we have little control over our sexual orientation and that, once acquired, our sexual orientation is largely impervious to change. See Note, supra, 57 S.Cal.L.Rev. at 817-821 (collecting sources); see also L. Tribe, supra note 23, at 945 n. 17. Scientific proof aside, it seems appropriate to ask whether heterosexuals feel capable of changing their sexual orientation. Would heterosexuals living in a city that passed an ordinance burdening those who engaged in or desired to engage in sex with persons of the opposite sex find it easy not only to abstain from heterosexual activity but also to shift the object of their sexual desires to persons of the same sex? It may be that some heterosexuals and homosexuals can change their sexual orientation through extensive therapy, neurosurgery or shock treatment. See L. Tribe, supra note 23, at 945 n. 17. But see Note, supra, 57 S.Cal.L. Rev. at 820-21 & nn. 147-149. But the possibility of such a difficult and traumatic change does not make sexual orientation “mutable” for equal protection purposes. To express the same idea under the alternative formulation, I conclude that allowing the government to penalize the failure to change such a central aspect of individual and group identity would be abhorrent to the values animating the constitutional ideal of equal protection of the laws.

The final factor the Supreme Court considers in suspect class analysis is whether the group burdened by official discrimination lacks the political power necessary to obtain redress from the political branches of government. See, e.g., Cleburne, 473 U.S. at 441 , 105 S.Ct. at 3255 ; Plyler, 457 U.S. at 216 n. 14, 102 S.Ct. at 2394 n. 14; Rodriguez, 411 U.S. at 28 , 93 S.Ct. at 1293-94 . Courts are understandably reluctant to extend heightened protection under equal protection doctrine to groups fully capable of securing their rights through the politi *727 cal process. It cannot be seriously disputed, however, that homosexuals as a group cannot protect their right to be free from invidious discrimination by appealing to the political branches.

The very fact that homosexuals have historically been underrepresented in and victimized by political bodies is itself strong evidence that they lack the political power necessary to ensure fair treatment at the hands of government. In addition, homosexuals as a group are handicapped by structural barriers that operate to make effective political participation unlikely if not impossible. First, the social, economic, and political pressures to conceal one’s homosexuality operate to discourage gays from openly protesting anti-homosexual governmental action. Ironically, by “coming out of the closet” to protest against discriminatory legislation and practices, homosexuals expose themselves to the very discrimination they seek to eliminate. As a result, the voices of many homosexuals are not even heard, let alone counted. Cf. J. Ely, supra note 21, at 163-64. “Because of the immediate and severe opprobrium often manifested against homosexuals once so identified publicly, members of this group are particularly powerless to pursue their rights openly in the political arena.” Rowland, 470 U.S. at 1014 , 105 S.Ct. at 1376-77 (Brennan, J., dissenting from denial of cert.).

Even when gays do come out of the closet to participate openly in politics, the general animus towards homosexuality may render this participation ineffective. Many heterosexuals, including elected officials, find it difficult to empathize with and take seriously the arguments advanced by homosexuals, in large part because of the lack of meaningful interaction between the heterosexual majority and the homosexual minority. Most people have little exposure to gays, both because they rarely encounter gays 29 and because — as I noted above — homosexuals are often pressured into concealing their sexual identity. Thus, elected officials sensitive to public prejudice and ignorance, and insensitive to the needs of the homosexual constituency, may refuse to even consider legislation that even appears to be pro-homosexual. See Note, supra, 98 Harv.L.Rev. at 1304 n. 96. Indeed, the Army itself argues that its regulations are justified by the need to “maintain the public acceptability of military service,” AR 635-200, 1115-2(a), because “toleration of homosexual conduct ... might be understood as tacit approval” and “the existence of homosexual units might well be a source of ridicule and notoriety.” Army’s Opening Brief at 17, 19 n. 9, 30-31 n. 18. These barriers to the exercise of political power both reinforce and are reinforced by the underrepresentation of avowed homosexuals in the decisionmak-ing bodies of government and the inability of homosexuals to prevent legislation hostile to their group interests. 30 See Frontiero, 411 U.S. at 686 & n. 17, 93 S.Ct. at 1770 & n. 17 (plurality) (underrepresentation of women in government caused in part by history of discrimination); Cleburne, 473 U.S. at 445 , 105 S.Ct. at 3257 (reasoning that the existence of legislation responsive to the needs of the mentally disabled belied the claim that they were politically powerless).

*728 In sum, all of the relevant factors drive me to the conclusion that homosexuals constitute a suspect class for equal protection purposes. Moreover, the principles that animate equal protection doctrine — the principles that gave rise to these factors in the first place — reinforce that conclusion. See also J. Ely, supra note 21, at 162-64 (classifications based on homosexuality merit heightened scrutiny); L. Tribe, supra note 23, at 944-45 n. 17 (same).

V

Having concluded that homosexuals constitute a suspect class, I now must subject the Army’s regulations facially discriminating against homosexuals to strict scrutiny. Consequently, I may uphold the regulations only if they are “ ‘necessary to promote a compelling governmental interest.’ ” Dunn v. Blumstein, 405 U.S. 330, 342 , 92 S.Ct. 995, 1003 , 31 L.Ed.2d 274 (1972) (quoting Shapiro, 394 U.S. at 634 , 89 S.Ct. at 1331 ); see also University of Calif. Regents v. Bakke, 438 U.S. 265, 357 , 98 S.Ct. 2733, 2782 , 57 L.Ed.2d 750 (1978) (Opinion of Brennan, White, Marshall & Blackmun, JJ.). The requirement of necessity means that no less restrictive alternative is available to promote the compelling governmental interest. See Dunn, 405 U.S. at 343 , 92 S.Ct. at 1003 ; Bakke, 438 U.S. at 357 , 98 S.Ct. at 2782 (Opinion of four justices).

I recognize that even under strict scrutiny, my review of military regulations must be more deferential than comparable review of laws governing civilians. See Goldman v. Weinberger, 475 U.S. 503 , 106 S.Ct. 1310, 1313 , 89 L.Ed.2d 478 (1986). While the Supreme Court does not “purport to apply a different equal protection test because of the military context, [it does] stress the deference due congressional choices among alternatives in exercising the congressional authority to raise and support armies and make rules for their governance.” Rostker v. Goldberg, 453 U.S. 57, 71 , 101 S.Ct. 2646, 2655 , 69 L.Ed.2d 478 (1981) (citing Schlesinger v. Ballard, 419 U.S. 498 , 95 S.Ct. 572 , 42 L.Ed.2d 610 (1975)). I question whether this special deference is appropriate in Watkins’ case given that Congress has chosen not to regulate homosexuality or any form of sexual conduct engaged in by military personnel save for one exception — Congress has chosen to criminalize sodomy by military personnel whether committed “with another person of the same or opposite sex.” 10 U.S.C. § 925 (emphasis added). Hence, if anything, section 925 reflects an absence of congressional intent to discriminate on the basis of sexual orientation.

In any case, even granting special deference to the policy choices of the military, I must reject many of the Army’s asserted justifications because they illegitimately cater to private biases. For example, the Army argues that it has a valid interest in maintaining morale and discipline by avoiding hostilities and “ ‘tensions between known homosexuals and other members [of the armed services] who despise/detest homosexuality.’ ” Army’s Opening Brief at 17 (quoting and incorporating into their argument Beller, 632 F.2d at 811); see also id. at 17-18, 19 n. 9, 30, 30-31 n. 18; Army’s Second Supp.Brief at 30-31 & n. 17; AR 635-200, 1Í 15-l(a). 31 The Army also *729 expresses its ‘doubts concerning a homosexual officer’s ability to command the respect and trust of the personnel he or she commands’ ” because many lower-ranked heterosexual soldiers despise and detest homosexuality. See Army’s Second Supp. Brief at 30-31 (quoting and incorporating Beller, 632 F.2d at 811); see also id. at 31 n. 17; Army’s Opening Brief at 17-18,19 n. 9, 30; AR 635-200, ¶ 15-l(a). Finally, the Army argues that the presence of gays in its ranks “might well be a source of ridicule and notoriety, harmful to the Army’s recruitment efforts” and to its public image. Army’s Opening Brief at 31 n. 18; see also id. at 15, 17, 19 n. 9, 30; AR 635-200, If 15-l(a).

These concerns strike an all-too-familiar chord. For much of our history, the military’s fear of racial tension kept black soldiers separated from whites. As recently as World War II both the Army chief of staff and the Secretary of the Navy justified racial segregation in the ranks as necessary to maintain efficiency, discipline, and morale. See G. Ware, William Hastie: Grace Under Pressure 99, 134 (1984). 32 Today, it is unthinkable that the judiciary would defer to the Army’s prior “professional” judgment that black and white soldiers had to be segregated to avoid interracial tensions. Indeed, the Supreme Court has decisively rejected the notion that private prejudice against minorities can ever justify official discrimination, even when those private prejudices create real and legitimate problems. See Palmore v. Sidoti, 466 U.S. 429 , 104 S.Ct. 1879 , 80 L.Ed.2d 421 (1984).

In Palmore , a state granted custody of a child to her father because her white mother had remarried a black man. The state rested its decision on the best interests of the child, reasoning that, despite improvements in race relations, the social reality was that the child would likely suffer social stigmatization if she had parents of different races. A unanimous Court, in an opinion by Chief Justice Burger, conceded the importance of the state’s interest in the welfare of the child, but nonetheless reversed with the following reasoning:

It would ignore reality to suggest that racial and ethnic prejudices do not exist or that all manifestations of those prejudices have been eliminated.... The question, however, is whether the reality of private biases and the possible injury they might inflict are permissible considerations for removal of an infant child from the custody of its natural mother. We have little difficulty concluding that they are not. The Constitution cannot control such prejudices but neither can it tolerate them. Private biases may be outside the reach of the law, but the law cannot, directly or indirectly, give them effect.

Id. at 433 , 104 S.Ct. at 1882. Thus, Pal-more forecloses the Army from justifying its ban on homosexuals on the ground that private prejudice against homosexuals would somehow undermine the strength of our armed forces if homosexuals were permitted to serve. See also Cleburne, 473 U.S. at 448 , 105 S.Ct. at 3258-60 (even under rationality review of discrimination against group that is neither suspect nor quasi-suspect, catering to private prejudice is not a cognizable state interest).

The Army’s defense of its regulations, however, goes beyond its professed fear of prejudice in the ranks. Apparently, the Army believes that its regulations rooting out persons with certain sexual tendencies are not merely a response to prejudice, but are also grounded in legitimate moral norms. In other words, the Army believes that its ban against homosexuals simply *730 codifies society’s moral consensus that homosexuality is evil. Yet, even accepting arguendo this proposition that anti-homosexual animus is grounded in morality (as opposed to prejudice masking as morality), and assuming further that the Army is an appropriate governmental body to articulate moral norms, equal protection doctrine does not permit notions of majoritari-an morality to serve as compelling justification for laws that discriminate against suspect classes.

A similar principle animates Loving v. Virginia, 388 U.S. 1 , 87 S.Ct. 1817 , 18 L.Ed.2d 1010 (1967), in which the Supreme Court struck down a Virginia statute outlawing marriages between whites and blacks. Although the Virginia legislature may have adopted this law in the sincere belief that miscegenation — the mixing of racial blood lines — was evil, 33 this moral judgment could not justify the statute’s discrimination on the basis of race. Like the Army’s regulations proscribing sexual acts only when committed by homosexual couples, the Virginia statute proscribed marriage only when undertaken by mixed-race couples. In both cases, the government did not prohibit certain conduct, it prohibited certain conduct selectively — only when engaged in by certain classes of people. Although courts may sometimes have to accept society’s moral condemnation as a justification even when the morally condemned activity causes no harm to interests outside notions of morality, see Hardwick, 478 U.S. at 196 , 106 S.Ct. at 2847 (accepting moral condemnation as justification under rationality review), our deference to majoritarian notions of morality must be tempered by the equal protection principles which require that those notions be applied evenhandedly. Laws that limit the acceptable focus of one’s sexual desires to members of the opposite sex, like laws that limit one’s choice of spouse (or sexual partner) to members of the same race, cannot withstand constitutional scrutiny absent a compelling governmental justification. This requirement would be reduced to a nullity if the government’s assertion of moral objections only to interracial couples or only to homosexual couples could itself serve as a tautological basis for the challenged classification.

The Army’s remaining justifications for discriminating against homosexuals may not be illegitimate, but they bear little relation to the regulations at issue. For example, the Army argues that military discipline might be undermined if emotional relationships developed between homosexuals of different military rank. Army’s Opening Brief at 17-18,19 n. 9, 30; AR 635-200, II 15-l(a). Although this concern might be a compelling and legitimate military interest, the Army’s regulations are poorly tailored to advance that interest. No one would suggest that heterosexuals are any less likely to develop emotional attachments within military ranks than homosexuals. Yet the Army’s regulations do not address the problem of emotional attachments between male and female personnel, which presumably subjects military dicip-line to similar stress. Surely, the Army’s interest in preventing emotional relationships that could erode military discipline would be advanced much more directly by a ban on all sexual contact between members of the same unit, whether between persons of the same or opposite sex. Cf. Cleburne, 473 U.S. at 449-50 , 105 S.Ct. at 3259-60 (rejecting certain asserted justifications under rationality review where the justification would extend to other groups but the challenged classification did not). Here the Army’s regulations disqualify all homosexuals whether or not they have developed any emotional or sexual relationships with other soldiers.

Also bearing little relation to the regulations is the Army’s professed concern with breaches of security. AR 635-200, *731 ¶ 15-l(a). Certainly the Army has a compelling interest in excluding persons who may be susceptible to blackmail. It is evident, however, that homosexuality poses a special risk of blackmail only if a homosexual is secretive about his or her sexual orientation. The Army’s regulations do nothing to lessen this problem. Quite the opposite, the regulations ban homosexuals only after they have declared their homosexuality or have engaged in known homosexual acts. The Army’s concern about security risks among gays could be addressed in a more sensible and less restrictive manner by adopting a regulation banning only those gays who had lied about or failed to admit their sexual orientation. 34 In that way, the Army would encourage, rather than discourage, declarations of homosexuality, thereby reducing the number of closet homosexuals who might indeed pose a security risk. Moreover, even if banning homosexuals could lessen security risks, there appears to be no reason for treating homosexuality as a nonwaivable disqualification from military service while treating other more serious potential sources of blackmail as waivable disqualifications. See AR 635-200, II 14-12(c) & (d) (making drug abuse and the commission of other serious military offenses waivable disqualifications).

CONCLUSION

The Army’s regulations violate the constitutional guarantee of equal protection of the laws because they discriminate against persons of homosexual orientation, a suspect class, and because the regulations are not necessary to promote a legitimate compelling governmental interest. I would thus reverse the district court’s rulings denying Watkins’ motion for summary judgment and granting summary judgment in favor of the Army, and remand with instructions to enter a declaratory judgment that the Army Regulations A.R. 635-200, Chapter 15, and 601-280, ¶ 2-21(c), are constitutionally void on their face, and to enter an injunction requiring the Army to consider Watkins’ reenlistment application without regard to his sexual orientation.

. Because I would grant Watkins the relief he seeks on the basis of his equal protection claim, I need not address in this concurring opinion Watkins’ other constitutional claims involving the free speech clause, the petition clause, and the due process entrapment doctrine.

. The equal protection component of the Fifth Amendment imposes precisely the same constitutional requirements on the federal government as the equal protection clause of the Fourteenth Amendment imposes on state governments. See, e.g., Weinberger v. Wiesenfeld, 420 U.S. 636 , 638 n. 2, 95 S.Ct. 1225 , 1228 n. 2, 43 L.Ed.2d 514 (1975).

. In this opinion I use the term "sexual orientation" to refer to the orientation of an individual’s sexual preference, not to his actual sexual conduct. Individuals whose sexual orientation creates in them a desire for sexual relationships with persons of the opposite sex have a heterosexual orientation. Individuals whose sexual orientation creates in them a desire for sexual relationships with persons of the same sex have a homosexual orientation.

In contrast, I use the terms "homosexual conduct” and “homosexual acts” to refer to sexual activity between two members of the same sex whether their orientations are homosexual, heterosexual, or bisexual, and we use the terms "heterosexual conduct” and “heterosexual acts” to refer to sexual activity between two members of the opposite sex whether their orientations are homosexual, heterosexual, or bisexual.

Throughout this opinion, the terms “gay" and “homosexual” will be used synonymously to denote persons of homosexual orientation.

.Discriminations that burden some despised or politically powerless groups are so likely to reflect antipathy against those groups that the classifications are inherently suspect and must be strictly scrutinized. See, e.g., Plyler v. Doe, 457 U.S. 202 , 216 n. 14, 102 S.Ct. 2382 , 2394 n. 14, 72 L.Ed.2d 786 (1982). Such groups are generally termed “suspect classes.” The Supreme Court has identified other groups whose history of past discrimination entitles them to intermediate scrutiny protection under equal protection doctrine. Such groups are termed "quasi-suspect" classes. See generally, Nowak, Rotunda & Young, Constitutional Law, Ch. 16, § 1, at 593 (2d ed. 1983).

. AR 635-200 provides:

15-2 Definitions ...

a. Homosexual means a person, regardless of sex, who engages in, desires to engage in, or intends to engage in homosexual acts.

b. Bisexual means a person who engages in, desires to engage in, or intends to engage in homosexual and heterosexual acts.

c. A homosexual act means bodily contact, actively undertaken or passively permitted, between soldiers of the same sex for sexual satisfaction.

15-3 Criteria

The basis for separation may include preser-vice, prior service, or current service conduct or statements. A soldier will be separated per this chapter if one or more of the following approved findings is made:

a.The soldier has engaged in, attempted to engage in, or solicited another to engage in a homosexual act unless there are further approved findings that—

(1) Such conduct is a departure from the soldier’s usual and customary behavior; and

(2) Such conduct is unlikely to recur because it is shown, for example, that the act occurred because of immaturity, intoxication, coercion, or a desire to avoid military service; and

(3) Such conduct was not accomplished by use of force, coercion, or intimidation by the soldier during a period of military service; and

(4) Under the particular circumstances of the case, the soldier’s continued presence in the Army is consistent with the interest of the Army in proper discipline, good order, and morale; and

(5) The soldier does not desire to engage in or intend to engage in homosexual acts.

Note: To warrant retention of a soldier after finding that he or she engaged in, attempted to engage in, or solicited another to engage in a homosexual act, the board’s findings must specifically include all five findings listed in a(l) through (5) above. In making these additional findings, boards should reasonably consider the evidence presented. For example, engagement in homosexual acts over a long period of time could hardly be considered “a departure from the soldier’s usual and customary behavior.” The intent of this policy is to permit retention only of nonhomosexual soldiers who, because of extenuating circumstances (as demonstrated by findings required by para 15-3a(l) through (5)) engaged in, attempted to engage in, or solicited a homosexual act.

b. The soldier has stated that he or she is a homosexual or bisexual, unless there is a further finding that the soldier is not a homosexual or bisexual.

c. The soldier has married or attempted to marry a person known to be of the same biological sex (as evidenced by the external anatomy of the person involved) unless there are further findings that the soldier is not a homosexual or bisexual (such as, where the purpose of the marriage or attempt to marry was the avoidance or termination of military service).

AR 635-200, ¶¶ 15-2 & 15-3 (emphasis in original).

Although it is the Army’s refusal to reenlist Watkins because of his homosexuality that is directly at issue, Watkins’ challenge to the Army’s regulation on discharge is relevant to this appeal for two reasons: (1) persons being validly discharged for homosexuality at the time *714 of reenlistment, as Watkins was, cannot reenlist under 601-280 ¶ 2-21(k); (2) enjoining the Army to consider Watkins’ reenlistment application without regard to his homosexuality will provide no effective relief if he would be subject to mandatory discharge because of homosexuality as soon as he was reenlisted. I thus consider Watkins’ challenge to the constitutionality of the Army’s discharge regulation as well as its reenlistment regulation.

. In stark contrast to the breadth and focus of the regulations, the only statute Congress has enacted regulating the private consensual sexual activity of military personnel covers only sodomy, not other forms of sexual conduct, and covers sodomy whether engaged in by homosexuals or heterosexuals. 10 U.S.C. § 925 (1982) provides:

(a) Any person subject to this chapter who engages in unnatural carnal copulation with another person of the same or opposite sex or with an animal is guilty of sodomy. Penetration, however slight, is sufficient to complete the offense.

(b) Any person found guilty of sodomy shall be punished as a court-martial may direct.

Although the statute does not define "sodomy” or “unnatural carnal copulation,” the statute does require proof of “penetration," which apparently limits sodomy to oral and anal copulation. See United States v. Harris, 8 M.J. 52, 53-59 (C.M.A.1979).

The Army has never made a finding that Watkins ever engaged in an act of sodomy in violation of section 925. Indeed, the Army twice investigated Watkins for allegedly committing sodomy in violation of section 925 and had to drop both investigations because of “insufficient evidence.”

. This reading of the regulations is supported by the Army’s treatment of Watkins himself. The only evidence that Watkins ever engaged in homosexual conduct is a statement he made during a 1968 investigation that he committed homosexual acts with two other servicemen. When these two servicemen denied engaging in homosexual acts with Watkins, the Army discontinued the investigation without making a finding that Watkins had committed homosexual acts. The Army did not decide to discharge Watkins (and deny him reenlistment) until 1981. In the meantime, Watkins openly and repeatedly acknowledged his homosexual orientation without admitting to any homosexual acts. It strains credulity to think that the Army *716 decided to discharge Watkins and deny him reenlistment solely on the basis of his contradicted statement in 1968 that he had committed homosexual acts. Plainly it is Watkins' homosexual orientation — rather than evidence of any conduct — that explains the Army’s decision to end Watkins' Army career.

. Under the Court’s analysis, because the Constitution’s protection of the right to privacy does not extend to homosexual sodomy, a judgment by the state that sodomy is immoral provides a sufficiently rational basis for sodomy laws to satisfy the requirements of substantive due process. See Hardwick at 196 , 106 S.Ct. at 2846.

. See also Hardwick, 478 U.S. at 201 , 106 S.Ct. at 2849 (Blackmun, J., dissenting) (Court “refused to consider" equal protection clause); Doe v. Casey, 796 F.2d 1508, 1522 (D.C.Cir.1986), aff’d in part, rev'd in part sub. nom, Webster v. Doe, — U.S. -, 108 S.Ct. 2047 , 100 L.Ed.2d 632 (1988) ("Although ... the Supreme Court’s recent decision in Bowers v. Hardwick [held] that homosexual conduct is not constitutionally protected, the Court did not reach the different issue of whether an agency of the federal government can discriminate against individuals merely because of sexual orientation." (Footnotes omitted and emphasis in the original.)); Swift v. United States, 649 F.Supp. 596 , 42 FEP Cases (BNA) 787, 790 (D.D.C.1987) ("this Circuit has declined to read [Hardwick ] as barring claims of discrimination based on sexual preference”); but cf. Padula v. Webster, 822 F.2d 97 (D.C.Cir.1987) (“reasoning in Hardwick forecloses ... suspect class status for practicing homosexuals”).

. One commentator and one district court have already agreed with Watkins //that the conduct-orientation dichotomy is a valid way of distinguishing Watkins’ case from Hardwick . As Professor Sunstein has written, "this feature [the conduct/orientation distinction] serves to distinguish [Watkins from] Hardwick in a persuasive way....” Sunstein, Sexual Orientation and the Constitution: A Note on the Relationship Between Due Process and Equal Protection, 55 U.Chi.L.Rev. 1161, 1162 n. 9 (1988).

In BenShalom v. Secretary of Army, 703 F.Supp. 1372 (1989), the District Court for the Eastern District of Wisconsin prevented the Army from denying reenlistment to Sergeant BenShalom under the same regulations Watkins challenges. The district court based this decision on both the First Amendment and the equal protection component of the Fifth Amendment. In analyzing BenShalom’s equal protection claim, the district court tracked the equal protection analysis of Watkins II, relying heavily on the conduct/orientation distinction.

. Thus, whether the Army’s regulations are "conduct-based” or "orientation-based,” Hard-wick cannot be read to foreclose Watkins’ equal protection claim. Professor Sunstein agrees, noting that "Hardwick ... was interpreted correctly in the majority opinion in Watkins [II], and misread in ... Judge Reinhardt’s opinion in Watkins [II].... [Because Hardwick involved due process rather than equal protection], Watkins can be distinguished from Hardwick even if the former decision were to be applied to a class of people including some, many or all who engage in the conduct at issue in Hardwick.” Sunstein, supra note 10, at 1162 & n. 9.

. See Hardwick, 478 U.S. at 188 n. 2, 106 S.Ct. at 2842 n. 2.

. "The only claim properly before the Court ... is Hardwick’s challenge to the Georgia statute as applied to consensual homosexual sodomy. We express no opinion on the constitutionality of the Georgia statute as applied to other acts of sodomy." Hardwick, 478 U.S. at 188 n. 2, 106 S.Ct. at 2842 n. 2. (emphasis added).

. See Hardwick, 478 U.S. at 191-92 , 106 S.Ct. at 2844 (quoting Palko v. Connecticut, 302 U.S. 319, 325 , 58 S.Ct. 149, 152 , 82 L.Ed. 288 (1937) and Moore v. East Cleveland, 431 U.S. 494, 503 , 97 S.Ct. 1932, 1937 , 52 L.Ed.2d 531 (1977) (Opinion of Powell, J.)).

. See also Anne Goldstein, History, Homosexuality, and Political Values: Searching for the Hidden Determinants of Bowers v. Hardwick 97 Yale LJ. 1073, 1084-85 (1988) (state laws relied upon by majority outlawed all sodomy, whether homosexual or heterosexual). Moreover, Congress has not distinguished between heterosexual and homosexual sodomy in proscribing acts of sodomy by members of the armed forces. See supra note 6.

. This example is loosely drawn from Moore v. City of East Cleveland, 431 U.S. 494 , 97 S.Ct. 1932 , 52 L.Ed.2d 531 (1977).

. See Moore, 431 U.S. at 509 , 97 S.Ct. at 1940 (Brennan, J., concurring) (indicating this was the case in East Cleveland).

. I should make clear that this was not shown to be the case in Moore. See 431 U.S. at 510 , 97 S.Ct. at 1941 (Brennan, J., concurring).

. See Moore, 431 U.S. at 505-06 , 97 S.Ct. at 1938-39 (plurality opinion).

. See 431 U.S. at 500 , 97 S.Ct. at 1936 (plurality opinion) (quoting city’s argument).

. Professor John Hart Ely, for example, has severely criticized the Supreme Court’s substantive due process analysis in Roe v. Wade, 410 U.S. 113 , 93 S.Ct. 705 , 35 L.Ed.2d 147 (1973), while at the same time expressing the view that governmental classifications burdening homosexuals merit heightened scrutiny under the equal protection clause. Compare J. Ely, Democracy and Distrust 248 n. 52 (1980), with id. at 162-64 .

. See generally J. Ely, supra note 21, at 101-02 (“unlike an approach geared to the judicial imposition of ‘fundamental values,' the representation-reinforcing [approach] ... is not inconsistent with, but to the contrary is entirely supportive of, the American system of representative democracy. It recognizes the unacceptability of the claim that appointed and life-tenured judges are better reflectors of conventional values than elected representatives, devoting itself instead to policing the mechanisms by which the system seeks to ensure that our elected representatives will actually represent.”).

. Under equal protection doctrine, heightened scrutiny not only applies to legal classifications that burden suspect or quasi-suspect classes but also applies to classifications that burden the exercise of fundamental or important substantive rights to engage in certain conduct. See, e.g., Plyler v. Doe, 457 U.S. 202 , 216-17 & nn. 14-15, 102 S.Ct. 2382 , 2394-95 & nn. 14-15, 72 L.Ed.2d 786 (1982); Maher v. Roe, 432 U.S. 464, 470-78 , 97 S.Ct. 2376, 2380-85 , 53 L.Ed.2d 484 (1977); L. Tribe, American Constitutional Law § 16-7, at 1002-03, § 16-31, at 1089-90 & n. 1 (1978).

. If Hatheway had decided that homosexuals do not constitute a suspect class, I would vote to have this en banc panel overrule it.

. Along with subsequent cases, DeSantis has established that there are only two ways of making this showing under § 1985(3): (1) proving that Congress has enacted statutes offering special protection to the class; or (2) proving that courts have offered special protection to the class by designating it a suspect or quasi-suspect class. Id., see also Schultz v. Sundberg, 759 F.2d 714, 718 (9th Cir.1985).

. The Fifth and Tenth circuits have also considered this question. Baker v. Wade, 769 F.2d 289, 292 (5th Cir.1985) (en banc), (stressing that statute at issue was "directed at certain conduct, not at a class of people”), cert. denied, 478 U.S. 1022 , 106 S.Ct. 3337 , 92 L.Ed.2d 742 (1986); National Gay Task Force v. Board of Educ., 729 F.2d 1270 , 1273 (10th Cir.1984) (statute at issue proscribes "public homosexual activity” by teachers), aff’d without opinion by an equally divided Court, 470 U.S. 903 , 105 S.Ct. 1858 , 84 L.Ed.2d 776 (1985). Both of these circuits held that discrimination based on homosexual conduct does not merit heightened scrutiny under the equal protection clause, but neither circuit attempted any serious analysis of the issue. See Baker v. Wade, 769 F.2d at 292 (noting merely that the plaintiff "has not cited any cases holding, and we refuse to hold, that homosexuals constitute a suspect or quasi-suspect classification”); National Gay Task Force, 729 F.2d at 1273 (stating summarily that classification based on choice of sexual partners could not be suspect because Supreme Court has not held gender to be a suspect classification); see also Rich v. Secretary of the Army, 735 F.2d 1220, 1229 (10th Cir.1984) (citing without explanation National Gay Task Force, Hatheway, and DeSan-tis for the proposition that a “classification based on one’s choice of sexual partners is not suspect").

. One district court has decided the question. See supra n. 10.

. See Army’s Second Supplemental Brief at 10.

. Because homosexuals are a minority and are frequently excluded from jobs, schools, churches, and heterosexual social circles, see supra, heterosexuals generally have relatively few opportunities to meet homosexuals and overcome their stereotypical thinking about homosexuality.

. The Army claims that homosexuals cannot be politically powerless because two states, Wisconsin and California, have passed statutes prohibiting discrimination against homosexuals. Two state statutes do not overcome the long and extensive history of laws discriminating against homosexuals in all fifty states. See, e.g., Note, supra, 57 S.Cal.L.Rev. at 803-07. Moreover, at the national level — the relevant political level for seeking protection from military discrimination — homosexuals have been wholly unsuccessful in getting legislation passed that protects them from discrimination.

The Army also argues that the repeal of sodomy statutes by many states proves that homosexuals are not politically powerless. However, sodomy statutes restrict the sexual freedom of heterosexuals as well as homosexuals. The repeal of sodomy statutes may thus reflect the liberalization of attitudes about heterosexual behavior more than it reflects the political power of homosexuals.

. A somewhat different rationale conceivably could also underlie certain cryptic statements the Army makes about its concerns regarding “close conditions affording minimal privacy,” “ 'potential for difficulties arising out of possible close confinement,'" and “the intimacy of barrack's life.” AR 635-200, ¶ 15-l(a); Army’s Opening Brief at 15 (quoting Belter, 632 F.2d at 812); Army’s Second Supp. Brief at 19 n. 9, 30. Conceivably, the Army could be concerned in part that the presence of gays in the ranks will create sexual tensions — as distinguished from tensions arising from prejudice — because of the practical necessity of housing gays with personnel of the same sex. The Army, however, never articulates this concern. Thus it gives no indication that it regards this concern as compelling or that it believes that weeding all homosexuals out of the military — even soldiers as exemplary as Sergeant Watkins — is necessary to advance a compelling military interest in reducing sexual tensions. Indeed, at points in its argument the Army implies that it is concerned about the close confinement of soldiers only insofar as such confinement might exacerbate hostilities and tensions assertedly created by the prejudice some heterosexuals have against homosexuals. See Army’s Opening Brief at 17, 31 n. 18. Even if the Army had raised the argument that excluding homosexuals from barracks reduces *729 sexual tension and had shown that reducing sexual tension serves a compelling interest, nothing in the record even suggests that a per se rule banning all homosexuals from the Army would be the least restrictive method of advancing this interest.

. It took an Executive Order in 1945 by President Truman, issued against the advice of almost every admiral and general, to integrate our armed forces. M. Miller, Plain Speaking: An Oral Biography of Harry S. Truman 79 (1983). It is also interesting to note that during World War II the Army deliberately minimized any publicity about the existence of black soldiers because it feared that such publicity would tarnish the Army’s public image. See G. Ware, supra, at 100.

. Indeed, the trial judge in Loving admonished Mildred and Richard Loving that interracial marriage was a violation of the Christian ethic of racial purity: “Almighty God created the races, white, black, yellow, malay and red, and he placed them on separate continents. And but for the interference with his arrangement there would be no cause for such marriages. The fact that he separated the races shows that he did not intend for the races to mix. Loving, 388 U.S. at 3 , 87 S.Ct. at 1819 .

. Watkins has forthrightly reported his homosexuality since his induction in 1967, and his homosexuality was always a matter of common knowledge. There is no suggestion in the record before us that Watkins ever feared public disclosure of his homosexualty.