This is the first case in which a federal appellate court has allowed a member of the armed services to bring a substantive due process challenge to the congressionally enacted “Don’t Ask, Don’t Tell” homosexual personnel policy for the military. With respect, I believe that our three-judge panel has erroneously reversed a district court order dismissing such suit and remanded for further fact-finding under an unsanctioned and malleable standard of review.

This case is far more than a harmless remand. Witt v. Department of the Air Force, 527 F.3d 806 (9th Cir. 2008), claims to rest its decision on the Supreme Court’s opinion in Lawrence v. Texas, 539 U.S. 558, 123 S.Ct. 2472, 156 L.Ed.2d 508 (2003), which decriminalized private and consensual homosexual conduct. Instead, however, Witt contravenes Supreme Court precedent, including Lawrence, in the area of substantive due process, creates a circuit split, and stretches the judicial power beyond its constitutional mandate. At the end of the day, Witt creates a forum in the judicial branch (rather than the political branches) to challenge the validity and the particular application of “Don’t Ask, Don’t Tell,” even though such policy infringes no constitutional right. Since today’s order denies rehearing of this problematic case by an en banc court, I must respectfully dissent.

i — i

h>

Unlike Lawrence, this is not a criminal case. Major Margaret Witt, an Air Force reservist nurse, was suspended from the Air Force when a military board found, after a two-day hearing, that she “had engaged in homosexual acts and had stated that she was a homosexual in violation of [10 U.S.C. § 654, commonly known as the ‘Don’t Ask, Don’t Tell’ policy].”1 Witt, 527 F.3d at 810. The military board recommended her discharge and the Secretary of the Air Force has so ordered, but it is unclear whether she has been formally discharged. See id. at 812. Major Witt admitted that she had “a committed and *1266long-term relationship” with “another woman from July 1997 through August 2003.” Id. at 809. As the panel noted, “Major Witt’s partner was never a member nor a civilian employee of any branch of the armed forces, and Major Witt states that she never had sexual relations while on duty or while on the grounds of any Air Force base.” Id. The two women lived together in Spokane, Washington, some 250 miles from the base where Major Witt was stationed. Id. at 809-10. Major Witt challenged her termination as a violation of her constitutional rights under the Equal Protection Clause, the procedural Due Process Clause, and the substantive Due Process Clause. Id. at 809. The district court granted the Air Force’s motion to dismiss Major Witt’s lawsuit. Id.

B

On appeal, a panel of this court upheld the dismissal of both the procedural due process claim and the equal protection claim, the former on standing grounds and the latter on the merits. However, the Witt panel went on to hold that there might be a viable substantive due process claim based on Lawrence and remanded the case back to the district court “to develop the record.” Id. at 821. Although previous Ninth Circuit decisions had found “Don’t Ask, Don’t Tell” constitutional under rational basis review, see, e.g., Philips v. Perry, 106 F.3d 1420, 1425-26 (9th Cir. 1997), the panel presumed, without any analysis or explanation, that Lawrence changed the equation.

After careful but, I respectfully submit, misguided consideration, the panel concluded that Lawrence introduced a new requirement of some kind of heightened scrutiny. The Lawrence opinion itself does not prescribe such scrutiny or even mention that it applied heightened scrutiny of any kind. Lacking a standard of review, the panel imported a multi-factor balancing test from another Supreme Court case, Sell v. United States, 539 U.S. 166, 123 S.Ct. 2174, 156 L.Ed.2d 197 (2003), that it thought enunciated a standard of review appropriate to Lawrence. In Sell, a case involving the government’s forcible administration of anti-psychotic drugs to a mentally ill defendant, the Court did develop a four-factor test, but only on the basis of prior cases with similar facts. See 539 U.S. at 179-181, 123 S.Ct. 2174 (deriving a standard from Washington v. Harper, 494 U.S. 210, 213, 110 S.Ct. 1028, 108 L.Ed.2d 178 (1990), which considered “whether a judicial hearing is required before the State may treat a mentally ill prisoner with anti-psychotic drugs against his will” and Riggins v. Nevada, 504 U.S. 127, 135-37, 112 S.Ct. 1810, 118 L.Ed.2d 479 (1992), which reversed State convictions because the defendant was unconstitutionally forced to take an anti-psychotic drug during trial). The Sell Court explicitly tied its test to those cases where the government had “involuntarily ... administered] antipsychotic drugs to a mentally ill defendant facing serious criminal charges in order to render that defendant competent to stand trial.” Id. at 179, 123 S.Ct. 2174.2

Notwithstanding the inapposite origin of the Sell test, the Witt panel adapted it to its own purposes. In Witt’s modified version, the inquiry requires “that when the *1267government attempts to intrude upon the personal and private lives of homosexuals, in a manner that implicates the rights identified in Lawrence,3 the government must advance an important governmental interest, the intrusion must significantly further that interest, and the intrusion must be necessary to further that interest.... In addition, we hold that this heightened scrutiny analysis is as-applied rather than facial.” Witt, 527 F.3d at 819.4

Applying this novel standard, the panel remanded for further fact-finding to determine whether “Don’t Ask, Don’t Tell” met the requirements of the modified Sell test with respect to Major Witt’s particular circumstances. The Witt opinion leaves no doubt about how fact-specific this inquiry is to be. The panel ordered the trial court on remand to determine “whether the application of [‘Don’t Ask, Don’t Tell’] specifically to Major Witt significantly furthers the government’s interest and whether less intrusive means would achieve substantially the government’s interest.” Id. at 821 (emphasis added). Witt remains the only federal appellate decision to subject “Don’t Ask, Don’t Tell” to such invasive review.5 It seems to me that if a court of appeals is to require a district court to second-guess the considered policy decision of Congress as applied to a particular military officer, it must do so with the clearest constitutional command and upon the most solid Supreme Court precedent. But, as I explain below, no such clarity, no such solidity, underlie Witt.

II

My first objection to Witt is to its application of Lawrence in the first place. The panel presumed, without any analysis, that Lawrence controlled this case. The Witt court thought that its first task with respect to the substantive due process claim was to “determine the proper level of scrutiny to apply.” Id. at 813.6 In other words, the panel ignored the necessary threshold question: “is this a Lawrence case at all?” Its assumption — that Witt is a Lawrence case — was wrong, and that mistake caused the panel both to misapply Lawrence and to contradict the well-settled substantive due process precedents of the Supreme Court.

A

In Lawrence, the Supreme Court struck down a Texas criminal statute penalizing certain homosexual conduct. It overruled Bowers v. Hardwick, 478 U.S. 186, 106 5.Ct. 2841, 92 L.Ed.2d 140 (1986), which upheld a similar Georgia law under the rational basis test. Citing “broad statements of the substantive reach of liberty under the Due Process Clause [of the Fourteenth Amendment] in earlier cases,” Lawrence, 539 U.S. at 565, 123 S.Ct. 2472, *1268the Court began by summarizing its objection to statutes criminalizing the consensual sexual activity at issue:

Their penalties and purposes touch[] upon the most private human conduct, sexual behavior, and in the most private of places, the home. The statutes ... seek to control a personal relationship that, whether or not entitled to formal recognition in the law, is within the liberty of persons to choose without being punished as criminals.

Id. at 567, 123 S.Ct. 2472 (emphasis added).

The Lawrence Court then explained its decision to overrule Bowers. It called into question what it saw as two critical pillars of that decision: the historical prevalence and continuing vitality of the moral condemnation of common homosexual conduct. Id. at 568-72, 123 S.Ct. 2472 (noting first that “there is no longstanding history in this country of laws directed at homosexual conduct as a distinct matter” and second that there is “an emerging awareness [in the past half century] that liberty gives substantial protection to adult persons in deciding how to conduct their private lives in matters pertaining to sex”). The Court approvingly referenced the argument in Justice Stevens’ dissent in Bowers that traditional moral condemnation is insufficient to uphold a law, even under the rational basis test, especially one penalizing “ ‘individual decisions by married [and unmarried] persons, concerning the intimacies of their physical relationship.’ ” Id. at 577-78, 123 S.Ct. 2472 (quoting Bowers, 478 U.S. at 216, 106 S.Ct. 2841 (Stevens, J., dissenting)).

1

Some of this language, taken alone, might appear to sweep somewhat broadly. As if to prevent lower courts from mistaking such an appearance for the actual holding, however, Lawrence explicitly limited the reach of the liberty interest that it protected in a concluding paragraph:

The present case does not involve minors. It does not involve persons who might be injured or coerced or who are situated in relationships where consent might not easily be refused. It does not involve public conduct or prostitution. It does not involve whether the government must give formal recognition to any relationship that homosexual persons seek to enter.

Id. at 578, 123 S.Ct. 2472 (emphasis added).

Witt, as several other lower court opinions have done, remarked upon the sometimes opaque language of the Laurrence opinion. See Witt, 527 F.3d at 816 (noting the “studied limits of the verbal analysis in Lawrence ”); see also United States v. Extreme Assocs., Inc., 352 F.Supp.2d 578, 591 (W.D.Pa. 2005) (noting the “lack of clarity in Lawrence ”), rev’d on other grounds, 431 F.3d 150 (3d Cir. 2005). But, as the limiting language quoted above shows, the occasional generality of the analysis does not mean that Lawrence applies to any statute that affects homosexual conduct to the exclusion of the rest of the case law on substantive due process. Such is common sense, and our sister circuits have recognized it. See, e.g., Muth v. Frank, 412 F.3d 808, 817 (7th Cir. 2005) (“Lawrence ... did not announce ... a fundamental right, protected by the Constitution, for adults to engage in all manner of consensual sexual conduct....”); Lofton v. Sec’y of the Dep’t of Children & Family Servs., 358 F.3d 804, 817 (11th Cir. 2004) (noting that “the shared homosexuality component” does not alone require application of Lawrence); United States v. Marcum, 60 M.J. 198, 205 (2004) (same).

I fear that the panel was misled by some of the most general and abstract statements in Lawrence, and therefore failed to exercise such common sense restraint. *1269With respect, I suggest that, in placing Witt under Lawrence’s umbrella, the panel made precisely the mistake that Lawrence’s limiting paragraph warned lower courts against. I submit that if one reads Lawrence with an eye to determining what cases it meant to control, this mistake becomes apparent.

2

The Court’s final summary, along with the limiting language, neatly shows how to read its broad pronouncements along with the specifics of the case before it: “[t]he petitioners are entitled to respect for their private lives. The State cannot demean their existence or control their destiny by making their private sexual conduct a crime. Their right to liberty ... gives them the full right to engage in their conduct without intervention of the government.” Lawrence, 539 U.S. at 578, 123 S.Ct. 2472.

The Supreme Court recites that the “petitioners are entitled to respect for their private lives,” reiterating the general thought animating its opinion. Proceeding from this statement to an applicable holding, however, it concludes that the State cannot tread on that entitlement — how?— “by making their private sexual conduct a crime.” Finally, the Court repeats the conclusion from the perspective of the individual. He or she has the right “to engage in [sexual] conduct without the intervention of the government.” “[IJntervention,” standing alone, is quite broad, but we must consider the context. The Court seems to have had in mind a certain kind of government intrusion — criminal sanction — and a certain kind of human activity — free participation in intimate behavior. The Witt panel professed its desire to focus on what the Lawrence Court did rather than to “dissectf ] isolated pieces of text,” Witt, 527 F.3d at 816, and yet Witt’s analysis derives almost entirely from Lawrence’s dicta. What Lawrence did was to strike down an outlier criminal statute punishing private, consensual homosexual conduct in the home. Lawrence, 539 U.S. at 578, 123 S.Ct. 2472. Some of our sister circuits have already arrived at the same conclusion. See, e.g., Lofton, 358 F.3d at 815 (“Lawrence’s holding was that substantive due process does not permit a state to impose a criminal prohibition on private consensual homosexual conduct.”).

Crucially, as if to anticipate challenges like the one before the Witt panel, Lawrence concludes by limiting its holding to the facts of the case. See Lawrence, 539 U.S. at 578, 123 S.Ct. 2472. I have quoted the entire passage above, see supra at 1268, but the most relevant limitation here is that the opinion did not apply itself to cases involving “public conduct” or require that “the government ... give formal recognition to any relationship that homosexual persons seek to enter,” Lawrence, 539 U.S. at 578, 123 S.Ct. 2472.7 If one combines the holding with this self-imposed limitation, one comes to the inexorable conclusion that a Lawrence case requires, at least, a criminal sanction on private conduct. Put in the negative, Lawrence did not deal with laws addressed to public conduct or non-criminal statutes.

3

Does such reading unjustifiably narrow Lawrence’s reach? It is, of course, true *1270that Lawrence contains some broad language, but we cannot stretch every scrap of dictum beyond the breaking point of the holding it underlies. I believe that careful scrutiny of the Lawrence opinion vindicates my reading of it.

All the statutes Lawrence cites in explaining its actual holding are criminal statutes. See generally 539 U.S. at 567-74, 123 S.Ct. 2472. Indeed, almost all of the substantive due process cases the Court cites at the start of its analysis for the most general propositions in the opinion also dealt with criminal statutes. See id. at 564-67, 123 S.Ct. 2472 (summarizing the reasoning of Griswold v. Connecticut, 381 U.S. 479, 85 S.Ct. 1678, 14 L.Ed.2d 510 (1965) (invalidating statute criminalizing the use of contraceptives); Eisenstadt v. Baird, 405 U.S. 438, 92 S.Ct. 1029, 31 L.Ed.2d 349 (1972) (invalidating statute criminalizing the distribution of contraceptives to unmarried persons); and Carey v. Population Services International, 431 U.S. 678, 97 S.Ct. 2010, 52 L.Ed.2d 675 (1977) (invalidating statute criminalizing the sale or distribution of contraceptives to persons younger than sixteen years old)).

Furthermore, in explaining why the Texas statute offended the Constitution, the Court focused like a laser on its criminal consequences. See, e.g., id. at 571, 123 S.Ct. 2472 (“The issue is whether the majority may use the power of the State to enforce[moral] views on the whole society through operation of the criminal law.” (emphasis added)); id. at 575, 123 S.Ct. 2472 (“When homosexual conduct is made criminal by the law of the State, that declaration in and of itself is an invitation to subject homosexual persons to discrimination ....” (emphasis added)); id. (“The stigma this criminal statute imposes ... is not trivial. The offense to be sure, is but a Class C misdemeanor .... [but] it remains a criminal offense with all that imports for the dignity of the persons charged. The petitioners will bear on their record the history of their criminal convictions.” (emphases added)); id. (“[I]f Texas convicted an adult for private, consensual homosexual conduct under the statute here in question the convicted person would come within the registration laws of at least four States .... ” (emphases added)); id. at 576 (noting “the consequential nature of the punishment and the state-sponsored condemnation attendant to the criminal prohibition .... [and that] the Texas criminal conviction carries with it the other collateral consequences always following a conviction” (emphases added)). These sentences form much of the reasoning preceding the holding that the Court summarized at the end of its opinion, namely that “[t]he State cannot demean [homosexuals’] existence or control their destiny by making their private sexual conduct a crime.” Id. at 578, 123 S.Ct. 2472 (emphasis added).

The reasoning of Laurrence also suggests that it was one of those cases, which periodically arise, in which the Court sees itself as striking down an anachronistic or extreme statute, that is, an irrational outlier. See, e.g., City of Cleburne v. Cleburne Living Ctr., 473 U.S. 432, 105 S.Ct. 3249, 87 L.Ed.2d 313 (1985) (striking down zoning ordinance requiring a special-use permit to build a group home for the mentally disabled). It was, after all, the increasingly unusual nature, and rare enforcement, of criminal statutes like the one in Texas that motivated the Lawrence Court in part. See Lawrence, 539 U.S. at 573, 123 S.Ct. 2472 (“The 25 States with laws prohibiting the relevant conduct referenced in the Bowers decision are reduced now to 13, of which 4 enforce their laws only against homosexual conduct.”). In explaining why its decision would not cause significant social upheaval — i.e., would not strike down too many laws — Lawrence noted the dwindling number of other *1271states that criminalized homosexual conduct. See id. at 570-71, 573, 123 S.Ct. 2472. In fact, the smallness of this number was one of the justifications the Court gave for overruling Bowers at all. Thus, that this number is indeed small and that states are unlikely to revive the type of criminal statute at issue does not mean that lower courts should be extending Lawrence to new situations.

In any event, Lawrence limits itself in numerous ways, see, e.g., id. at 578, 123 S.Ct. 2472; I merely suggest that by ignoring those limits in this case, we have erroneously embroiled ourselves in a debate which belongs predominantly in the political branches.

B

Once one keeps these limits in mind, it becomes apparent that Witt strayed far beyond the proper reach of Lawrence. The case before the panel involved both public conduct and a non-criminal statute. The opinion makes much of the fact that the homosexual acts for which, in part, Major Witt was dismissed occurred in the privacy of the couple’s shared home. But nothing in the “Don’t Ask, Don’t Tell” policy forbids anyone from doing anything in the home on pain of criminal or even of civil penalties. Indeed, the whole point of the policy is to keep such private behavior private. If no one asks and no one tells, no one in the military cares. “Don’t Ask, Don’t Tell” is about how the military manages its personnel; the policy only matters if an employee’s homosexual conduct or acknowledgment of homosexuality becomes public. What happened in this case, and what must happen for “Don’t Ask, Don’t Tell” to apply, is that homosexual conduct, originally private or not, became public. And Lawrence simply does not apply to non-criminal public conduct. Id. at 578, 123 S.Ct. 2472.

Furthermore, “Don’t Ask, Don’t Tell” imposes no criminal sanction-the intrusion that so worried the Court in Lawrence. Major Witt has been and remains at liberty to carry on a homosexual relationship, even though the Air Force plans to discharge her. Lawrence may guarantee people the liberty to engage privately in “sexual practices common to a homosexual lifestyle,” id. at 578, 123 S.Ct. 2472, but it never suggests a constitutional right to job security. See Lofton, 358 F.3d at 817 (refusing to apply Lawrence to the adoption of children in part because “[t]he relevant state action is not criminal prohibition, but grant of a statutory privilege. And the asserted liberty interest is not the negative right to engage in private conduct without facing criminal sanctions, but the affirmative right to receive official and public recognition.”).

A final corollary of the Court’s focus on criminal sanctions for consensual activity in the home is that Lawrence does not apply in the military context. The opinion excludes public conduct from its holding. It is difficult to imagine a context more public than the personnel policies of the armed services. Indeed, the initial conclusion the Court drew from the broad language of early substantive due process cases supports this interpretation: “as a general rule, [the State or a court should not] define the meaning of the relationship or ... set its boundaries absent injury to a person or abuse of an institution the law protects.” Lawrence, 539 U.S. at 567, 123 S.Ct. 2472. American jurisprudence reflects consistent deference to the considerations of the military, as one of the institutions the law protects. See, e.g., Goldman v. Weinberger, 475 U.S. 503, 507-08, 106 S.Ct. 1310, 89 L.Ed.2d 478 (1986); Rostker v. Goldberg, 453 U.S. 57, 66, 101 S.Ct. 2646, 69 L.Ed.2d 478 (1981) (“The operation of a healthy deference to legislative and executive judgments in the *1272area of military affairs is evident in several recent decisions of this Court.”); Parker v. Levy, 417 U.S. 733, 743, 94 S.Ct. 2547, 41 L.Ed.2d 439 (1974) (“This Court has long recognized that the military is, by necessity, a specialized society separate from civilian society.”). Just this term, the Supreme Court explicitly reaffirmed this tradition of deference in a case involving “ ‘complex, subtle, and professional decisions as to the composition, training, equipping, and control of a military force,’ which are ‘essentially professional military judgments.’ ” Winter v. Natural Res. Def. Council, — U.S.-, 128 S.Ct. 2996, 171 L.Ed.2d 910 (2008) (quoting Gilligan v. Morgan, 413 U.S. 1, 10, 93 S.Ct. 2440, 37 L.Ed.2d 407 (1973), and also citing Goldman, 475 U.S. at 507, 106 S.Ct. 1310).

By remanding to the district court for further proceedings, Witt is in considerable tension with this traditional deference, which obtains even when it comes to constitutional rights. See, e.g., Goldman, 475 U.S. at 509-10, 106 S.Ct. 1310 (rejecting a First Amendment challenge to a military policy barring an Orthodox Jew from wearing his yarmulke while in uniform). The fact-specific and subjective inquiry that Witt ordered the district court to undertake on remand recalls the “independent evaluation of the evidence” by a district court that the Supreme Court declared “quite wrong” in Rostker, 453 U.S. at 83, 101 S.Ct. 2646 (insisting that the district court should have “adopt[ed] an appropriately deferential examination of Congress’ evaluation of [the] evidence” (emphasis added)). Indeed, I am also reminded of Chief Justice Warren’s warning that “courts are ill-equipped to determine the impact upon discipline that any particular intrusion upon the military might have.” Earl Warren, The Bill of Rights and the Military, 37 N.Y.U. L.Rev. 181, 187 (1962).8

All of these reasons show quite clearly that Lawrence does not apply to the “Don’t Ask, Don’t Tell” policy. Since Lawrence did not disturb our pre-Law-rence precedents holding that such policy survives rational basis review, I believe the panel should have simply applied our existing law.

Ill

Even if Lawrence did apply to Witt, the panel significantly overstated the implications of such application. It began by noting that Lawrence did not read like a typical rational basis case, which seems fair enough. But I respectfully suggest that the panel made too much of that by importing from Sell, a case which had nothing to do with Lawrence, a multi-factor heightened scrutiny test. Lawrence is but one decision in the sprawling jurisprudence regarding the substantive reach of the Due Process Clause, and understanding its role requires placing it in context.

A

Since the Supreme Court initiated the third iteration of substantive due process in Griswold, its eases have moved, sometimes haltingly, toward a distinction between abridgements of “fundamental rights,” which merit strict scrutiny, and *1273other laws restricting the freedom of individuals to act, which receive rational basis review only. Crucially, the Court has policed the boundaries between these two regimes closely, leaving no doubt that lower courts should not broaden the reach of substantive due process beyond the strict confines of precedent. Glucksberg, 521 U.S. at 720, 117 S.Ct. 2258 (“[W]e ‘ha[ve] always been reluctant to expand the concept of substantive due process because guideposts for responsible decisionmaking in this unchartered area are scarce and open-ended.’ ”) (quoting Collins v. City of Barker Heights, 503 U.S. 115, 125, 112 S.Ct. 1061, 117 L.Ed.2d 261 (1992)).

To guide lower courts, the Supreme Court has explained its “established method” for determining what counts as a fundamental right, nowhere more famously than when it reversed our decision in Glucksberg:

First, we have regularly observed that the Due Process Clause specially protects those fundamental rights and liberties which are, objectively, deeply rooted in this Nation’s history and tradition and implicit in the concept of ordered liberty, such that neither liberty nor justice would exist if they were sacrificed. Second, we have required in substantive-due-process cases a careful description of the asserted fundamental liberty interest. Our Nation’s history, legal traditions, and practices thus provide the crucial guideposts for responsible deci-sionmaking, that direct and restrain our exposition of the Due Process Clause. As we stated recently in Flores, the Fourteenth Amendment forbids the government to infringe ... “fundamental” liberty interests at all, no matter what process is provided, unless the infringement is narrowly tailored to serve a compelling state interest.

Id. at 720-21, 117 S.Ct. 2258 (internal quotation marks and citations omitted), rev’g Compassion in Dying v. Washington, 79 F.3d 790 (9th Cir. 1996) (en banc).

It is certainly true, as the opinion in Witt points out, that the Supreme Court has occasionally made exceptions to this dual structure in specific cases. Sell is one such case; the abortion cases are another example, since they apply an analysis unique to their subject-matter. But it is up to the Supreme Court to carve out such niches, not a panel of the United States Court of Appeals for the Ninth Circuit. See Rodriguez de Quijas v. Shearson/Am. Express, Inc., 490 U.S. 477, 484, 109 S.Ct. 1917, 104 L.Ed.2d 526 (1989) (admonishing a federal appellate court for anticipating a change in Supreme Court case law and for refusing to apply controlling precedent). The opinion in Lawrence, unlike that in Sell, simply did not clearly lay out such a new standard of review.

Exactly where Lawrence fits into the framework of Glucksberg and the Court’s modern substantive due process jurisprudence certainly has proven a vexing question. But one can say with confidence that, measured against Glucksberg’s established method, Lawrence did not announce a new fundamental right. See Lawrence, 539 U.S. at 594, 123 S.Ct. 2472 (Scalia, J., dissenting) (“The Court today does not ... once ... describe homosexual sodomy as a ‘fundamental right.’ ”). And, more importantly, Lawrence gave no license to ignore Glucksberg’s cautionary admonition to hew closely to precedent in substantive due process cases. In particular, the Supreme Court reminds us that

[b]y extending constitutional protection to an asserted right or liberty interest, we, to a great extent, place the matter outside the arena of public debate and legislative action. We must therefore exercise the utmost care whenever we are asked to break new ground in this field, lest the liberty protected by the *1274Due Process Clause be subtly transformed into the policy preferences of the Members of this Court.

Glucksberg, 521 U.S. at 720, 117 S.Ct. 2258 (internal quotation marks and citation omitted).

Our panel, regrettably, did not heed this advice but instead drifted well off course. Not once does Witt cite Glucksberg or relevant due process law beyond Lawrence and associated cases. Not once does Witt pause to consider whether it had forgotten the limits of both Laurrence and Glucks-berg and ventured into the wilderness. Instead, it simply applied three parts of the four-factor test of Sell, a case Lawrence never cites, as a kind of makeshift heightened scrutiny.9 The only conclusion to draw from such heightened scrutiny is either that Witt created a new fundamental right or that it elevated a liberty interest to some other kind of protected status. Either innovation is without sanction of the Supreme Court and defies the Court’s admonitions to tread cautiously in the area of substantive due process.10

The result of the panel’s innovation — its remand to the district court to develop the record on an as-applied basis — flies in the face of both Congress’ careful consideration and the Supreme Court’s emphasis on deference to military policies. Such judicial adventurism lays the groundwork for a continuing series of fact-bound challenges to the specific application of a law that Witt only calls into constitutional question in the first place by stretching substantive due process beyond repair. I know no principle of law or logic that justifies such a result.

B

Contrast the panel’s approach with that of the Eleventh Circuit in Lofton, which recognized that Lawrence did not create a fundamental right. Lofton, 358 F.3d at 815. Lofton read Lawrence in the context of the Supreme Court’s substantive due process jurisprudence to evaluate what the Court “did” in that case. Performing the task that Witt overlooked, Lofton checked Lawrence against cases such as Glucks-berg which articulate the two-part test for the recognition of fundamental rights under the Due Process Clause, a test conspicuously absent from Lawrence. Id. at 815-17; see also Glucksberg, 521 U.S. at 719-20, 117 S.Ct. 2258. Given that Lawrence did not run through that test, but simply struck down the Texas statute under rational basis review, and the Court’s advisable caution in this area, see Glucksberg, 521 U.S. at 720, 117 S.Ct. 2258, the Lofton court unsurprisingly concluded that Lawrence did not articulate any new fundamental right.

Several other courts of appeals have concluded likewise, reading the opinion *1275alongside the body of substantive due process case law. See, e.g., Seegmiller v. LaVerkin City, 528 F.3d 762, 770-71 (10th Cir. 2008); Williams v. Att’y Gen. of Ala., 378 F.3d 1232, 1235-43 (11th Cir. 2004); see also Sylvester v. Fogley, 465 F.3d 851, 857-58 (8th Cir. 2006) (dicta); Muth, 412 F.3d at 817-18 (same). But cf. Cook, 528 F.3d at 53-54 (holding that, although Lawrence did not recognize a fundamental right, it did recognize a “protected liberty interest”); Marcum, 60 M.J. at 205 (holding that Lawrence did not recognize a fundamental right but does require “searching constitutional inquiry”). The government’s petition for rehearing also lists several district, bankruptcy, and state courts that have followed suit, some in holdings and some in dicta.

C

It is no secret that Lawrence pursues rational basis review with unusual vigor. But if the panel wished to identify cases analogous in this respect, it might have pointed either to City of Cleburne or to Romer v. Evans, 517 U.S. 620, 116 S.Ct. 1620, 134 L.Ed.2d 855 (1996), instead of to Sell. In both of those cases, as in Lawrence, the Supreme Court struck down, under the rational basis test, laws that on inspection seemed to reflect little more than “bare animus” and “irrationality.” Compare Lawrence, 539 U.S. at 577, 123 S.Ct. 2472 (accepting as controlling the view that “the fact that the governing majority in a State has traditionally viewed a particular practice as immoral is not a sufficient reason for upholding a law prohibiting the practice”), with City of Cleburne, 473 U.S. at 450, 105 S.Ct. 3249 (“[T]he short of it is that [the statute] in this case appears to us to rest on an irrational prejudice against the mentally retarded .... ”), and Romer, 517 U.S. at 634, 116 S.Ct. 1620 (“[A] bare ... desire to harm a politically unpopular group [is not a] legitimate governmental interest.” (internal quotation marks omitted)). The same cannot be said of the “Don’t Ask, Don’t Tell” policy, which Congress enacted into law after an exhaustive review of the military’s needs and the pros and cons of adopting the policy. See Cook, 528 F.3d at 58-60 (reviewing Congress’ deliberations regarding “Don’t Ask, Don’t Tell”).

Such observations dramatize the panel’s mistake. A basic rule of substantive due process jurisprudence is that, barring subject-area-specific standards such as those the Court fashioned in the abortion cases or in Sell, fundamental rights get strict scrutiny, and everything else gets rational basis. Thus, if Lawrence did not establish the heightened scrutiny standard that the panel jury-rigged from portions of Sell— and it did not — then Supreme Court law requires the use of the plain vanilla rational basis test. By inventing a specialized standard of review, Witt broke that rule.

IV

No matter how strongly some of us may feel about the underlying issues in this case, the Supreme Court’s precedents in substantive due process law compel not only our usual obedience, but also our self-conscious restraint. We have no mandate to follow either our reasons or our convictions down paths the Constitution and the Court have left for Congress to chart. Lawrence did not change that, nor did it provide a forum for lower courts to measure the policy decisions of Congress against the circumstances of a particular litigant.

Whatever we think of its merits, the “Don’t Ask, Don’t Tell” policy is not an outlying anachronism, but a considered legislative judgment. If Congress now reconsiders it in the light of new developments or evidence, then it acts entirely within its constitutional powers. But until Congress acts, the federal judiciary should not preempt its policy choices, and certain*1276ly not at the cost of tearing substantive due process law from the guideposts to which the Supreme Court has fastened it.

For the foregoing reasons, I respectfully dissent from our unfortunate decision not to rehear this case en banc.

I reach the same conclusion as Judge O’Scannlain, that we ought to rehear this case en banc and affirm. The district court was correct in dismissing Major Witt’s challenge for failure to state a claim upon which relief could be granted. I write separately because that conclusion is compelling even taking the law more favorably to the panel’s view than Judge O’Scannlain would.

Suppose, for the sake of argument, that the line of authority beginning with Gris-wold v. Connecticut1 and resting at present with Lawrence v. Texas2 establishes a broad constitutional right, enforceable in civil as well as criminal proceedings, to liberty among consenting adults to have whatever sort of sexual contact they choose.3 And suppose further, for the sake of argument, that the burden on the government to justify interference with this constitutional right to sexual liberty is intermediate or strict scrutiny, rather than scrutiny to determine whether the government restraint has a rational basis.4 I do not suggest that either of these propositions is or should be true or false, but use them as hypothetical bases for discussing the present case.

Even under such a broad and aggressive interpretation of Lawrence> the panel would still be mistaken. The reason why is that the general constitutional right to sexual liberty competes against the especially high level of deference we are required to extend to Congress and the President regarding military affairs, and few liberties prevail against that deference. As explained below, I do not think the panel’s decision can stand unless Goldman v. Weinberger falls.5

“Don’t Ask, Don’t Tell,” or “DADT” as the panel calls it, is not a policy adopted by the military. It is among the major policy initiatives during the first two years of the Clinton Administration. New laws passed by Congress receive such extensive scrutiny by individual legislators, witnesses military and otherwise testifying at hearings, journalists, and the general public as this one.6 No doubt sex triggers more public attention than, say, improvements to air*1277ports. At the end of the extensive public controversy, Congress passed and the President signed a law embodying the policy challenged in this lawsuit, 10 U.S.C. § 654, promulgated November 13, 1993.7

Congress made fifteen findings explaining the reasons for the law. Among them are that “[s]uccess in combat requires military units that are characterized by high morale, good order and discipline, and unit cohesion,”8 that “military society is characterized by ... numerous restrictions on personal behavior, that would not be acceptable in civilian society,”9 that military standards “regulate a member’s life for 24 hours each day,”10 that living conditions are sometimes “characterized by forced intimacy with little or no privacy,” 11 and that “presence in the armed forces of persons who demonstrate a propensity or intent to engage in homosexual acts would create an unacceptable risk to the high standards of morale, good order and discipline, and unit cohesion that are the essence of military capability.”12 The statute requires separation (subject to numerous exceptions) from the armed forces of persons who have engaged in homosexual acts or declared their homosexuality.13

It is not our business to decide whether Congress and the President were correct or mistaken in their judgment about homosexuality and military effectiveness. I intimate no view of my own. Many homosexuals have been excellent soldiers, as is evidently the case with Major Witt, and some militaries seem to function fine with homosexuals in the ranks. On the other hand, further injecting passions of love and jealousy into the emotional maelstrom of armed nineteen year olds who may soon die for their country may reasonably be seen by Congress and the President as too risky, especially when combined with minority sexual orientations.

I cannot tell just what the panel intends for the district court to do on remand. The panel remanded the case to the district court to determine whether, under a heightened level of scrutiny, application of this law to Major Witt “significantly furthers the government’s interest.”14 Congress and the President concluded that it does, because the rule is general, for the entire military. The panel cannot mean that the district court should repeat the extensive congressional hearings that preceded adoption of the law, to determine whether the court agrees with the policy adopted. But the panel does not say what sort of evidence the district court is supposed to consider, or precisely what factual question the evidence is supposed to answer.

Do model officers such as Major Witt get an exception to the rule? Popular officers? Officers whose units appear to have suffered no decline in morale and unit cohesion? The panel does not justify making any of these things matter. Suppose all facts elicited for all these questions and all other particularized questions about Major Witt come out in her favor. She could be assigned to a different unit and a different location, perhaps in a war zone, tomorrow,15 or her personal relationships *1278and the relationships and feelings within her unit could change. Or even if none of those things happen, other homosexuals in the unit, either less popular or less capable than Major Witt, may become envious that she has more sexual liberty than they do, generating an attitude particularly corrosive to discipline. Congress and the President established a rule of general applicability, and morale probably requires that people be treated alike. As the only other circuit to have considered this issue post-Lawrence concluded, “[ejvery as-applied challenge brought by a member of the armed forces against the Act, implicates this interest [in morale and unit cohesion].” 16 In the Goldman case discussed below, Justice Stevens’s concurring opinion expressly rejected individualized balancing of the constitutionally protected interest against the military interest, such as the panel seems to require, because “we must test validity of the ... rule not merely as it applies to Captain Goldman but also as it applies to all service personnel....” 17

The panel, relying on civilian cases of criminal prosecution for homosexual conduct, Lawrence v. Texas,18 and forced medication of a criminal defendant, Sell v. United States,19 gives mere lip service to the military aspect of the case.

Regardless of what liberties the Constitution guarantees for civilians, the military context changes the analysis. As the Court held in Rostker v. Goldberg, “the tests and limitations to be applied may differ because of the military context.”20 In Rostker, the Supreme Court rejected a due process challenge to drafting men but not women, justifying its holding on the ground that “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.”21 Congress appears to have shaped a finding in the statute before us to fit the Rostker language, the finding that “the Constitution of the United States commits exclusively to the Congress the powers to raise and support armies ... and make rules for the government and regulation of the land and naval forces.”22

Quite a few constitutional rights do not apply to the military as they do in civilian life. For example, though mandatory conscription is not currently in force, men but not women must register for the draft, and an equal protection challenge may be ruled out by Rostker v. Goldberg,23 When men are drafted, they are not free to decline the employment on Thirteenth Amendment grounds, as civilians are.24 And soldiers may be vigorously prosecuted for desertion, though civilians are generally free to leave an employment.25 If a new Secretary in, say, the Department of Education issued an order that all employees *1279must stand and salute when a higher ranking employee enters the room, no doubt a First Amendment challenge would succeed, perhaps with an a fortiori reference to the proposition that even a child need not salute the flag in school during wartime,26 yet it is hard to imagine a soldier defeating an insubordination charge on the ground that he had a First Amendment right to express his true feeling of disrespect and not make a false show of respect for a superior officer who entered the room.27

This case should be treated as an a fortiori application of the Supreme Court’s holding in Goldman v. Weinberger,28 An Orthodox Jewish rabbi serving as a clinical psychologist in a base hospital was recommended for non-retention because he wore a yarmulke (5 1/2" circle of dark cloth) on his head on duty. The right to free exercise of religion is certainly as protected by the Constitution as whatever right to sexual liberty Lawrence may create, and the military headgear policy was merely a regulation, not a statute, yet the Court rejected Captain Goldman’s Free Exercise challenge to application of the rule. The Court relied on its well established general principle that “ ‘the military is, by necessity, a specialized society separate from civilian society,’ ”29 and “ ‘must insist on a respect for duty and a discipline without counterpart in civilian life.’ ”30 The Court went so far as to say that “[t]he essence of military service ‘is the subordination of the desires and interests of the individual to the needs of the service,’ ”31 while defining the service’s “needs” to embrace pretty much whatever the military thought desirable to promote “discipline” and “uniformity.”32

Congress picked up the language from Goldman in its findings in the statute before us, that “the armed forces ... exist as a specialized society ... characterized by ... numerous restrictions on personal behavior, that would not be acceptable in civilian society.”33 It is more difficult to imagine threats to order and discipline from the wearing of a yarmulke by a clinical psychologist in an on-shore base hospital, than from the emotions stirred up by sexual conduct, and free exercise of religion is among the most fundamental constitutional rights, yet religion gives way to the military interest in discipline and uniformity. If a man does not have a constitutional right to wear an unobtrusive 5" circle of dark cloth on his head as his religion requires, because of the threat to discipline and uniformity, it is hard to see how an individual could nevertheless be entitled to practice or declare a sexual orientation that the military, Congress, and the President have concluded endangers military effectiveness.

Major Witt has a legitimate and important interest in associating and enjoying sexual relationships with those whom she chooses, as she chooses. Yet even these sorts of human relationships, among the *1280most private and important, are subject to strict military regulation. For example, fraternization between officers and enlisted persons is a military offense.34 Laws against adultery have generally fallen into desuetude for civilians, but are routinely enforced by the military.35 If a captain and a corporal fall in love, or if a soldier has sexual intercourse with another soldier’s spouse, the threat to military discipline and unit cohesion is serious regardless of the sexes of the participants. Our soldiers make great sacrifices for our country. At the very least, they sacrifice much of their liberty.36 I am unable to see, under Rostker and Goldman, how Major Witt’s liberty under Lawrence can trump the decision Congress and the President made to limit that liberty in the military.

*1265Order; Dissent by Judge O’SCANNLAIN; Dissent by Judge KLEINFELD; Dissent by Judge KOZINSKI.

ORDER

The panel voted to deny the Appellees’ Petition for Rehearing.

The full court has been advised of the Petition for Rehearing En Banc. A judge of the court requested a vote on whether to rehear the case en banc. However, the en banc call failed to receive a majority of votes of the nonrecused active judges in favor of en banc consideration. Fed. R.App. P. 35.

The Appellees’ Petition for Rehearing and the Petition for Rehearing En Banc are DENIED.

The panel has done a fine job in an exceptionally difficult and fraught area of the law. And Major Witt’s case compellingly illustrates the sometimes arbitrary and destructive operation of the “Don’t Ask, Don’t Tell” policy. At the same time, Judge O’Scannlain raises serious doubts about the scope of Lawrence, and Judge Kleinfeld makes a strong case for deference to the political branches on military matters.

I’m not sure who’s right, but I am sure that this is the type of case we need to take en banc because it “involves a question of exceptional importance.” Fed. R.App. P. 35(a)(2). The deference due in the realm of military affairs, see, e.g., Winter v. Natural Res. Def. Council, Inc., — U.S.-, 129 S.Ct. 365, — L.Ed.2d-(2008), reinforces the deference we owe to the political branches when they act in concert: “When the President acts pursuant to an express or implied authorization of Congress, his authority is at its maximum, for it includes all that he possesses in his own right plus all that Congress can delegate.” Youngstown Sheet & Tube Co. v. Sawyer, 343 U.S. 579, 635, 72 S.Ct. 863, 96 L.Ed. 1153 (1952) (Jackson, J., concurring).

When we stand against the combined might of the other branches of government, we should ensure that our own authority is at its maximum. En banc rehearing — whatever the outcome — would have shown that we gave this matter the sustained attention it merits. Moreover, there is strength in numbers: The conclusions of an en banc court would reflect many more points of view and could not easily be dismissed as outliers.

Finally, there’s much to be said here for making haste slowly. Delaying disposition of the case while we reconsidered it en banc might have given the political branches a chance to revisit the DADT policy in light of experience over the last 15 years— including that in Major Witt’s case. Fed. R.App. P. 35(a)(2) is there for a reason and we would have been well advised to follow its directive. 15926