(concurring):
I concur in the result reached by the court. I am concerned, however, with *1337the reasoning of Part III of its opinion relating to petitioner’s contention that the local counts (of which he was convicted) and the federal bank robbery count, which was dismissed prior to submission to the jury, were not properly joined in a single indictment under Canty v. United States.1 On that premise, petitioner urged us to conclude that the district court had no jurisdiction under D.C.Code § 11-502 to try him on the local charges.
In Canty, the defendant was convicted, inter alia, of bank robbery under the United States Code2 and assault with a deadly weapon under the D.C.Code.3 Both charges stemmed from his having robbed a bank with a gun. He was sentenced to consecutive terms of 6 to 18 years on the bank robbery count and 3 to 9 years on the local count, making a combined sentence of from 9 to 27 years. If the defendant had been charged and convicted of the federal bank robbery count alone, he would have been subject to a maximum term of 25 years.4 This court vacated the conviction and sentence on the assault count, holding that the Government was prohibited from venturing outside the scheme established by the Federal Bank Robbery Act in order “to circumvent [that] scheme’s carefully crafted hierarchy of penalties.”5
The immediate concern of the court in Canty was to limit punishment for crimes cognizable under the federal act. But as the majority opinion points out, there is language in Canty to suggest that charges under the federal act and the local code covering the same conduct may not be joined.6 Indeed, Canty’s characterization of the federal act as “a comprehensive scheme for prosecuting and punishing persons who rob federally-insured banks”7 supports a broad inference that Congress did not intend the local code to apply at all to offenses dealt with in that act. Such a reading would not only prevent joinder of federal and local counts in federal district court, it would also rule out prosecution in D.C. court of local code offenses relating to bank robberies.
A construction of the Federal Bank Robbery Act as superseding catchall provisions of the D.C.Code is not prevented by any express statement of Congress to the contrary. The majority’s rejection of it, despite the broader implications of Canty, appears to have two major bases. The first of these is a review of the Federal Bank Robbery Act’s legislative history and relevant case law8 which concludes that federal jurisdiction over offenses dealt with in the Act does not preclude state prosecution of the same offenses. The short answer to this approach is that the District of Columbia is not a state. As will become evident from subsequent discussion, there are grounds on which to distinguish between questions of a statute’s exclusivity as to the states and its exclusivity as to D.C.9 A second basis for the majority’s construction is the anomaly that would exist under a reading favorable to petitioners:
if someone committed an armed robbery in the District of Columbia of any type of establishment other than a bank or S&L, he would be subject to the maximum sentence of life imprisonment under the local statute, whereas if that person confined himself to the armed robbery of federally-insured banks or savings and loan associations located in D.C., he could be punished only under the less harsh federal statute.10
*1338The opinion reasons that as Congress cannot reasonably have intended such a result, the Federal Bank Robbery Act and the D.C.Code must be read as “meshing together.” The prosecutor in D.C. retains discretion to prosecute bank robberies under either federal or local law, or under both at the same time, as he chose to do in this case.
But nowhere does the majority opinion acknowledge the anomaly that is created by its reading of the statutes as “meshing together.” Dual applicability of the statutes will mean that a bank robber in D.C. will be subject to a maximum sentence of life imprisonment, whereas in any other jurisdiction maximum punishment under congressional enactment would be only twenty-five years. In adopting the majority’s reading we are saying in effect that Congress intended to subject D.C. bank robbers to greater potential punishment than bank robbers anywhere else in the country. In my view, a court might reasonably refuse to reach such a conclusion in the absence of an express statement of legislative intent to the contrary, reversing the presumption indulged in by the majority opinion.
In suggesting this, I do not pretend that the inequity cited by the court, and relied on in its construction of the statute, does not exist, or even that it is insignificant. But it is at least neutral as to the political status of those affected. The inequity created by the court’s reading is not. As we said in United States v. Thompson,
What is involved here is not just any arbitrary classification between two similarly situated groups. The residents of Washington occupy a profoundly anomalous position in the federal system, [because they do not send a voting representative to Congress,] and any classification which discriminates against them is particularly suspect.11
We concluded that “it is not enough for [discriminatory] classifications [affecting D.C. residents] to be merely rational or even plausible; the justification offered must be convincing.”12 These principles informed our approach to the question of whether the bail provisions of the D.C. Court Reform Act applied to those convicted in D.C. of violating federal criminal laws. Finding no adequate justification for subjecting the bail release applications of D.C. offenders to standards different from those applicable elsewhere in the federal court system, we based our construction of the statute on a presumption that the Government’s reading was erroneous.
There exists a possible ground for distinguishing Thompson. There we dealt with discrimination against those in D.C. convicted of federal crimes; in this case we are concerned with discrimination against those convicted of violating laws exclusively local in application. This distinction forms the basis for an argument that the equal protection concerns stressed in Thompson simply are not relevant here.
In addition to its responsibilities as the national law maker, Congress is empowered by the Constitution “[t]o exercise exclusive Legislation in all Cases whatsoever” in the District of Columbia.13 As the Supreme Court declared recently in Palmore v. United States,
The power is plenary. Not only may statutes of Congress of otherwise nationwide application be applied to the District of Columbia, but Congress may also exercise all the police and regulatory powers which a state legislature or municipal government would have in legislating for state or local purposes.14
*1339This unique aspect of Congress’ power was relied on in rejecting the equal protection arguments of petitioners in United States v. McDonald,15 which upheld a provision of the federal Youth Offenders Act denying coverage to those convicted to violating provisions of the D.C.Code. It was also at the core of the court’s analysis in United States v. Greene,16 which upheld the local felony murder statute’s inclusion of U.S.Code offenses.
The principle enunciated in Palmore supports the proposition which I stated in an opinion concurring and dissenting in McDonald, in a passage quoted with approval by the majority opinion in this case:
in enacting a separate body of law for the District Congress apparently also has the power to enact laws substantially identical to its national legislation, to provide differing penalties for their violation despite that identity, and to require that both sets of laws be enforced in the District of Columbia.17
But this does not mean that in exercising its dual functions Congress is exempt from the guarantee of equal protection implicit in the due process clause of the Fifth Amendment.18 Skirting Thompson requires a further conceptual step, which amounts to considering “Congress as two distinct legislative bodies, one exercising the police power of a sovereign state over the District of Columbia, and the other the limited power of the federal government.” 19 This legal fiction, I argued in McDonald, is not required by necessity.20 And it moves us further in a direction contrary to basic principles of democratic government, presenting the stark apparition of a “local” legislature which does not contain a single voting member elected by the people of the locality. So long as the residents of D.C. are denied a vote in congressional elections, it would seem incumbent on the courts to enforce the minimal controls on legislative activity stemming from the Constitution to their fullest extent, rather than to fabricate devices which render those controls inoperative.21
I press this analysis with considerable diffidence at this point.22 It has been rejected, by implication at least, in McDonald, Greene, and more recently in United States v. Caldwell,23 which plainly held, albeit in a footnote, that the Federal Bank Robbery Act does not bar prosecution of bank robbery and related offenses under the D.C.Code. Perhaps there are reasons which have yet to be put forward, or which I have yet to properly appreciate, that do make Thompson irrelevant here. And even if Thompson is relevant, there may be justification for the discriminatory'classifications in cases such as this that did not *1340exist in Thompson. But after careful review of the reasoning of the majority opinion, and the opinions on which it relies, petitioner’s contentions continue to present for me questions of serious import which have not been satisfactorily answered.
Estopped by prior decisions, in particular the holding in Caldwell, from carrying Canty to what I believe is its logical conclusion, I am left to dispute the majority’s reading of that decision on a more limited basis. The majority opinion states, and correctly so, that Canty prohibits multiple convictions for armed robbery and bank robbery resulting from the same activity. But the opinion goes on to state that the “requirements of Canty were of no further relevance once the federal count of the indictment was dismissed.”24 This leaves open the possibility that Shepard could have been sentenced to a term of life in prison under the D.C.Code (as against a maximum of twenty-five years under the federal act), so long as he was not convicted of the federal charge for which he was indicted.
Leaving open this possibility is inconsistent with Canty’s broad charge that the prosecutor not be allowed to “circumvent the [federal] scheme’s carefully crafted hierarchy of penalties.”25 Caid-well requires us to recognize two broad options on the part of the prosecutor. He may invoke the federal act in district court, with or without joinder of counts under the local code, or he may ignore the federal statute entirely and proceed under the local code in D.C. Superior Court. If he chooses the latter alternative, the defendant is vulnerable to the full penalty provided for by the local code. Once the prosecutor chooses the former course, however, it seems to me that the circumvention of the federal scheme becomes a relevant concern, regardless of whether the prosecutor ultimately attains multiple convictions, as he did in Canty. The central thrust of Canty is that the pyramiding of penalties in excess of the ceiling under the federal act must be avoided. The multiple convictions involved in Canty are simply one device by which the Government might achieve this impermissible goal. Joinder of charges with dismissal prior to judgment is another. Canty is properly read to prevent the use of either toward that end.26
My difference with the majority does not affect the result. Shepard was sentenced under the local armed robbery statute to a term of 4 to 20 years. Since this is less than the term he could have received under the Federal Bank Robbery Act, in this particular case Canty, as I understand it, was not violated.
. 152 U.S.App.D.C. 103, 469 F.2d 114 (1972).
. 18 U.S.C. § 2113(a).
. D.C.Code § 22-502.
. 18 U.S.C. § 2113(d).
. 469 F.2d at 128.
. Draft Opinion at 19. See United States v. Canty, supra note 1, at 127, 129.
. 469 F.2d at 127.
. 350 U.S. 497, 76 S.Ct. 477, 100 L.Ed. 640. (1956). See Majority Opinion 169 U.S.App.D.C. at ---, 515 F.2d at 1333-1334.
. See note 21 & text accompanying notes 11-21 infra.
. Draft Opinion at 20.
. 147 U.S.App.D.C. 1, 452 F.2d 1333, 1340 (1971).
. Id at 1341.
. U.S.Const., art. I, § 8, cl. 17.
. 411 U.S. 389, 397, 93 S.Ct. 1670, 1676, 36 L.Ed.2d 342 (1973).
. 156 U.S.App.D.C. 338, 481 F.2d 513 (1973).
. 160 U.S.App.D.C. 21, 489 F.2d 1145 (1973).
. 481 F.2d at 523.
. See generally Washington v. Legrant, 394 U.S. 618, 89 S.Ct. 1322, 22 L.Ed.2d 600 (1968) (consolidated with and reported sub nom. Shapiro v. Thompson); Bolling v. Sharpe, 347 U.S. 497, 74 S.Ct. 693, 98 L.Ed. 884 (1954).
. United States v. McDonald, supra note 15, at 523 (Bazelon, C. J., concurring in part and dissenting in part).
. Id. at 524.
. This analysis should make clear the basis of' my view that a holding for petitioners would be consistent with decisions that the Federal Bank Robbery Act is not exclusive as to the states. See text accompanying note 9 supra.
. I have expressed substantially the same position previously in United States v. Greene, supra note 16, at 1159 (Statement of Bazelon, C. J., as to why he would grant rehearing en banc), and United States v. McDonald, supra note 15, at 518 (Bazelon, C. J., concurring in part and dissenting in part). And in an opinion for the majority in United States v. Brown, 157 U.S.App.D.C. 311, 483 F.2d 1314 (1973), I found that Thompson -based considerations “might also apply” to the issue before us in that case: whether D.C.Code bail provisions should be read to apply to local offenders in federal court. The Government’s reading was rejected on a rationale which did not depend on the Thompson presumption, although constitutional concerns formed the background against which the issue was decided.
. No. 72-1513 (D.C.Cir. Dec. 31, 1974), slip op. at 65 n. 176.
. Draft Opinion at 19.
. 469 F.2d at 128.
. This position is not in conflict with our recent holding in United States v. Knight, 166 U.S.App.D.C. 21, 509 F.2d 354 (1974). Appellants in that case were convicted and sentenced under both the federal mail robbery statute and the local armed robbery provision. Each was sentenced to from three to ten years under the federal code; one was sentenced to from 4 to 12 years under the local code and the other to from 10 to 30 years — all sentences to run concurrently. Maximum sentence under the mail robbery statute* was ten years. The court found that Canty prevented the dual convictions, but at the same time it held that on remand the district court had discretion to sentence either under the federal statute or under the local code — without restrictions keyed to the 10-year ceiling under the federal scheme. This holding turned on a determination of congressional intent with regard to the particular provisions involved and therefore' does not govern the instant case; indeed the court itself in Knight suggested a possible basis for distinguishing the cases in acknowledging that ‘[i]n the case of the mail robbery statute, 18 U.S.C. § 2114, the scheme is not as comprehensive as that of the federal bank robbery statute.” 509 F.2d at 362. To the degree that the court in Knight considered the mail robbery statute comparable in all relevant respects to the Bank Robbery Act, its result is simply inconsistent with the disposition in Canty. For the Canty court, in remanding to the district court, vacated the local assault count and allowed sentencing only under the Bank Robbery Act and the local count of carrying a dangerous weapon, which was specifically found to lie outside the coverage of the federal scheme. This action was an implicit denial of the discretion recognized in Knight.