with whom Justice Brennan joins, dissenting.
The Court today places its imprimatur on one of the most potent remaining public expressions of “ancient canards about the proper role of women,” Phillips v. Martin Marietta Corp., 400 U. S. 542, 545 (1971) (Marshall, J., concurring). It upholds a statute that requires males but not females to register for the draft, and which thereby categorically excludes women from a fundamental civic obligation. Because I believe the Court’s decision is inconsistent with the Constitution’s guarantee of equal protection of the laws, I dissent.
I
A
The background to this litigation is set out in the opinion of the Court, ante, at 59-64, and I will not repeat that discussion here. It bears emphasis, however, that the only question presented by this case is whether the exclusion of women from registration under the Military Selective Service Act, 50 U. S. C. App. § 451 et seq. (1976 ed. and Supp. III) (MSSA), contravenes the equal protection component of the Due Process Clause of the Fifth Amendment. Although the purpose of registration is to assist preparations for drafting civilians into the military, we are not asked to rule on the constitutionality of a statute governing conscription,1 With the advent of the All-Volunteer Armed Forces, the MSSA was specifically amended to preclude conscription as of July 1, 1973, Pub. L. 92-129, § 101 (a) (35), 85 Stat. 353, 50 U. S. C. App. § 467 (c), and reactivation of the draft would therefore re*87quire a legislative amendment. See S. Rep. No. 96-826, p. 155 (1980). Consequently, we are not called upon to decide whether either men or women can be drafted at all, whether they must be drafted in equal numbers, in what order they should be drafted, or, once inducted, how they are to be trained for their respective functions. In addition, this case does not involve a challenge to the statutes or policies that prohibit female members of the Armed Forces from serving in combat.2 It is with this understanding that I turn to the task at hand.
B
By now it should be clear that statutes like the MSSA, which discriminate on the basis of gender, must be examined under the “heightened” scrutiny mandated by Craig v. Boren, 429 U. S. 190 (1976).3 Under this test, a gender-based classification cannot withstand constitutional challenge unless the classification is substantially related to the achievement of an important governmental objective. Kirchberg v. Feenstra, 450 U. S. 455, 459, 459-460 (1981); Wengler v. Druggist Mutual Ins. Co., 446 U. S. 142, 150 (1980); Califano v. Westcott, 443 U. S. 76, 84 (1979); Orr v. Orr, 440 U. S. 268, 278 (1979); Craig v. Boren, supra, at 197. This test applies whether the *88classification discriminates against males or females. Caban v. Mohammed, 441 U. S. 380, 391 (1979); Orr v. Orr, supra, at 278-279; Craig v. Boren, supra, at 204.4 The party defending the challenged classification carries the burden of demonstrating both the importance of the governmental objective it serves and the substantial relationship between the discriminatory means and the asserted end. See Wengler v. Druggist Mutual Ins. Co., supra, at 151; Caban v. Mohammed, supra, at 393; Craig v. Boren, supra, at 204. Consequently before we can sustain the MSSA, the Government must demonstrate that the gender-based classification it employs bears “a close and substantial relationship to [the achievement of] important governmental objectives,” Personnel Administrator of Massachusetts v. Feeney, 442 U. S. 256, 273 (1979).
C
The MSSA states that “an adequate armed strength must be achieved and maintained to insure the security of this Nation.” 50 U. S. C. App. §451 (b). I agree with the majority, ante, at 70, that “[n]o one could deny that . . . the Government’s interest in raising and supporting armies is an ‘important governmental interest.’ ” Consequently, the first part of the Craig v. Boren test is satisfied. But the question remains whether the discriminatory means employed itself substantially serves the statutory end. In concluding that it does, the Court correctly notes that Congress enacted (and reactivated) the MSSA pursuant to its constitutional authority to raise and maintain armies.5 The majority also notes, *89ante, at 64, that “the Court accords great weight to the decisions of Congress,’ ” quoting Columbia Broadcasting System, Inc. v. Democratic National Committee, 412 U. S. 94, 102 (1973), and that the Court has accorded particular deference to decisions arising in the context of Congress’ authority over military affairs. I have no particular quarrel with these sentiments in the majority opinion. I simply add that even in the area of military affairs, deference to congressional judgments cannot be allowed to shade into an abdication of this Court’s ultimate responsibility to decide constitutional questions. As the Court has pointed out:
“[T]he phrase 'war power’ cannot be invoked as a talis-manic incantation to support any exercise of congressional power which can be brought within its ambit. '[E]ven the war power does not remove constitutional limitations safeguarding essential liberties.’ ” United States v. Robel, 389 U. S. 258, 263-264 (1967), quoting Home Bldg. & Loan Assn. v. Blaisdell, 290 U. S. 398, 426 (1934).
See United States v. L. Cohen Grocery Co., 255 U. S. 81, 88-89 (1921); Hamilton v. Kentucky Distilleries & Warehouse Co., 251 U. S. 146, 156 (1919); Ex parte Milligan, 4 Wall. 2, 121-127 (1866).
One such “safeguard] [of] essential liberties” is the Fifth Amendment’s guarantee of equal protection of the laws.6 When, as here, a federal law that classifies on the basis of gender is challenged as violating this constitutional guarantee, it is ultimately for this Court, not Congress, to decide whether there exists the constitutionally required “close and *90substantial relationship” between the discriminatory means employed and the asserted governmental objective. See Powell v. McCormack, 395 U. S. 486, 549 (1969); Baker v. Carr, 369 U. S. 186, 211 (1962). In my judgment, there simply is no basis for concluding in this case that excluding women from registration is substantially related to the achievement of a concededly important governmental interest in maintaining an effective defense. The Court reaches a contrary conclusion only by using an “[a]nnounced degre[e] of 'deference’ to legislative judgmen[t]” as a “facile abstraction] ... to justify a result.” Ante, at 69, 70.
II
A
The Government does not defend the exclusion of women from registration on the ground that preventing women from serving in the military is substantially related to the effectiveness of the Armed Forces. Indeed, the successful experience of women serving in all branches of the Armed Services would belie any such claim. Some 150,000 women volunteers are presently on active service in the military,7 and their number is expected to increase to over 250,000 by 1985. See Department of Defense Authorization for Appropriations for Fiscal Year 1981: Hearings on S. 2294 before the Senate. Committee on Armed Services, 96th Cong., 2d Sess., 1657, 1683 (1980) (1980 Senate Hearings); Women in the Military: Hearings before the Military Personnel Subcommittee of the House Committee on Armed Services, 96th Cong., 1st *91and 2d Sess., 13-23 (1979 and 1980) (Women in the Military-Hearings). At the congressional hearings, representatives of both the Department of Defense and the Armed Services testified that the participation of women in the All-Volunteer Armed Forces has contributed substantially to military effectiveness. See, e. g., 1980 Senate Hearings, at 1389 (Lt. Gen. Yerks), 1682 (Principal Deputy Assistant Secretary of Defense Danzig); Women in the Military Hearings, at 13-23 (Assistant Secretary of Defense Pirie). Congress has never disagreed with the judgment of the military experts that women have made significant contributions to the effectiveness of the military. On the contrary, Congress has repeatedly praised the performance of female members of the Armed Forces, and has approved efforts by the Armed Services to expand their role. Just last year, the Senate Armed Services Committee declared:
“Women now volunteer for military service and are assigned to most military specialties. These volunteers now make an important contribution to our Armed Forces. The number of women in the military has increased significantly in the past few years and is expected to continue to increase.” S. Rep. No. 96-826, p. 157 (1980).
Accord, S. Rep. No. 96-226, p. 8 (1979).8 These statements thus make clear that Congress’ decision to exclude women from registration — and therefore from a draft drawing on the pool of registrants — cannot rest on a supposed need to prevent women from serving in the Armed Forces. The justification for the MSSA’s gender-based discrimination must *92therefore be found in considerations that are peculiar to the objectives of registration.
The most authoritative discussion of Congress’ reasons for declining to require registration of women is contained in the Report prepared by the Senate Armed Services Committee on the Fiscal Year 1981 Defense Authorization Bill. S. Rep. No. 96-826, supra, at 156-161. The Report’s findings were endorsed by the House-Senate Conferees on the Authorization Bill. See S. Conf. Rep. No. 96-895, p. 100 (1980). Both Houses of Congress subsequently adopted the findings by passing the Conference Report. 126 Cong. Rec. 23126, 23261 (1980). As the majority notes, ante, at 74, the Report’s “findings are in effect findings of the entire Congress.” The Senate Report sets out the objectives Congress sought to accomplish by excluding women from registration, see S. Rep. No. 96-826, supra, at 157-161, and this Court may appropriately look to the Report in evaluating the justification for the discrimination.
B
According to the Senate Report, “[t]he policy precluding the use of women in combat is . . . the most important reason for not including women in a registration system.” S. Rep. No. 96-826, supra, at 157; see also S. Rep. No. 96-226, supra, at 9. In reaffirming the combat restrictions, the Report declared:
“Registering women for assignment to combat or assigning women to combat positions in peacetime then would leave the actual performance of sexually mixed units as an experiment to be conducted in war with unknown risk' — a risk that the committee finds militarily unwarranted and dangerous. Moreover, the committee feels that any attempt to assign women to combat positions could affect the national resolve at the time of mobilization, a time of great strain on all aspects of the Nation’s resources.” S. Rep. No. 96-826, supra, at 157.
*93Had appellees raised a constitutional challenge to the prohibition against assignment of women to combat, this discussion in the Senate Report might well provide persuasive reasons for upholding the restrictions. ' But the validity of the combat restrictions is not an issue we need decide in this case.9 Moreover, since the combat restrictions on women have already been accomplished through statutes and policies that remain in force whether or not women are required to register or to be drafted, including women in registration and draft plans will not result in their being assigned to combat roles. Thus, even assuming that precluding the use of women in combat is an important governmental interest in its own right, there can be no suggestion that the exclusion of women from registration and a draft is substantially related to the achievement of this goal.
The Court’s opinion offers a different though related explanation of the relationship between the combat restrictions and Congress’ decision not to require registration of women. The majority states that “Congress . . . clearly linked the need for renewed registration with its views of the character of a subsequent draft.” Ante, at 75. The Court also states that “Congress determined that any future draft, which would be facilitated by the registration scheme, would be characterized by a need for combat troops.” Ante, at 76. The Court then reasons that since women are not eligible for assignment to combat, Congress’ decision to exclude them from registration is not unconstitutional discrimination inasmuch as “[m]en and women, because of the combat restrictions on women, are simply not similarly situated for purposes of a draft or registration for a draft.” Ante, at 78. There is a certain logic to this reasoning, but the Court’s approach is fundamentally flawed.
*94In the first place, although the Court purports to apply the Craig v. Boren test, the “similarly situated” analysis the Court employs is in fact significantly different from the Craig v. Boren approach. Compare Kirchberg v. Feenstra, 450 U. S., at 459-460 (employing Craig v. Boren test), with id., at 463 (Stewart, J., concurring in result) (employing “similarly situated” analysis). The Court essentially reasons that the gender classification employed by the MSSA is constitutionally permissible because nondiscrimination is not necessary to achieve the purpose of registration to prepare for a draft of combat troops. In other words, the majority concludes that women may be excluded from registration because they will not be needed in the event of a draft.10
This analysis, however, focuses on the wrong question. The relevant inquiry under the Craig v. Boren test is not whether a gender-neutral classification would substantially advance important governmental interests. Rather, the question is whether the gender-based classification is itself substantially related to the achievement of the asserted governmental interest. Thus, the Government’s task in this case is to demonstrate that excluding women from registration substantially furthers the goal of preparing for a draft of combat troops. Or to put it another way, the Government must show that registering women -would substantially impede its efforts to prepare for such a draft. Under our precedents, the Government cannot meet this burden without showing that a gender-neutral statute would be a less effective means of attaining this end. See Wengler v. Druggists Mutual Ins. Co., 446 U. S., at 151. As the Court explained in Orr v. Orr, 440 U. S., at 283 (emphasis added):
“Legislative classifications which distribute benefits and burdens on the basis of gender carry the inherent risk of *95reinforcing sexual stereotypes about the ‘proper place’ of women and their need for special protection. . . . Where, as here, the [Government’s] . . . purposes are as well served by a gender-neutral classification as one that gender classifies and therefore carries with it the baggage of sexual stereotypes, the [Government] cannot be permitted to classify on the basis of sex.”
In this case, the Government makes no claim that preparing for a draft of combat troops cannot be accomplished just as effectively by registering both men and women but drafting only men if only men turn out to be needed.11 Nor can the Government argue that this alternative entails the additional cost and administrative inconvenience of registering women. This Court has repeatedly stated that the administrative convenience of employing a gender classification is not an adequate constitutional justification under the Craig v. Boren test. See, e. g., Craig v. Boren, 429 U. S., at 198; Frontiero v. Richardson, 411 U. S. 677, 690-691 (1973).
The fact that registering women in no way obstructs the governmental interest in preparing for a draft of combat troops points up a second flaw in the Court’s analysis. The Court essentially reduces the question of the constitutionality of male-only registration to the validity of a hypothetical program for conscripting only men. The Court posits a draft in which all conscripts are either assigned to those specific combat posts presently closed to women or must be available for rotation into such positions. By so doing, the Court is able to conclude that registering women would be no more than a “gestur[e] of superficial equality,” ante, at 79, since women are necessarily ineligible for every position to be filled in its hypothetical draft. If it could indeed be guaranteed *96in advance that conscription would be reimposed by Congress only in circumstances where, and in a form under which, all conscripts would have to be trained for and assigned to combat or combat rotation positions from which women are categorically excluded, then it could be argued that registration of women would be pointless.
But of course, no such guarantee is possible. Certainly, nothing about the MSSA limits Congress to reinstituting the draft only in such circumstances. For example, Congress may decide that the All-Volunteer Armed Forces are inadequate to meet the Nation’s defense needs even in times of peace and reinstitute peacetime conscription. In that event, the hypothetical draft the Court relied on to sustain the MSSA’s gender-based classification would presumably be of little relevance, and the Court could then be forced to declare the male-only registration program unconstitutional. This difficulty comes about because both Congress12 and the Court have lost sight of the important distinction between registration and conscription. Registration provides “an inventory of what the available strength is within the military qualified pool in this country.” Reinstitution of Procedures for Registration Under the Military Selective Service Act: Hearing before the Subcommittee on Manpower and Personnel of the Senate Armed Services Committee, 96th Cong., 1st Sess., 10 (1979) (Selective Service Hearings) (statement of Gen. Rogers). Conscription supplies the military with the personnel needed to respond to a particular exigency. The fact that registration is a first step in the conscription process does not *97mean that a registration law expressly discriminating between men and women may be justified by a valid conscription program which would, in retrospect, make the current discrimination appear functionally related to the program that emerged.
But even addressing the Court’s reasoning on its own terms, its analysis is flawed because the entire argument rests on a premise that is demonstrably false. As noted, the majority simply assumes that registration prepares for a draft in which every draftee must be available for assignment to combat. But the majority’s draft scenario finds no support in either the testimony before Congress, or more importantly, in the findings of the Senate Report. Indeed, the scenario appears to exist only in the Court’s imagination, for even the Government represents only that “in the event of mobilization, approximately two-thirds of the demand on the induction system would be for combat skills.” Brief for Appellant 29 (emphasis added). For my part, rather than join the Court in imagining hypothetical drafts, I prefer to examine the findings in the Senate Report and the testimony presented to Congress.
C
Nothing in the Senate Report supports the Court’s intimation that women must be excluded from registration because combat eligibility is a prerequisite for all the positions that would need to be filled in the event of a draft. The Senate Report concluded only that “[i]f mobilization were to be ordered in a wartime scenario, the primary manpower need would be for combat replacements.” S. Rep. No. 96-826, p. 160 (1980) (emphasis added). This conclusion was in keeping with the testimony presented at the congressional hearings. The Department of Defense indicated that in the event of a mobilization requiring reinstitution of the draft, the primary manpower requirement would be for combat troops and support personnel who can readily be deployed into combat. See 1980 Senate Hearings, at 1395 (Principal *98Deputy Assistant Secretary of the Army Clark), 1390 (Lt. Gen. Yerks). But the Department indicated that conscripts would also be needed to staff a variety of support positions having no prerequisite of combat eligibility, and which therefore could be filled by women. Assistant Secretary of Defense (Manpower, Reserve Affairs, and Logistics) Pirie explained:
“Not only will we need to expand combat arms, and as I said, that is the most pressing need, but we also will need to expand the support establishment at the same time to allow the combat arms to carry out their function successfully. The support establishment now uses women very effectively, and in wartime I think the same would be true.” Registration of Women: Hearing on H. R. 6569 before the Subcommittee on Military Personnel of the House Committee on Armed Services, 96th Cong., 2d Sess., 17 (1980) (1980 House Hearings).
In testifying about the Defense Department’s reasons for concluding that women should be included in registration plans, Pirie stated:
“It is in the interest of national security that, in an emergency requiring the conscription for military service of the Nation’s youth, the best qualified people for a wide variety of tasks in our Armed Forces be available. The performance of women in our Armed Forces today strongly supports the conclusion that many of the best qualified people for some military jobs in the 18-26 age category will be women.” Id., at 7.
See 1980 Senate Hearings, at 171 (Secretary of the Army Alexander), 182 (Secretary of the Navy Clay tor).13 The De*99fense Department also concluded that there are no military-reasons that would justify excluding women from registration. The Department’s position was described to Congress in these terms:
"Our conclusion is that there are good reasons for registering [women]. Our conclusion is even more strongly that there are not good reasons for refusing to register them.” Id., at 1667-1668 (Principal Deputy Assistant Secretary of Defense Danzig) (emphasis added).
All four Service Chiefs agreed that there are no military reasons for refusing to register women, and uniformly advocated requiring registration of women. The military’s position on the issue was summarized by then Army Chief of Staff General Rogers: “[W]omen should be required to register for the reason that [Marine Corps Commandant] General Wilson mentioned, which is in order for us to have an inventory of what the available strength is within the military qualified pool in this country.” Selective Service Hearings, at 10; see id., at 10-11 (Adm. Hayward, Chief of Naval Operations; Gen. Allen, Air Force Chief of Staff; Gen. Wilson, Commandant, Marine Corps).
*100Against this background, the testimony at the congressional hearings focused on projections of manpower needs in the event of an emergency requiring reinstitution of the draft, and, in particular, on the role of women in such a draft. To make the discussion concrete, the testimony examined a draft scenario dealing with personnel requirements during the first six months of mobilization in response to a major war in Europe. The Defense Department indicated three constraints on the maximum number of women the Armed Services could use in the event of such a mobilization:
“(1) legislative prohibitions against the use of women in certain military positions, (2) the policy to reserve certain assignments, such as ground combat roles, for men only, and (3) the need to reserve a substantial number of noncombat positions for men in order to provide a pool of ready replacements for ground combat positions.” 1980 House Hearings, at 6 (Assistant Secretary Pirie).
After allowing for these constraints, the Defense Department reached the following conclusion about the number of female draftees that could be absorbed:
“If we had a mobilization, our present best projection is that we could use women in some 80,000 of the jobs that we would be inducting 650,000 people for. The reason for that is because some 80,000 of those jobs, indeed more than 80,000 of those jobs are support related and not combat related.
“We think women could fill those jobs quite well.” 1980 Senate Hearings, at 1688 (Principal Deputy Assistant Secretary of Defense Danzig).
See id., at 1661, 1665, 1828; 1980 House Hearings, at 6, 16-17 (Assistant Secretary of Defense Pirie).14 Finally, the De*101partment of Defense acknowledged that amending the MSSA to authorize registration and induction of women did not necessarily mean that women would be drafted in the same numbers as men. Assistant Secretary Pirie explained:
“If women were subject to the draft, the Department of Defense would determine the maximum number of women that could be used in the Armed Forces, subject to existing constraints and the needs of the Military Services to provide close combat fillers and replacements quickly. We estimate that this might require at least 80,000 additional women over the first six months. If there were not enough women volunteers, a separate draft call for women would be issued.” Id., at 6.
See 1980 Senate Hearings, at 1661 (Principal Deputy Assistant Secretary of Defense Danzig).
This review of the findings contained in the Senate Report and the testimony presented at the congressional hearings demonstrates that there is no basis for the Court’s representation that women are ineligible for all the positions that would need to be filled in the event of a draft. Testimony about personnel requirements in the event of a draft established that women could fill at least 80,000 of the 650,000 positions for which conscripts would be inducted. Thus, with respect to these 80,000 or more positions, the statutes and policies barring women from combat do not provide a reason for distinguishing between male and female potential conscripts; the two groups are, in the majority’s parlance, “similarly situated.” As such, the combat restrictions cannot by themselves supply the constitutionally required justification for the MSSA’s gender-based classification. Since the classification precludes women from being drafted to fill positions for which they would be qualified and useful, the Govern*102ment must demonstrate that excluding women from those positions is substantially related to the achievement of an important governmental objective.
Ill
The Government argues, however, that the “consistent testimony before Congress was to the effect that there is no military need to draft women.” Brief for Appellant 31 (emphasis in original). And the Government points to a statement in the Senate Report that “[b]oth the civilian and military leadership agreed that there was no military need to draft women. . . . The argument for registration and induction of women ... is not based on military necessity, but on considerations of equity.” S. Rep. No. 96-826, p. 158 (1980). In accepting the Government’s contention, the Court asserts that the President’s decision to seek authority to register women was based on “equity,” and concludes that “Congress was certainly entitled, in the exercise of its constitutional powers to raise and regulate armies and navies, to focus on the question of military need rather than ‘equity.’ ” Ante, at 80. In my view, a more careful examination of the concepts of “equity” and “military need” is required.
As previously noted, the Defense Department’s recommendation that women be included in registration plans was based on its conclusion that drafting a limited number of women is consistent with, and could contribute to, military effectiveness. See supra, at 97-102. It was against this background that the military experts concluded that “equity” favored registration of women. Assistant Secretary Pirie explained:
“Since women have proven that they can serve successfully as volunteers in the Armed Forces, equity suggests that they be liable to serve as draftees if conscription is reinstated.” 1980 House Hearings, at 7.
By “considerations of equity,” the military experts acknowledged that female conscripts can perform as well as male con*103scripts in certain positions, and that there is therefore no reason why one group should be totally excluded from registration and a draft. Thus, what the majority so blithely dismisses as “equity” is nothing less than the Fifth Amendment’s guarantee of equal protection of the laws which “requires that Congress treat similarly situated persons similarly,” ante, at 79. Moreover, whether Congress could subsume this constitutional requirement to “military need,” in part depends on precisely what the Senate Report meant by “military need.”
The Report stated that “[b]oth the civilian and military leadership agreed that there was no military need to draft women.” S. Rep. No. 96-826, supra, at 158. An examination of what the “civilian and military leadership” meant by “military need” should therefore provide an insight into the Report’s use of the term. Several witnesses testified that because personnel requirements in the event of a mobilization could be met by drafting men, including women in draft plans is not a military necessity. For example, Assistant Secretary of Defense Pirie stated:
“It is doubtful that a female draft can be justified on the argument that wartime personnel requirements cannot be met without them. The pool of draft eligible men ... is sufficiently large to meet projected wartime requirements.” 1980 House Hearings, at 6.
See 1980 Senate Hearings, at 1665 (Principal Deputy Assistant Secretary of Defense Danzig). Similarly, Army Chief of Staff General Meyer testified:
“I do not believe there is a need to draft women in peacetime. In wartime, because there are such large numbers of young men available, approximately 2 million males in each year group of the draft age population, there would be no military necessity to draft females except, possibly, doctors, and other health pro*104fessionals if there are insufficient volunteers from people with those skills.” Id., at 749.
To be sure, there is no “military need” to draft women in the sense that a war could be waged without their participation.15 This fact is, however, irrelevant to resolving the constitutional issue.16 As previously noted, see supra, at 94-95, it is not appellees’ burden to prove that registration of women substantially furthers the objectives of the MSSA.17 Rather, *105because eligibility for combat is not a requirement for some of the positions to be filled in the event of a draft, it is incumbent on the Government to show that excluding women from a draft to fill those positions substantially furthers an important governmental objective.
It may be, however, that the Senate Report’s allusion to “military need” is meant to convey Congress’ expectation that women volunteers will make it unnecessary to draft any women. The majority apparently accepts this meaning when it states: “Congress also concluded that whatever the need for women for noncombat roles during mobilization, whether 80,000 or less, it could be met by volunteers.” Ante, at 81. But since the purpose of registration is to protect against unanticipated shortages of volunteers, it is difficult to see how excluding women from registration can be justified by conjectures about the expected number of female volunteers.18 I fail to see why the exclusion of a pool of persons who would be conscripted only if needed can be justified by reference to the current supply of volunteers. In any event, the Defense Department’s best estimate is that in the event of a mobilization requiring reinstitution of the draft, there will not be *106enough women volunteers to fill the positions for which women would be eligible. The Department told Congress:
“If we had a mobilization, our present best projection is that we could use women in some 80,000 of the jobs we would be inducting 650,000 people for.” 1980 Senate Hearings, at 1688 (Principal Deputy Assistant Secretary of Defense Danzig) (emphasis added).19
Thus, however the “military need” statement in the Senate Report is understood, it does not provide the constitutionally required justification for the total exclusion of women from registration and draft plans.
IV
Recognizing the need to go beyond the “military need” argument, the Court asserts that “Congress determined that staffing noncombat positions with women during a mobilization would be positively detrimental to the important goal of military flexibility.” Ante, at 81-82. None would deny that preserving “military flexibility” is an important governmental interest. But to justify the exclusion of women from registration and the draft on this ground, there must be a further showing that staffing even a limited number of noncombat positions with women would impede military flexibility. I find nothing in the Senate Report to provide any basis *107for the Court’s representation that Congress believed this to be the ease.
The Senate Report concluded that “military reasons . . . preclude very large numbers of women from serving.” S. Rep. No. 96-826, p. 158 (1980) (emphasis added). The Report went on to explain:
“Military flexibility requires that a commander be able to move units or ships quickly. Units or ships not located at the front or not previously scheduled for the front nevertheless must be able to move into action if necessary. In peace and war, significant rotation of personnel is necessary. We should not divide the military into two groups — one in permanent combat and one in permanent support. Large numbers of non-combat positions must be available to which combat'troops can return for duty before being redeployed.” Ibid.
This discussion confirms the Report’s conclusion that drafting “very large numbers of women” would hinder military flexibility. The discussion does not, however, address the different question whether drafting only a limited number of women would similarly impede military flexibility. The testimony on this issue at the congressional hearings was that drafting a limited number of women is quite compatible with the military’s need for flexibility. In concluding that the Armed Services could usefully employ at least 80,000 women conscripts out of a total of 650,000 draftees that would be needed in the event of a major European war, the Defense Department took into account both the need for rotation of combat personnel and the possibility that some support personnel might have to be sent into combat. As Assistant Secretary Pirie testified:
“If women were subject to the draft, the Department of Defense would determine the maximum number of women that could be used in the Armed Forces, subject to existing constraints and the needs of the Military *108Services to provide close combat fillers and replacements quickly. We estimate that this might require at least 80,000 additional women over the first 6 months.” 1980 House Hearings, at 6 (emphasis added).
See App. 278 (deposition of Principal Deputy Assistant Secretary of Defense Danzig).20
Similarly, there is no reason why induction of a limited number of female draftees should any more divide the military into “permanent combat” and “permanent support” groups than is presently the case with the All-Volunteer Armed Forces. The combat restrictions that would prevent a female draftee from serving in a combat or combat rotation position also apply to the 150,000-250,000 women volunteers in the Armed Services. If the presence of increasing but controlled numbers of female volunteers has not unacceptably “divide[d] the military into two groups,” it is difficult to see how the induction of a similarly limited additional, number of women could accomplish this result. In these circumstances, I cannot agree with the Court’s attempt to “interpret” the Senate Report’s conclusion that drafting very large numbers of women would impair military flexibility, as proof that Congress reached the entirely different conclusion that drafting a limited number of women would adversely affect military flexibility.
*109Y
The Senate Report itself recognized that the “military-flexibility” objective speaks only to the question whether “very large numbers” of women should be drafted. For the Report went on to state:
“It has been suggested that all women be registered, but only a handful actually be inducted in an emergency. The committee finds this a confused and ultimately unsatisfactory solution.” S. Rep. No. 96-826, p. 158 (1980).
The Report found the proposal “confused” and “unsatisfactory” for two reasons.
“First, the President’s proposal [to require registration of women] does not include any change in section 5 (a) (1) of the [MSSA], which requires that the draft be conducted impartially among those eligible. Administration witnesses admitted that the current language of the law probably precludes induction of women and men on any but a random basis, which should produce roughly equal numbers of men and women. Second, it is conceivable that the courts, faced with a congressional decision to register men and women equally because of equity considerations, will find insufficient justification for then inducting only a token number of women into the Services in an emergency.” Id., at 158-159 (emphasis in original).
The Report thus assumed that if women are registered, any subsequent draft would require simultaneous induction of equal numbers of male and female conscripts. The Report concluded that such a draft would be unacceptable:
“It would create monumental strains on the training system, would clog the personnel administration and support systems needlessly, and would impede our defense preparations at a time of great national heed.
“Other administrative problems such as housing and *110different treatment with regard to dependency, hardship and physical standards would also exist.” Id., at 159.21
See also S. Rep. No. 96-226, p. 9 (1979). Relying on these statements, the majority asserts that even “assuming that a small number of women could be drafted for noncombat roles, Congress simply did not consider it worth the added burdens of including women in draft and registration plans.” Ante, at 81. In actual fact, the conclusion the Senate Report reached is significantly different from the one the Court seeks to attribute to it.
The specific finding by the Senate Report was that “[i]f the law required women to be drafted in equal numbers with men, mobilization would be severely impaired because of strains on training facilities and administrative systems.” S. Rep. No. 96-826, supra, at 160 (emphasis added). There was, however, no suggestion at the congressional hearings that simultaneous induction of equal numbers of males and female conscripts was either necessary or desirable. The Defense Department recommended that women be included in registration and draft plans, with the number of female draftees and the timing of their induction to be determined by the military’s personnel requirements. See supra, at 100-101.22 In endorsing this plan, the Department gave no indication that such a draft would place any strains on training and administrative facilities. Moreover, the Director of the Selective Service System testified that a registration and induction *111process including both males and females would present no administrative problems. See 1980 Senate Hearings, at 1679 (Bernard Rostker); App. 247-248 (deposition of Bernard Rostker).
The Senate Report simply failed to consider the possibility that a limited number of women could be drafted because of its conclusion that § 5 (a)(1) of the MSSA does not authorize drafting different numbers of men and women and its speculation on judicial reaction to a decision to register women. But since Congress was free to amend §5 (a)(1), and indeed would have to undertake new legislation to authorize any draft, the matter cannot end there. Furthermore, the Senate Report’s speculation that a statute authorizing differential induction of male and female draftees would be vulnerable to constitutional challenge is unfounded. The unchallenged restrictions on the assignment of women to combat, the need to preserve military flexibility, and the other factors discussed in the Senate Report provide more than ample grounds for concluding that the discriminatory means employed by such a statute would be substantially related to the achievement of important governmental objectives. Since Congress could have amended § 5 (a)(1) to authorize differential induction of men and women based on the military’s personnel requirements, the Senate Report’s discussion about “added burdens” that would result from drafting equal numbers of male and female draftees provides no basis for concluding that the total exclusion of women from registration and draft plans is substantially related to the achievement of important governmental objectives.
In sum, neither the Senate Report itself nor the testimony presented at the congressional hearings provides any support for the conclusion the Court seeks to attribute to the Report— that drafting a limited number of women, with the number and the timing of their induction and training determined by the military’s personnel requirements, would burden training and administrative facilities.
*112VI
After reviewing the discussion and findings contained in the Senate Report, the most I am able to say of the Report is that it demonstrates that drafting very large numbers of women would frustrate the achievement of a number of important governmental objectives that relate to the ultimate goal of maintaining “an adequate armed strength ... to insure the security of this Nation," 50 U. S. C. App. § 451 (b). Or to put it another way, the Senate Report establishes that induction of a large number of men but only a limited number of women, as determined by the military’s personnel requirements, would be substantially related to important governmental interests. But the discussion and findings in the Senate Report do not enable the Government to carry its burden of demonstrating that completely excluding women from the draft by excluding them from registration substantially furthers important governmental objectives.
In concluding that the Government has carried its burden in this case, the Court adopts “an appropriately deferential examination of Congress’ evaluation of [thel evidence,” ante, at 83 (emphasis in original). The majority then proceeds to supplement Congress’ actual findings with those the Court apparently believes Congress could (and should) have made. Beyond that, the Court substitutes hollow shibboleths about “deference to legislative decisions” for constitutional analysis. It is as if the majority has lost sight of the fact that “it is the responsibility of this Court to act as the ultimate interpreter of the Constitution.” Powell v. McCormack, 395 U. S., at 549. See Baker v. Carr, 369 U. S., at 211. Congressional enactments in the area of military affairs must, like all other laws, be judged by the standards of the Constitution. For the Constitution is the supreme law of the land, and all legislation must conform to the principles it lays down. As the Court has pointed out, “the phrase 'war power’ cannot be invoked as a talismanic incantation to support any exercise of *113congressional power which can be brought within its ambit.” United States v. Robel, 389 U. S., at 263-264.
Furthermore, “[w]hen it appears that an Act of Congress conflicts with [a constitutional] provisio[n], we have no choice but to enforce the paramount commands of the Constitution. We are sworn to do no less. We cannot push back the limits of the Constitution merely to accommodate challenged legislation.” Trop v. Dulles, 356 U. S. 86, 104 (1958) (plurality opinion). In some 106 instances since this Court was established it has determined that congressional action exceeded the bounds of the Constitution. I believe the same is true of this statute. In an attempt to avoid its constitutional obligation, the Court today “pushes back the limits of the Constitution” to accommodate an Act of Congress.
I would affirm the judgment of the District Court.
Given the Court’s lengthy discourse on the background to this litigation, it is interesting that the Court chooses to bury its sole reference to this fact in a footnote. See ante, at 60, n. 1.
By statute, female members of the Air Force and the Navy may not be assigned to vessels or aircraft engaged in combat missions. See 10 U. S. C. § 6015 (1976 ed., Supp III), §8549. Although there are no statutory restrictions on the assignment of women to combat in the Army and the Marine Corps, both services have established policies that preclude such assignment.
Appellees do not concede the constitutional validity of these restrictions on women in combat, but they have taken the position that their validity is irrelevant for purposes of this case.
I join the Court, see ante, at 69, in rejecting the Solicitor General’s suggestion that the gender-based classification employed by the MSSA should be scrutinized under the "rational relationship” test used in reviewing challenges to certain types of social and economic legislation. See, e. g., Schweiker v. Wilson, 450 U. S. 221 (1981); U. S. Railroad Retirement Bd. v. Fritz, 449 U. S. 166 (1980).
Consequently, it is of no moment that the constitutional challenge in this case is pressed by men who claim that the MSSA’s gender classification discriminates against them.
The Constitution grants Congress the power “To raise and support Armies,” “To Provide and maintain a Navy,” and “To make Rules for the Government and Regulation of the land and naval Forces.” U. S. Const., Art. I, § 8, cls. 12-14.
Although the Fifth Amendment contains no Equal Protection Clause, this Court has held that “the Fifth Amendment’s Due Process Clause prohibits the Federal Government from engaging in discrimination that is ‘so unjustifiable as to be violative of due process.’ ” Schlesinger v. Ballard, 419 U. S. 498, 500, n. 3 (1975), quoting Bolling v. Sharpe, 347 U. S. 497, 499 (1954).
With the repeal in 1967 of a statute limiting the number of female members of the Armed Forces to 2% of total enlisted strength, the number of women in the military has risen steadily both in absolute terms and as a percentage of total active military personnel. The percentage has risen from 0.78% in 1966, to over 5% in 1976, and is expected to rise to 12% by 1985. See U. S. Dept. of Defense, Use of Women in the Military 5-6 (2d ed. 1978), reprinted at App. 98, 111-113; M. Binkin & S. Bach, Women and the Military 13-21 (1977).
In summarizing the testimony presented at the congressional hearings, Senator Cohen stated:
“[B]asically the evidence has come before this committee that participation of women in the All-Volunteer Force has worked well, has been praised by every military officer who has testified before the committee, and that the jobs are being performed with the same, if not in some cases, with superior skill.” 1980 Senate Hearings, at 1678.
As noted, see n. 2, supra, appellees elected not to challenge the constitutionality of the combat restrictions.
I would have thought the logical conclusion from this reasoning is that there is in fact no discrimination against women, in which case one must wonder why the Court feels compelled to pledge its purported fealty to the Craig v. Boren test.
Alternatively, the Government could employ a classification that is related to the statutory objective but is not based on gender, for example, combat eligibility. Under the current scheme, large subgroups of the male population who are ineligible for combat because of physical handicaps or conscientious objector status are nonetheless required to register.
The Court quotes Senator Warner’s comment: “ T equate registration with the draft,’ ” ante, at 75. The whole of Senator Warner’s statement merits quotation because it explains why Congress refused to acknowledge the distinction between registration and the draft. Senator Warner stated: “Frankly I equate registration with the draft because there is no way you can establish a registration law on a coequal basis and then turn right around and establish a draft law on a nonequal basis. I think the court would knock that down right away.” 1980 Senate Hearings, at 1197.
Pirie explained the reasoning behind the Defense Department’s conclusion in these terms:
“Large numbers of military women work in occupations such as electronics, communications, navigation, radar repair, jet engine mechanics, drafting, *99surveying, ordnance, transportation and meteorology and do so very effectively, as has been shown by numerous DOD studies and tests. The work women in the Armed Forces do today is essential to the readiness and capability of the forces. In case of war that would still be true, and the number of women doing similar work would inevitably expand beyond our peacetime number of 250,000.
“Women have traditionally held the vast majority of jobs in fields such as administrative/clerical and health care/medical An advantage of registration for women is that a pool of trained personnel in these traditionally female jobs would exist in the event that sufficient volunteers were not available. It would make far greater sense to include women in a draft call and thereby gain many of these skills than to draft only males who would not only require training in these fields but would be drafted for employment in jobs traditionally held by females. A further advantage would be to release males currently holding noncombatant jobs for reassignment to combat jobs.” 1980 House Hearings, at 6.
The Defense Department arrived at this number after it “surveyed the military services, and asked them how many women they could use *101[in the event of a mobilization of] 650,000, and received answers suggesting that they could use about 80,000.” 1980 Senate Hearings, at 1665 (Principal Deputy Assistant Secretary of Defense Danzig).
A colloquy between Senator Jepsen and Principal Deputy Assistant Secretary of Defense Danzig reveals that some Members of Congress understood “military need” in this sense.
“Mr. Danzig. . . .
“We surveyed the military services, and asked them how many women they could use among those 650,000, and received answers suggesting that they could use 80,000.
“Let me indicate when I say they could use[,] I do not mean to imply that they would have to use women. Our Department of Defense view is that women would be useful in a mobilization scenario. If women were not available, I do not think the republic would crumble. Men could be used instead.
“Senator Jepsen. So there is no explicit military requirement involved ?
“Mr. Danzig. My problem, Senator, and I don’t mean to be semantic about it, is with the use of the words, ‘explicit requirement.’ If you said to me, for example, does the military require people with brown eyes to serve, I would tell you no, because people with blue eyes, et cetera, could do the job.
“On the other hand, I wouldn’t deny that they could do the job and that we would find them useful.” 1980 Senate Hearings, at 1665; see id., at 1853-1856.
Deputy Assistant Attorney General Simms explained as much to Congress in his testimony at the hearings. He stated:
“[T]he question of military necessity for drafting women is irrelevant to the constitutional issue, which is whether or not there is sufficient justification by whatever test the courts may apply for not registering women.”
Id., at 1667.
If we were to assign appellees this burden, then all of the Court’s prior “mid-level” scrutiny equal protection decisions would be drawn into question. For the Court would be announcing a new approach under *105which the party challenging a gender-based classification has the burden of showing that elimination of the classification substantially furthers an important governmental interest.
As Assistant Secretary of Defense Pirie explained:
“Perhaps sufficient women volunteers would come forward to meet this need, perhaps not. Having our young women register in advance would put us in a position to call women if they do not volunteer in sufficient numbers,” quoted at 126 Cong. Rec. 13885-13886 (1980).
See 1980 Senate Hearings, at 1828 (Principal Deputy Assistant Secretary of Defense Danzig).
Past wartime recruitment experience does not bear out the Court’s sanguine view. With the advent of the Korean War, an unsuccessful effort was made to recruit some 100,000 women to meet the rapidly expanding manpower requirements. See Use of Women in the Military, supra n. 7, at 5, App. 111.
A colloquy between Representative Hillis and Assistant Secretary of Defense Pirie at the House Hearings makes clear that the 80,000 number is in addition to the number of women serving in the All-Volunteer Armed Forces.
“Mr. Pirie. Mr. Hillis, we estimate that we would need 650,000 individuals to be inducted over the first six months.
“Mr. Hillis. How many of those would be women?
“Mr. Pirie. At least 80,000 of these individuals would be women, Mr. Hillis.
“Mr. Hillis. That is even if we had the 250,000 [women in active service expected by 1985], you are talking about another 80,000, which » projects into about 330,000.
“Mr. Pirie. Yes, sir.” 1980 House Hearings, at 22.
Senator Warner questioned the Service Chiefs about the “impact on your service as a consequence of a draft, which would be based on a total provision of equality between male and female.” Selective Service Hearings, at IS (emphasis added). Two of the Service Chiefs answered Senator Warner's question about the effect of a draft of equal numbers of men and women. Their answers merit quotation.
“General Allen [Air Force]. It would not have any unfavorable effect on the Air Force. We would have no objection to such a draft.” Ibid.
“General WilsoN [Marine Corps]....
“. . . [W]e would be perfectly happy to have women drafted. That is up to the 5 percent goal which I believe we can handle in the Marine Corps.” Ibid.
The Report further explained:
“If the Congress were to mandate equal registration of men and women, therefore, we might well be faced with a situation in which the combat replacements needed in the first 60 days — say 100,000 men — would have to be accompanied by 100,000 women. Faced with this hypothetical, the military witnesses stated that such a situation would be intolerable.” S. Rep" No. 96-826, at 159.
As stated in the Senate Report, “Selective Service Plans provide[d] for drafting only men during the first 60 days, and only a small number of women would be included in the total drafted for the first 180 days.” Id., at 158.