dissenting:
The Navy discharged Petty Officer Mark A Philips because he said that he is gay and that he engaged in private, off-base, consensual same-sex sexual activity. Because I conclude the military’s policy of differentiating between the private sexual activities of its heterosexual and homosexual service members is not rationally related to a legitimate government interest, I respectfully dissent.
*1433I.
The military’s “don’t ash/don’t tell” policy is not directed to prohibiting public homosexual acts or sexual misconduct by gay men and lesbians. Elaborate military regulations and the Uniform Code of Military Justice (“UCMJ”), 10 U.S.C. §§ 836-940, already punish sexual misconduct such as sexual harassment, sexual assault, fraternization, and indecent exposure, and sodomy is prohibited under the UCMJ for both heterosexual and homosexual service members. 10 U.S.C. § 925. However, a simple statement that one is gay, and honest answers to questions about one’s sexual activity, can mandate discharge under the policy.
Philips contends that the policy violates his right to equal protection by subjecting him to mandatory discharge for engaging in sexual acts with a man while a heterosexual man would not be discharged for engaging in the same sexual acts with a woman. He argues that the only basis for the policy is the military’s desire to accommodate the biases and prejudices of heterosexual service members and that there is no legitimate reason for treating heterosexual service members differently than homosexual service members. The majority concludes that the policy is constitutional because it advances legitimate military goals. As Philips was separated for engaging in same-sex sexual activity (as well as for stating that he is gay), we must review the constitutionality of the policy as it relates to the differential treatment of same-sex sexual activity and opposite-sex sexual activity.1
Everyone agrees that the current policy treats homosexual service members and heterosexual service members differently. A statement by a gay service member that he is gay will result in discharge, whereas a statement by a heterosexual service member that he is heterosexual will not. A statement by a gay service member- that he has had sexual relations with a member of the same sex will result in discharge, whereas a statement by a heterosexual service member that he has had sexual relations with a member of the opposite sex will not. Evidence that a service member has engaged in sexual activity with a member of the same sex will result in discharge,2 whereas evidence that a service member has engaged in the same sexual activity with a member of the opposite sex will not. It is clear that Philips would not have been discharged had his sexual partners been women rather than men.
That homosexual and heterosexual service members are treated differently does not in itself render the policy unconstitutional. Where a law neither burdens a fundamental right nor targets a suspect class, the courts will uphold the differential classification “so *1434long as it bears a rational relation to some legitimate end.” Romer v. Evans, — U.S. -, -, 116 S.Ct. 1620, 1627, 134 L.Ed.2d 855 (1996) (applying meaningful rational-basis review to Colorado’s discriminatory classification of gays and lesbians); High Tech Gays v. Defense Indus. Sec. Clearance Office, 895 F.2d 563, 574 (9th Cir.1990). While rational-basis review “is not a license for courts to judge the wisdom, fairness, or logic of legislative choices,” Heller v. Doe, 509 U.S. 312, 319, 113 S.Ct. 2637, 2642, 125 L.Ed.2d 257 (1993) (quoting FCC v. Beach Communications, Inc., 508 U.S. 307, 313, 113 S.Ct. 2096, 2101, 124 L.Ed.2d 211 (1993)), neither is it an abdication of the court’s responsibility to strike down arbitrary classifications.
[E]ven in the ordinary equal protection case calling for the most deferential of standards, we insist on knowing the relation between the classification adopted and the object to be attained. The search for the link between classification and objective gives substance to the Equal Protection Clause; it provides guidance and discipline for the legislature, which is entitled to know what sorts of laws it can pass; and it marks the limits of our own authority.
Romer, — U.S. at -, 116 S.Ct. at 1627 (critically examining the government’s proffered rationales and the relationship between means and ends). We must examine the possible justifications for the policy in light of the factual context in the record. See Cleburne v. Cleburne Living Center, 473 U.S. 432, 448, 105 S.Ct. 3249, 3258, 87 L.Ed.2d 313 (1985) (“Because in our view the record does not reveal any rational basis for believing that the Featherston home would pose any special threat to the city’s legitimate interests, we affirm the judgment below insofar as it holds the ordinance invalid as applied in this case.”) (emphasis added); Heller, 509 U.S. at 321, 113 S.Ct. at 2643 (“[E]ven the standard of rationality ... must find some footing in the realities of the subject addressed by the legislation.”); Pruitt v. Cheney, 963 F.2d 1160, 1166 (9th Cir.1992) (“[I]n High Tech Gays, upon plaintiffs’ showing of discrimination, we required the government to establish on the record that its policy had a rational basis.”), cert. denied, 506 U.S. 1020, 113 S.Ct. 655, 121 L.Ed.2d 581 (1992).
II.
In order to survive rationality review, there must be a legitimate reason why homosexual but not heterosexual conduct is barred: The majority, with circular logic, implicitly concludes only that “homosexual conduct” is prohibited in order to prohibit “homosexual conduct.”3 The primary justification proffered for the “don’t ask/don’t tell” *1435policy is “unit cohesion.” “Good morale,” “discipline,” and the ability to recruit and retain military personnel are related sub-interests. While unit cohesion is surely a legitimate government interest, the current policy simply does not farther this interest in a rational and reasonably related way. There is no reason to believe that engaging in private, consensual, off-base sexual activity with a member of the same sex somehow makes one a worse soldier than engaging in the same conduct with a member of the opposite sex. Indeed, government witnesses testified that gay soldiers are as effective as heterosexual soldiers. The government has acknowledged that gay and lesbian service members are no less able to perform their duties. Thus, “unit cohesion” is not harmed by gay service members’ ability to contribute to their unit.4
The only way, then, that “unit cohesion” could conceivably be affected by the presence of gay men and lesbians in the military is by the negative reactions of service members opposed to homosexuality. This is evident from the structure of the new prohibitions.5 If a service member keeps his homosexual orientation secret, then he is allowed to remain in the military. However, if a service member acknowledges that he is gay, then he is a threat to “unit cohesion” and must be discharged. The only material difference in these two situations is that information regarding the service member’s homosexuality has been- communicated to other service members, who might react negatively to the information and threaten unit cohesion.
Even assuming that such a reaction would occur,6 its accommodation is not a legitimate *1436government interest. Disapproval of homosexuality on the part of heterosexual service members is an impermissible reason for discriminating against gay service members. See Romer, — U.S. at -, 116 S.Ct. at 1627-29 (implicitly rejecting dissent’s contention that moral disapproval of homosexuality creates a legitimate state interest justifying discrimination against homosexuals). Otherwise, discrimination against an unpopular class could always be justified by reference to the moral disapproval of the majority-
The goal of promoting “unit cohesion” is illegitimate if it is based solely on biases that cannot be tolerated under the laws. “Private biases may be outside the reach of the law, but the law cannot, directly or indirectly, give them effect.” Palmore v. Sidoti, 466 U.S. 429, 433, 104 S.Ct. 1879, 1882, 80 L.Ed.2d 421 (1984).7 “[M]ere negative attitudes ... are not permissible bases” for disfavoring a class of citizens. Cleburne, 473 U.S. at 448, 105 S.Ct. at 3259. “ ‘[I]f the constitutional conception of ‘equal protection of the laws’ means anything, it must at the veiy least mean that a bare ... desire to harm a politically unpopular group cannot constitute a legitimate governmental interest.’ ” Romer, — U.S. at-, 116 S.Ct. at 1628 (quoting United States Dep’t of Agric, v. Moreno, 413 U.S. 528, 534, 93 S.Ct. 2821, 2826, 37 L.Ed.2d 782 (1973)). Just as the desire to accommodate other citizens’ personal or religious objections to homosexuality did not suffice to uphold Amendment 2, see Romer, — U.S. at -, 116 S.Ct. at 1629, the desire to protect a child from the racial prejudices of others did not provide a legitimate reason for favoring a same-race couple over an interracial couple in awarding custody of a child, see Palmore, 466 U.S. at 433, 104 S.Ct. at 1882, and the negative attitudes of property owners did not provide a legitimate justification for allowing rejection of a zoning permit for a home for mentally retarded individuals, see Cleburne, 473 U.S. at 448, 105 S.Ct. at 3258-59, the desire to accommodate the attitudes of heterosexual service members opposed to homosexuality does not provide a legitimate reason for excluding gay men and lesbians from the military.8
The military’s “retention and recruitment,” “morale,” and “discipline” justifications are similarly invalid. The only way the military’s exclusion of gay men and lesbians conceivably furthers “retention and recruitment” is by accommodating the attitudes of service members opposed to homosexuality who refuse to serve in the military if gay men and lesbians are permitted to serve. Similarly, the only way the military’s exclusion of gay men and lesbians conceivably furthers “morale” and “discipline” is by accommodating the negative attitudes of those service members who oppose having gay men and lesbians in their ranks. These are not legitimate government interests.
*1437The remaining justifications proffered for the exclusion of gay men and lesbians are “privacy” and “sexual tension.” At oral argument, the government asserted that “privacy” refers to the military’s interest in protecting the privacy of heterosexual service members, whereas “sexual tension” refers to the military’s interest in avoiding sexual attraction of homosexual service members to heterosexual members.
Risk of “sexual tension” as a rationale is not reasonable. The military has no interest in regulating the private feelings and thoughts of its service members: it acknowledges that sexual orientation is a personal and private matter not subject to regulation. Further, the military’s professed interest in avoiding “sexual tension” is not rationally furthered by the “don’t ask/don’t tell” policy. As one district court judge has stated, “[T]he military’s regulations permit homosexuals to serve so long as they, in effect, remain celibate and conceal their sexual orientation. The ‘sexual attractions’ and ‘tension’ which the Federal defendants speculate will result from the presence of homosexuals will, therefore, not be curtailed under the current policy.” Holmes, 920 F.Supp. at 1531. It is irrational to suggest that service members who acknowledge their homosexuality or who engage in private, off-base same-sex sexual activities will create or be subject to greater “sexual tensions” than closeted gay service members or heterosexual service members.9 Thus, differential classification on this basis is neither rational nor legitimate.
While concerns of “privacy” have a greater surface plausibility, they too do not survive rational basis scrutiny. If by “privacy,” the military means service members’ desire not to associate with gay men and lesbians, this is an impermissible interest under Cleburne, Pruitt, Palmore, and Romer.10 If by privacy, the military means keeping heterosexual service members’ bodies from the gaze of gay and lesbian service members, the policy is also invalid. As in Romer, the “breadth of the [policy] is so far removed from th[is] particular justification that we find it impossible to credit” the justification. Romer, — U.S. at -, 116 S.Ct. at 1629. Openly gay men and lesbians are not permitted to hold any position in the military—whether lawyer, clerk, nurse, or pilot—regardless of how far removed from being stationed in close quarters in a combat situation. And, as service members are already expected to deal with greatly reduced privacy in the military, it is difficult to believe that a concern for service members’ privacy is sufficient reason for disqualifying an entire class of citizens from serving in the military.
Even crediting the justification in the abstract, it is not rationally furthered by the current policy. Under the policy, “secret” homosexuals are permitted to serve; only “open” homosexuals are not.11 To the extent *1438the military purports to be concerned with protecting heterosexual service members from sleeping and showering with persons who find them sexually attractive, the policy does not rationally further this interest. As one district court judge has put it,
[I]f indeed there are homosexuals who wish to peek at naked bodies, they might do so quite as readily when their orientation is a secret as when it is open. The only difference will be that heterosexuals will not know which of their servicemates are homosexuals, and heterosexuals will have reason to have a generalized suspicion of everyone in the showers, hardly a circumstance likely to increase “cohesion.” To suggest to heterosexuals that the secrecy policy will “accommodate” their privacy interests is to attempt to mislead them.
Able v. United States, 880 F.Supp. 968, 978 (E.D.N.Y.1995), vacated and remanded on other grounds, 88 F.3d 1280 (2nd Cir.1996); accord Holmes, 920 F.Supp. at 1631. Similarly, the fact that a service member has engaged in private, off-base, same-sex sexual activity in no way makes him more of a “threat” to .heterosexual service members’ privacy than a service member with a secret homosexual “orientation”—-he is no more likely to stare at heterosexual service members’ naked bodies or be attracted to them. While the policy need not be perfectly tailored to the asserted government interest to withstand constitutional scrutiny, there must be some rational basis for distinguishing between the disfavored and favored classes. Here, there is none. Accord Holmes, 920 F.Supp. at 1531.
III.
The majority relies heavily upon our prior cases upholding challenges to military policies discharging service members who en— gage in sexual relations with members of the same sex for support of its holding that the current policy does not violate the Equal Protection Clause. While these eases are instructive, they must be analyzed discretely as to their effect on the constitutionality of the new policy. In addition to being undercut by subsequent Supreme Court decisions holding that accommodation of individuals’ disapproval of an unpopular group on moral or religious grounds or bias of any kind does not constitute a legitimate government interest, see Romer, — U.S. at -, 116 S.Ct. at 1628; Cleburne, 473 U.S. at 448, 105 S.Ct. at 3258-59; Palmore, 466 U.S. at 438, 104 S.Ct. at 1882, these earlier cases are premised on different policy justifications and definitions of proscribed conduct than those that underlie the current policy. These cases are founded in large part on acceptance of the military’s justification of avoiding tension “between known homosexuals and other members who ‘despise/detest homosexuality.’ ” See Pruitt, 963 F.2d at 1163 (quoting Belter v. Middendorf, 632 F.2d 788, 811 (9th Cir.1980), cert. denied, 452 U.S. 905, 101 S.Ct. 3030, 69 L.Ed.2d 405 (1981)). As explained above, Romer, Palmore, and Cle-burne undercut this rationale, holding that accommodation of the biases of others is not a legitimate government interest.
Hatheway and Schowengerdt12 rely on Better13 for the proposition that the military has a legitimate interest in discharging those who engage in homosexual acts in order to avoid tension between gay and lesbian service members and other members who “despise/detest homosexuality.” Beller, 632 F.2d at 811. As such, the reasoning of each of these eases is undercut by subsequent holdings in Romer, Cleburne, and Palmore. Cf. Pruitt, 963 F.2d at 1165 (noting that Better and Hatheway cannot be given unexa*1439mined effect after Palmore and Cleburne). Meinhold, similarly, merely cites Better, Hatheway, and Schowengerdt for the proposition that the military can discharge a service member for homosexual conduct. 34 F.3d at 1477, n. 7. To the extent Meinhold distinguishes between disapproval of homosexual “status” and disapproval of homosexual “conduct,” its reasoning is similarly unavailing post-ifomer.
IV.
Finally, that the purported justifications arise in the military context does not endow them with greater legitimacy. Although Romer, Cleburne, and Palmore arose in the civilian context, I cannot accept Judge Noo-nan’s proposition that the desire to accommodate animosity towards a disfavored group is more acceptable in the military context than in the civilian context. See supra (Noonan, J., concurring). Judge Noonan puts the Constitution in his hip pocket and brushes its commands aside when he says that “not every provision of the Constitution can be enforced across the board within the military.” Supra at 1432 (Noonan, J., concurring).
As courts and commentators have noted, the “unit cohesion” rationale proffered in support of the “don’t ask/don’t tell” policy is disturbingly similar to the arguments used by the military to justify the exclusion from and segregation of African Americans in military service. “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.” Watkins, 875 F.2d at 729 (Norris, J., concurring); see also Thomasson, 80 F.3d at 950 (Hall, J., dissenting); Holmes, 920 F.Supp. at 1533; Kenneth L. Karst, The Pursuit of Manhood and the Desegregation of the Armed Forces, 38 UCLA L.Rev. 499, 572-74 (1991) (noting parallels between arguments made in support of the segregation and exclusion of African Americans from the military and in support of the exclusion of gay men and lesbians). Despite the deference due the military, there is no doubt that were the government today to exclude African-Americans from the military, the courts would easily reject the military’s assertions of “unit cohesion,” “morale,” and “discipline” and strike down the policy as violative of equal protection. While racial classifications are subject to a stricter level of scrutiny, these asserted interests, which are based on animosity towards the disfavored class, are no more acceptable when used to support the exclusion of gay men and lesbians from the military.
The dangers of unquestioning deference to the military are demonstrated by decisions such as Korematsu v. United States, 323 U.S. 214, 65 S.Ct. 193, 89 L.Ed. 194 (1944), which upheld the internment of Japanese Americans during World War II against constitutional challenge. As commentators and courts have noted, “[t]hat decision, justified by deference to the military’s race-based judgment about the threat posed by Japanese Americans, is one of the Court’s most embarrassing moments, and has been thoroughly repudiated by history.” David Cole & William N. Eskridge, Jr., From Hand-Holding to Sodomy: First Amendment Protection of Homosexual (Expressive) Conduct, 29 Harv.C.R.-C.L.L.Rev. 319, 343 (1994). These inexcusable moments in our history should give us pause whenever deference to the military is urged in support of discrimination against a class of citizens.
Judge Noonan states explicitly what the majority holds implicitly—that “[w]e lack the constitutional capacity to substitute our commands for the military’s.” Supra at 1432 (Noonan, J., concurring). Although claiming only ‘deference’ to the military, he ultimately negates our ability to review any military decision despite its constitutional infirmity: “It is not the task of the judiciary to second guess when competent military officers conclude that a given practice is necessary for the good of the service.” Id. However, the military is not above the constitution. See Parisi v. Davidson, 405 U.S. 34, 55, 92 S.Ct. 815, 826-27, 31 L.Ed.2d 17 (1971) (“When the military steps over [the] bounds [of civil liberties], it leaves the area of its expertise and forsakes its domain. The matter then becomes one for civilian courts to resolve, con*1440sistent with the statutes and with the Constitution.”) (Douglas, J., concurring) (citations omitted). While “judicial deference ... is at its apogee when legislative action under the congressional authority to raise and support armies and make rules and regulations for their governance is challenged,” Goldman v. Weinberger, 475 U.S. 503, 508, 106 S.Ct. 1310, 1313, 89 L.Ed.2d 478 (1986) (quoting Rostker v. Goldberg, 453 U.S. 57, 70, 101 S.Ct. 2646, 2655, 69 L.Ed.2d 478 (1981)), “[w]e of course do not abdicate our ultimate responsibility to decide the constitutional question.” Rostker, 453 U.S. at 67, 101 S.Ct. at 2653.
If not for the military context, there is no doubt that Philips’ discharge pursuant to the “don’t ask/don’t tell” policy would violate equal protection. Even taking into account the unique military context, I see no legitimate government interest that the policy rationally furthers. Gay men and lesbians are no less able and no more prone to misconduct than their heterosexual counterparts. There is nothing about engaging in private, off-base, same-sex conduct that harms the military more than engaging in private, off-base, opposite-sex conduct. The only conceivable rationales for the exclusionary policy—unit cohesion, morale, and privacy—are dependent not on the capabilities of gay men and lesbians as service members but on the negative reactions of other service members. Even in the military context, an entire class may not be singled out for disfavored treatment because of animosities towards that class.
While certain internal military matters are totally exempt from judicial review, the military does not contend that the “don’t ask/don’t tell policy” is such a matter. Indeed, the validity of regulations similar to those before us were reviewed by our court in Meinhold, Pruitt, and Better. Thus, “deference” to the military cannot mean that we may not subject the “don’t ask/don’t tell” policy to meaningful judicial review. Deference does not mean unquestioningly accepting the military’s asserted justifications.14 Judicial deference to the military simply means sensitivity to the special circumstances of the military and appropriate respect for their particular role as our protectors, not abdication of our role as adjudicators of constitutional claims. See Knutson v. Wisconsin Air National Guard, 995 F.2d 765, 769 (7th Cir.1993). In Goldman, the controlling concurrence of Justice Stevens deferred to the military’s interest in uniformity of dress precisely because the rule at issue was “based on a neutral, completely objective standard” and was “not motivated by hostility against” any particular group, however defined. 475 U.S. at 513, 106 S.Ct. at 1316. The same cannot be said here.
“Inequality is inherent in the structure of the military services.” Supra at 1430 (Noo-nan, J., concurring). I agree that there are ranks and privileges attendant upon them, but that does not mean that the military may allow unequal access to those ranks without a permissible rational basis for doing so. Military regulations cannot provide, for example, that only white males can aspire to ranks above captain in the Army or Air Force or lieutenant in the Navy. The military now concedes that competence of homosexuals is not an issue and that aptitude for military service is not an issue. It concedes equivalence with heterosexuals on all tests and scales that matter. By the military’s own regulations closeted homosexuals are not disqualified from service. Why may acknowledged or practicing homosexuals not serve? Only because some heterosexuals may harbor prejudices against them.
The military has carved out a particular class and subjected it to adversely discriminatory treatment. The “don’t ask/don’t tell” policy “inflicts on [gay men and lesbians] immediate, continuing, and real injuries that *1441outrun and belie any legitimate justifications that may be claimed for it.” Romer, — U.S. at -, 116 S.Ct. at 1629. The policy, pursuant to which Philips, an exemplary public servant, was discharged, violates equal protection. Philips’ service record illustrates the irrationality of the policy. I respectfully dissent.
. Because Philips was discharged for engaging in "homosexual acts” within the meaning of the statute, I use the terms "homosexual,” "gay,” and "lesbian” to refer to individuals who engage in same-sex sexual activity. I note, however, that identifying oneself as gay or lesbian need not involve sexual activity. "Gay,” "lesbian," and "homosexual” can refer to an individual's sexual and emotional attraction to members of the same sex regardless of sexual activity. That the policy treats same-sex conduct differently depending on whether or not an individual has a "propensity" to engage in homosexual conduct, 10 U.S.C. § 654(b)(1)(E), suggests that the military’s concern is with gay and lesbian identity' rather than specific sexual acts. Although the military purports to be concerned only with homosexual conduct, even a service member’s statement that he is gay requires discharge unless the service member can "demonstrate that he or she is not a homosexual as defined in the statute.” S. Rpt. No. 103-112, at 294. "[A] member cannot rebut the presumption simply through a promise to adhere to military standards of conduct in the future; nor can the member rebut the presumption by a statement to the effect that he or she has a propensity towards homosexuality but has not acted on it.” Id.
. In the foregoing situations, a’ service member who states that he is gay or who engages in same-sex acts may avoid discharge only if he can demonstrate, among other things, that despite his statements or acts, he does not have a "propensity” to engage in homosexual acts. 10 U.S.C. § 654(b)(1)(E) & (2). Thus, presumably, if a service member can demonstrate that he was only joking about being gay, or that he is really heterosexual despite a one-time same-sex encounter, he can avoid discharge. Of course, such a means of rebutting the presumption is of little comfort to a service member who is in fact gay. For the reasons discussed infra that the military's discrimination against service members on the basis of "homosexual acts” cannot survive equal protection scrutiny, neither can the military’s discrimination on the basis of “homosexual propensity” survive equal protection scrutiny.
. As sodomy and sexual misconduct are already prohibited by the UCMJ for both heterosexual and homosexual service members, and as the government acknowledges that homosexual service members are no more likely to engage in sexual misconduct than heterosexual service members, “homosexual conduct" cannot be used as proxy for sexual misconduct and the policy cannot rationally be justified as a way of deterring sexual misconduct. The Romer Court implicitly but necessarily rejected the argument that discrimination on the basis of homosexuality does not violate equal protection simply because Bowers v. Hardwick has held that the criminalization of homosexual sodomy does not offend due process. Romer, — U.S. at -, 116 S.Ct. at 1627-29; id. at -, 116 S.Ct. at 1629 (Scalia, J., dissenting) (criticizing the majority for this reason); see also Watkins v. United States Army, 875 F.2d 699, 716-20 (9th Cir.1989) (en banc) (Norris, J., concurring) (rejecting the Army’s argument that Bowers v. Hardwick forecloses a gay service member's equal protection claim), cert. denied, 498 U.S. 957, 111 S.Ct. 384, 112 L.Ed.2d 395 (1990); Cass R. 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) (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."). Likewise, the Romer Court declined to exclude gay men and lesbians who engage in same-sex sexual relations from the protection of its ruling. The opinion did not differentiate between men and women who merely had a "homosexual orientation" and those who engaged in "homosexual conduct;” indeed, it struck down in its entirety Amendment 2, which encompassed both "homosexual ... orientation [and] conduct.” Romer, - U.S. at -, 116 S.Ct. at -, 1623. Thus, although the majority relies on our prior cases which found no violation of equal protection because the underlying conduct could be regulated under the Due Process Clause, see Hatheway v. Secretary of the Army, 641 F.2d 1376, 1382 (9th Cir.1981), cert. denied, 454 U.S. 864, 102 S.Ct. 324, 70 L.Ed.2d 164 (1981), or assumed that the military could disfavor those who engaged in "homosexual conduct" even though it could not disfavor *1435individuals with a “homosexual orientation,” see Meinhold v. U.S. Dep’t of Defense, 34 F.3d 1469, 1477 (9th Cir.1994), these cases have been undercut by Romer. See supra Section III.
. In Philips’ Commanding Officer's Recommendation for Administrative Separation by Reason of Homosexual Conduct as Evidenced by His Statement directed to the Chief of Personnel, he stated:
In support of the Department of Defense current policy regarding administrative separations based on homosexuality, I concur with the recommendations of the administrative separation board and recommend Petty Officer Philips for an Honorable discharge.
Since reporting on board TPU Puget Sound in December 1992 for administrative processing, Petty Officer Philips has consistently proven to be a 4.0 Sailor. Always willing to accept additional responsibilities, he has excelled at every endeavor.
Even though his future in the military has been uncertain since December 1992, he has continued to serve the Navy to the utmost of his ability and performed remarkably well.
If ever DOD policy changes and Petty Officer Philips is called back to active duty, I have utmost confidence that he will have a highly productive career and serve the Navy with pride and professionalism.
. It is also evident from the legislative history of the statute. Witnesses testified that "the question of open versus closeted behavior is the key here. I agree that open homosexuality works against unit cohesion,” S. Hrg. No. 103-845, at 730 (emphasis added); "the introduction of an open homosexual into a small unit immediately polarizes that unit," S. Rpt. No. 103-112, at 280 (emphasis added); "the presence of open homosexuality would have an unacceptable detrimental and disruptive impact." Id. at 278 (emphasis added). Such an effect could stem only from the negative reactions of service members prejudiced against homosexuals.
.That such a reaction would in fact occur is far from certain. The ability of service members to put duty before prejudice has been tested before. When President .Truman integrated the armed forces in 1948, he did so despite widespread opposition from the military and even greater opposition from society than the current opposition to allowing gay men and lesbians in the military. See Holmes v. California Army Nat'l Guard, 920 F.Supp. 1510, 1533 (N.D.Cal.1996); Watkins, 875 F.2d at 729 n. 32 (Norris, J., concurring); RAND’s National Defense Research Institute, Sexual Orientation and U.S. Military Personnel Policy: Options and Assessment 22 (1993) (hereinafter "RAND Study”). Although " ‘[djire consequences were predicted for maintaining discipline, building group morale, and achieving military organization goals, ... [njone of these predictions has come true.’ ” Dahl v. Secretary of U.S. Navy, 830 F.Supp. 1319, 1330 (E.D.Cal.1993) (quoting a 1988 study commissioned by the DOD concerning gays in the military). The RAND study, which was prepared for the DOD, concludes that the experience of integrating the military indicates that "civilian and military leadership can effectively overcome the initial resistance to change and can minimize the worst fears of opponents about the damaging effects on unit performance.” RAND Study at 22. The experience of countries that permit gay men and lesbians to serve in the military also supports the conclusion that unit performance would not be adversely affected. RAND Study at 15 ("None of the militaries studied for this report believe their *1436effectiveness as an organization has been impaired or reduced as a result of the inclusion of homosexuals.”). Similarly, United States fire and police departments visited by RAND researchers "report that, overall, the effectiveness of the organization has not been diminished by the presence of homosexuals on the force.” RAND Study at 19. And, despite the military's fear of unit breakdown, many gay service members have been accepted and supported by their units. Cf. Richard Posner, Sex and Reason 319 (1992) ("The most important reason for doubting that dropping the ban on homosexuals in the military would cause serious morale problems is simply that a large number of homosexuals already serve without significant difficulties.”). For these reasons, the evidence that unit cohesion would be harmed by the inclusion of gay men and lesbians is weak. However, even if true the military’s judgment that unit cohesion would be harmed, the only way that unit cohesion would be harmed is through the unit's animosity towards gay men and lesbians. A justification that rests on prejudice and hatred of others is illegitimate.
. Although the racial classifications in Palmore were subject to strict scrutiny, Cleburne holds that accommodating the private biases of others is equally invalid under rational basis review. 473 U.S. at 448, 105 S.Ct. at 3258-59.
. Although the Fourth Circuit has accepted the military's assertions that unit discipline and solidarity justify the policy, Thomasson v. Perry, 80 F.3d 915, 929 (4th Cir.1996) (en banc), cert. denied, - U.S. -, 117 S.Ct. 358, 136 L.Ed.2d 250 (1996), the dissent in that case has much the better of the argument: the military’s assertion of "unit cohesion” is merely another name for accommodating animosity towards gay men and lesbians. See id. at 951 (Hall, J., dissenting). “ 'Unit cohesion’ is a facile way for the ins to put a patina of rationality on their efforts to exclude the outs.” Id. at 952 (Hall, J., dissenting).
. Sexual relationships between heterosexual service members of the same rank are permitted; such relationships seem to provide more potential for “sexual tension" than same-sex relationships between a homosexual service member and a non-service member. Indeed, heterosexual sexual harassment in the military is a serious and widely prevalent problem. According to a GAO survey, in 1991 between 93-97% of women in military academies experienced some form of sexual harassment. GAO, DOD Service Academies: More Actions Needed to Eliminate Sexual Harassment 3 (January 31, 1994). This again points up the irrationality of the current policy.
. The government's arguments disturbingly parallel the arguments put forth in opposition to the proposed racial integration of the military in the 1940s: "Men on board ship live in particularly close association; in their messes, one man sits beside another; their hammocks or bunks are close together.... How many white men would choose, of their own accord, that their closest associates in sleeping quarters, at mess, and in a gun’s crew should be of another race? ... The General Board believes that the answer is 'Few, if any.’" Thomasson, 80 F.3d at 952 (Hall, J., dissenting) (quoting committee that studied the proposed racial integration of the Navy).
.The majority in Thomasson does not accept this characterization of the policy, reasoning that the military’s failure to ask recruits whether or not they are gay is simply a matter of conserving scarce resources. However, this conclusion contradicts the explicit terms of the policy. The policy officially acknowledges that "sexual orientation” is not a bar to military service. DOD Directive 1332.14(H)(1)(a) (1994); NAVADMIN 033/94, at 3; S. Hrg. No. 103-845, at 702. Thus, a service member who considers himself gay but tells no one is allowed in the military; it is not simply that the military has failed to catch him. However, a service member who tells the military that he is gay is subject to discharge. 10 U.S.C. § 654(b)(2).
. Schowengerdt v. United States, 944 F.2d 483 (9th Cir.1991), cert. denied, 503 U.S. 951, 112 S.Ct. 1514, 117 L.Ed.2d 650 (1992). In that case, we also noted that we were not addressing an equal protection objection to the policy, which “is not the same as a substantive due process objection.” Id. at 490 n. 8. Here, of course, we are addressing an equal protection challenge.
. Better also appears to be based, in part, on acceptance of the military’s assertions that "[a] homosexual might force his desires upon others” and "[h]omosexuals may be less productive/effective than their heterosexual counterparts.... ” 632 F.2d at 811 n. 22. As the military has now acknowledged, gay service members are no more prone to misconduct or less effective than their heterosexual counterparts. Accordingly, these assertions were not offered as rationale for the current policy.
. The Supreme Court has reviewed numerous constitutional claims against the military. See, e.g., Goldman v. Weinberger, 475 U.S. 503, 106 S.Ct. 1310, 89 L.Ed.2d 478 (1986) (first amendment challenge to Air Force dress code); Rostker v. Goldberg, 453 U.S. 57, 101 S.Ct. 2646, 69 L.Ed.2d 478 (1981) (equal protection challenge to male-only draft registration); Brown v. Glines, 444 U.S. 348, 100 S.Ct. 594, 62 L.Ed.2d 540 (1980) (first amendment challenge to Air Force regulation regarding circulation of petitions); Frontiero v. Richardson, 411 U.S. 677, 93 S.Ct. 1764, 36 L.Ed.2d 583 (1973) (equal protection challenge to military benefits statutes which discriminated against women).