White House Vigil for the Era Committee v. William P. Clark, Secretary of the Interior

WALD, Circuit Judge,

concurring in the judgment in part and dissenting in part:

These cases are never easy. The nation has a paramount interest in the safety of its President, and judges must respect the experience and knowledge of the law enforcement agencies charged with protecting the President, his family, and others who live and work in the White House. In reviewing decisions about their security we must proceed with care.

*1542 Nonetheless, our duty requires us ultimately to weigh for ourselves the merits of a first amendment challenge to agency regulations in this sensitive area. The appellees in this case, all of whom frequently demonstrate on the White House sidewalk, argue that the government has needlessly infringed their right “peaceably to assemble, and to petition the Government for a redress of grievances” at this center of executive power, and the district court sustained that view. While I agree with the majority that most of the Park Service regulations are not unconstitutional as written, in my view the majority opinion does not adequately dignify the constitutional interests at stake in this claim, and does not examine those interests with sufficient particularity. The majority tautologically informs us that if a regulation “lies within the zone prescribed by the first amendment it is constitutional,” Maj. Op. at 1532, but unfortunately tells us very little about where the boundaries of that “zone” are found.

Prior cases of the Supreme Court and this court concerning time, place, and manner restrictions do not permit such uncritical deference to agency decisionmaking. Because we are deciding the content of constitutional principles with significant effect on future cases, I believe our differences merit the discussion that follows.

I. The Legal Standards

I have problems with the majority’s legal analysis in three respects. First, I believe it discounts the district court’s primary obligation to do the constitutional balancing for itself, and as a result overrates the appropriate scope of our appellate review. Second, it does not heed closely enough the requirement that time, place, and manner restrictions be “narrowly tailored.” Recent Supreme Court cases on which the majority relies expressly reaffirm this test. Yet I read in the majority opinion hints that the government — as a result of those cases — now has greater latitude than previously to burden protected expression. As I parse those cases, however, they refine but do not revamp settled principles in first amendment law that govern our decision here. Finally, I disagree with the majority’s analysis of the limited responses available to federal judges when they decide regulations are unconstitutional in part.

A

The majority explains its refusal to review the district court’s findings under the usual Fed.R.Civ.P. 52 “clearly erroneous” standard by asserting that

The issue for decision on this appeal is not factual, it is legal: did the Park Service draft regulations that were “narrowly tailored to serve a significant government interest”? The agency in this case was the institution charged with the principal resolution of factual issues; the court’s role was limited to determining whether the regulations which the agency adopted were within the boundaries of constitutionality prescribed by the first amendment.

Maj. Op., text at n. 83 (emphasis in original). The majority appears to reason that (1) the district court should have reviewed agency factfinding under a more deferential standard; (2) the district court’s own “findings” are really conclusions of law, since it had no independent factfinding responsibility; and (3) therefore those “findings” are reviewable on appeal under the liberal standard for assessing legal error. I do not agree that this is an accurate statement of what a trial or appellate judge’s responsibility is in such cases.

Judge Leventhal offered a balanced appraisal of judicial responsibility in first amendment cases.

[Tjhis case is not a normal review of an executive action or administrative proceeding. When the executive or the administrative process abridges constitutional rights, it is subject to closer scrutiny than otherwise, and ultimately it is *1543 the court rather than the agency that must balance the competing interests. The question in this case is not whether some support for the regulations may be adduced, by reference to evidence in the record and a claim of reasonable inferences or concerns, but is whether the regulations at issue here are “unnecessarily restrictive for the purpose they are designed to serve.”

We now have ... a judicial determination based upon factual evidence adduced at a trial, indeed at a rather extensive and complete trial. To this determination we owe greater deference than to the untested administrative judgments with which we have been previously confronted. This decision being appealed was rendered by a district judge after consideration of both constitutional considerations, for which a judge has special concern, and the security considerations brought forward by the government officials. Moreover, the district judge had the benefit of the live testimony of the various witnesses whose assertions could be tested and probed on cross-examination. Thus, unless we discern clear error in the district court’s findings of fact, or a mistake in its legal approach, we have no warrant for reversal.

A Quaker Action Group v. Morton, 516 F.2d 717, 723-24 (D.C.Cir.1975) (footnote omitted) (quoting A Quaker Action Group v. Morton, 460 F.2d 854, 860 (D.C.Cir. 1971)); see also Women Strike for Peace v. Morton, 472 F.2d 1273, 1289 (D.C.Cir.1972) (opinion of Wright, J.). 1

In this case, the district court conducted a trial de novo, during which it exhaustively reviewed evidence relevant to the constitutionality of the regulations, including evidence offered by the governmental agencies. It was required to give the government’s witnesses the attention their expertise warranted. But that evidence still had to undergo the judge’s independent appraisal and judgment, and his factual findings deserve our approval unless clearly erroneous.

B

As the majority rightly notes, time, place, and manner restrictions must be justified without regard to the content of the message expressed; must be narrowly tailored to serve a significant government interest; and must leave open ample alternative channels of communication. See, e.g., Clark v. Community for Creative Non-Violence, — U.S. -, 104 S.Ct. 3065, 3069 , 82 L.Ed.2d 221 (1984). The crux of this case involves the second element of the test.

United States v. O’Brien, 391 U.S. 367 , 88 S.Ct. 1673 , 20 L.Ed.2d 672 (1968), required that the incidental restrictions a regulation imposes on protected expression be no broader than is essential to the furtherance of the governmental interest at stake. Id. at 382 , 88 S.Ct. at 1681 . 2 That standard has been further broken down into two subsidiary inquiries. First, the challenged regulation must not unnecessarily contain provisions that entirely fail to advance the relevant governmental interest. If a regulation prohibits an identifiable class of expressive activity that does not pose any threat of the evil against which the regulation is directed, the courts will declare the *1544 regulation unconstitutional as it applies to that class of expression. See Ely, Flag Desecration: A Case Study in the Roles of Categorization and Balancing in First Amendment Analysis, 88 Harv.L.Rev. 1482, 1485-90 (1975).

The second aspect of O’Brien’s “narrow tailoring” requirement looks to see if an alternative regulation would serve the government’s interest nearly as efficiently but would be demonstrably less intrusive on protected expression. Of course, any regulation could be made a little less intrusive on speech, at the cost of a little more protection for first amendment concerns. In this case, for example, adding six inches to the maximum permissible sign dimensions would surrender some marginal protection for security interests for a marginal benefit to free expression, but this sort of whittling is not what the “narrowly tailored” requirement is about. Instead, the court must look to see if the burden on speech is approaching an unreasonable level, or a serious loss to speech is being imposed for a disproportionately small governmental gain. 3

The government here offers two purposes for its regulations: a compelling interest involving security of the White House occupants and the law enforcement officers and individuals on its sidewalks, and a more limited aesthetic interest in an unobstructed view of the White House for visitors.

Before considering the regulations in detail, however, I want to register my disagreement with an insistent theme in the majority opinion that the Supreme Court’s recent decisions have changed the character or the mood of appropriate judicial scrutiny for time, place, and manner restrictions. See, e.g., Maj. Op., text at n. 95. I believe those decisions are consistent with the O’Brien framework outlined above.

The recent Supreme Court pronouncements in Clark v. Community for Creative Non-Violence, — U.S. -, 104 S.Ct. 3065 , 82 L.Ed.2d 221 (1984) and Regan v. Time, Inc., — U.S.-, 104 S.Ct. 3262 , 82 L.Ed.2d 487 (1984), emphasize that a time, place, and manner regulation is not unconstitutional simply because some alternative regulation would, on the facts of the case before the court, satisfy the government’s aims equally well and yet not restrict the expressive rights of that particular challenger. Regan yields a corollary to that principle: A time, place, and manner regulation is not unconstitutional as applied to situations that do not threaten the governmental interest at stake if that application is an unavoidable consequence of regulating other conduct that does threaten the interest at stake. In other words, if a regulation cannot reasonably be drafted so as to prohibit all the conduct the state really needs to suppress, without marginally prohibiting some expressive activity that is harmless, it will pass muster. It seems to me the majority’s approach to interpreting Clark and Regan blurs these well-defined principles into a far more diffuse deference to the government.

In Clark , demonstrators for the homeless challenged the constitutionality of a Park Service regulation forbidding overnight camping in the park. The Court rejected the argument that the Park Service should be required to adopt some other regulatory scheme to protect the parklands from overuse without forbidding sleeping by those demonstrators, such as restricting the size, duration, or frequency of demon *1545 strations. The Court observed that such measures “would still curtail the total allowable expression in which demonstrators could engage, whether by sleeping or otherwise,” and concluded that “these suggestions represent no more than a disagreement with the Park Service over how much protection the core parks require or how an acceptable level of preservation is to be attained.” Id. 104 S.Ct. at 3072. The proposed alternative regulations were less speech-restrictive on the facts presented in Clark ; but viewed more generally, they were simply another regulatory scheme that would have allowed more expressive activity in some situations, and less expressive activity in others.

In Regan , Time magazine challenged an anti-counterfeiting statute prohibiting, among other things, reproducing United States currency in color. See id. 104 S.Ct. at 3264-65 (quoting 18 U.S.C. § 474 If 6, § 504). The Court rejected Time’s arguments that the color ban was too broad because it included photographs so distorted that they were entirely incapable of aiding counterfeiting. Id. at 3274 & nn. 14-15.

Writing for a plurality of four, Justice White commented:

That the limitations may apply to some photographs that are themselves of no use to counterfeiters does not invalidate the legislation. The less-restrictive-alternative analysis invoked by Time has never been a part of the inquiry into the validity of a time, place, and manner regulation. It is enough that the color restriction substantially serves the Government’s legitimate ends.

Id. at 3271-72 (footnote omitted). In a footnote to this paragraph, Justice White stated that “[i]f Time is exempted from the color requirement, so must all others who wish to use such reproductions. While Time may consistently use negatives and plates that are of little use to counterfeiters, there is no way of ensuring that others will adhere to that practice.” Id. at 3272 n. 12. 4 In sustaining the color provision, Justice White thus relied on the notion that the government could not frame a narrower statute that adequately protected against the evil to be prevented. 5

It is crucial, I believe, to consider the exact context in which Justice White wrote his rejection of least speech-restrictive analysis in time, place, and manner restrictions. In Regan , because no alternative was available that could have prohibited only the speech that the government had a legitimate interest in suppressing, the Court ruled the statute could be enforced *1546 in all its applications. 6 Clark ruled that a challenger to a time, place, and manner restriction cannot win by merely conjuring up an alternative regulation that does not prohibit its conduct, regardless of the alternative regulation’s effects on the expression of others. Seen in that light, the basic structure of the O’Brien test is still alive and well, and must be applied to these regulations.

C

The majority informs us that “it is not the province of the court to ‘finetune’ the regulations so as to institute the single regulatory option the court personally considers most desirable,” and that the role of courts “is to uphold regulations which are constitutional and to strike down those which are not.” Maj. Op., text at n. 83. Based on these unexceptionable generalizations, the majority expresses disapproval that “not only did [the district court] uphold some restrictions and reject others, it modified the content of individual provisions.” Maj.Op., text at n. 32; see also id. n. 97. I do not share in that disapproval, for it seems to me the district court did exactly what it had to, assuming that its judgment of partial unconstitutionality was a correct one. The trial judge here simply indicated at what point he believed a regulation strayed over the bounds of constitutionality. In so doing he did not “rewrite” the regulation, but only elucidated what the results of his constitutional balancing permitted. The agency is always free to withdraw the regulations altogether rather than amend or apply them to conform to his views. Indeed, it can prepare new and different ones. The trial court's judgment is a clearcut one — that the regulation as written is or is not constitutional or that it may be applied to some but not other situations. For example, in United States v. Grace, 461 U.S. 171 , 103 S.Ct. 1702 , 75 L.Ed.2d 736 (1984), the Supreme Court considered the constitutionality of a statute barring certain demonstrations in the Supreme Court building or on its grounds, which included the surrounding sidewalks. The statute did not distinguish between the Supreme Court sidewalk and the rest of the building grounds, but the Court had no difficulty in finding the statute unconstitutional only as it applied to the sidewalk. See id. 103 S.Ct. at 1706, 1710 . 7 Here, the district judge declared the parcels and sign attendance regulations unconstitutional as they applied to parcels and signs within five feet of the person owning or controlling them, but otherwise constitutional. See White House Vigil for the ERA Com *1547 mittee v. Clark, No. 83-1243 (D.D.C. Apr. 26, 1984) (order). If one was “rewriting,” so was the other. I believe both courts were acting in time-honored fashion to decide whether the unconstitutional portion of a statute can be pared away without unduly disrupting the intended regulatory plan.

II. The Regulations

A

The first group of contested regulations concern the material, size, and placement of signs on the White House sidewalk. The regulations provide that:

No signs or placards shall be permitted on the White House sidewalk except those made of cardboard, posterboard or cloth having dimensions no greater than three feet in width, twenty feet in length, and one-quarter inch in thickness. No supports shall be permitted for signs or placards except those made of wood having cross-sectional dimensions no greater than three-quarter of an inch by three-quarter of an inch. Stationary signs or placards shall be no closer than three feet from the White House sidewalk fence. All signs and placards shall be attended at all times that they remain on the White House sidewalk. Signs or placards shall be considered to be attended only when they are in physical contact with a person. No signs or placards shall be tied, fastened, or otherwise attached to or leaned against the White House fence, lamp posts or other structures on the White House sidewalk.

36 C.F.R. § 50.19 (e)(9) (1983).

According to the demonstators, the main problem with the “materials” provision, requiring all signs to be made of cardboard, posterboard, or cloth, is that it bans plywood signs which are more sturdy and durable and pose no security hazards. The government said that plywood signs could be used to scale the White House fence or as shields or weapons in fights, and that splinters from such signs would be dangerous in the event of an explosion.

After considering the evidence, the district court found that, given its decision to uphold the regulation limiting the thickness of all signs to one-fourth inch,

The total ban on signs made of wood is unjustified. It appears clear to the court that any sign with a thickness not exceeding Vi inch, including one made of wood, would be of no assistance to anyone bent upon scaling the fence. As a matter of fact, it would be an impediment and utterly foolhardy for one to attempt to make use of such a sign for that purpose. Contrary testimony is incredible. The possibility that sheet explosives might be concealed in wooden signs is adequately dealt with by the thickness requirement of not more than Vi inch on all signs. The testimony is that the thinnest sheet explosive is itself Vi inch, and that a sign of that thickness regardless of its composition has no capacity to conceal such explosives. *1551 No signs or placards shall be held, placed ■ or set down on the center portion of the White House sidewalk, comprising ten yards on either side of the center point of the sidewalk; Provided, however, that individuals may demonstrate while carrying signs on that portion of the sidewalk if they continue to move along the sidewalk.

*1547 White House Vigil for the ERA Committee v. Clark, No. 83-1243, slip op. at 23-24 (D.D.C. Apr. 26, 1984) (emphasis added). The district court also found no evidence that any sign had ever been used to scale the White House fence, and that in any event it can be — and has been — readily scaled without the assistance of a flimsy plywood sheet. See id. at 22-23 . After a careful review of the evidence, I cannot see any grounds for overriding the district court’s conclusion that based on weight and credibility of the testimony, as well as undisputed historical facts, the government made out no case that plywood signs, one-fourth inch thick, posed a danger to the security of White House occupants. 8 The potential use of plywood signs as weapons *1548 and the minimization of flying debris in an explosion involve security on the sidewalk itself. While that interest is certainly a legitimate one, it is far from unique to the White House sidewalk. The government’s evidence on any special sidewalk dangers of plywood signs, as opposed to those made of plexiglass, was entirely speculative and not sufficient to overcome the appellees’ considerable testimony that cloth and cardboard signs would not withstand prolonged use and would collapse or disintegrate in bad weather. 9 I therefore agree with the district court that the government’s evidence failed to establish any significant relationship between its interest in maintaining order on the sidewalk and this part of the materials ban. Because in this instance the regulation prohibits a discrete, readily segregated class of expressive activity, i.e., the display of plywood signs that present no adequately documented danger to the interests asserted by the government, I would uphold the trial court’s ruling that the materials ban is unconstitutional as applied to plywood signs.

On the other hand, I agree with the majority that the restriction on size and placement of signs and on the type of sign supports deserved to be upheld. The district court invalidated the restriction on large signs on the ground that the Secret Service can adequately survey crowds on *1549 the sidewalk through elevated observation posts and closed circuit television cameras. See White House Vigil for the ERA Committee v. Clark, No. 83-1243, slip op. at 21-22 (D.D.C. Apr. 26,1984). Law enforcement officers testified that while a few large signs might not create a security hazard, many of them would obscure the ability of officers patrolling on the sidewalk to observe the activities of persons in front of the White House. See, e.g., Lindsey Tr., R. 162 at 164-65, 233-36. Even accepting the district court’s conclusion that demonstrators’ signs would not block perusal of the sidewalk from other vantage points, I nonetheless believe the district court gave too little weight to the government’s legitimate interest in assuring that the officers actually on the scene — those making immediate decisions about crowd management, rather than those on platforms or watching television monitors— also be aware of what is going on. Cf. White House Vigil for the ERA Committee v. Watt, 717 F.2d 568, 572 (D.C.Cir. 1983) (discussing dangers of very large signs). This interest also implicates building security to some degree, since the government seeks to prevent the firing of any dangerous objects towards the White House, not merely to observe it. Those in the best position to act immediately will often be the sidewalk officers. Finally, this regulation does not by itself impose an overwhelming burden on expression: a banner measuring three feet by twenty feet is a large and very visible one by most viewers’ standards.

Sign supports are restricted to those no larger than three-quarters of an inch by three-quarters of an inch and made of wood. Government security experts testified that the commonly used aluminum hollow tubular sign supports could be used to fire projectiles or conceal explosives. See Parr Tr., R. 163 at 35-37. This testimony was somewhat undercut by other testimony that the smallest rocket launcher now developed is substantially larger than the largest support authorized by the regulation. Nonetheless, the regulation was not designed to protect against projectile danger alone. Considering the stakes at issue, the district court failed to adequately credit the government’s interest in preventing serious harm. The danger that hollow supports might conceal potential weapons like blast marbles and flares, possibly even explosives, was, according to the testimony, a more realistic concern. 10

In addition, concerns that large supports could be used as weapons have prompted the District of Columbia to regulate the size of sign supports on the public streets, and that danger must be taken into account here as well. See Hensdill Dep., R. 145D at 63-64. The testimony did show that the regulation would inconvenience some demonstrators. Large banners can more easily be displayed with stronger supports. 11 But even crediting all of the appellees’ evidence, as the district court apparently did, I do not think that the burden this regulation imposes on expression outweighs the dangers the regulation addresses.

The regulations also ban stationary signs within three feet of the White House fence and attaching signs to, or leaning signs against, the fence or other structures on *1550 the White House sidewalk. The justification for these regulations is that signs close to the fence create a triangular space bounded by the sidewalk, the solid ledge at the base of the White House fence, and the sign, in which explosives or contraband could be concealed. See, e.g., Parr Tr., R. 162 at 52-53. The government’s principal expert witness on security matters admitted that the three-foot regulation bans some conduct that presents no threat to security, such as simply holding a sign aloft in the zone without obscuring the area under the ledge. 12 But the government claimed that a regulation requiring signs to be held aloft in this area would require continuous surveillance of each demonstrator, 13 and would likely require frequent admonitions from police officers to enforce, possibly precipitating confronta-. tions between demonstrators and police.

The challengers alleged substantial burdens stemming from the placement restrictions. They prevented demonstrators with signs but not others from sitting on the ledge, whether or not the signs are located in a way that enables anyone to conceal objects behind them.

A regulation is valid even if it unavoidably prohibits harmless conduct in order to cover similar but dangerous conduct. Despite qualms, I believe that on essentially undisputed facts, the government showed that the three-foot regulation was directed to a substantial danger, and that a narrower regulation might well be less effective because of difficulties with its enforcement. For me, however, the question is a very close one.

I have less trouble with the leaning ban. Appellees urge that police officers could look around signs, moving leaning signs to inspect beneath them, or use mechanical devices and dogs to “sniff” for explosives. But that answer does not seem adequate under the OBrien-Clark test where a substantial harm is shown.

Finally, the government supports its regulation requiring that signs be attended at all times by arguing that unattended signs create an opportunity for others to place explosives in or under them. The district court’s factual findings on this issue are unclear, 14 but it recognized the substantiality of the government’s concern by upholding the regulation to the extent of requiring that demonstrators remain within five feet of their signs. The government objects to this compromise, pointing out that in a crowd, police officers cannot quickly determine what sign belongs to what individual. The district court’s resolution might work well most of the time, but serious harm could be threatened where a large crowd on the sidewalk made it impossible to identify a sign owner within five feet. The government is entitled to promulgate regulations broad enough to reach all instances in which a grave danger is threatened, even though they incidentally affect speech. On that principle this regulation may be sustained.

B

The center zone restriction provides:

*1551 36 C.P.R. § 50.19(e)(9) (1983).

I have no quarrel with the majority’s view that aesthetic concerns may justify some time, place, and manner restrictions on expression. However, it is equally clear that courts must be especially careful in scrutinizing restrictions on first amendment expression that the government seeks to justify on eye-pleasing grounds. Aesthetic concerns will in close cases involving first amendments rights weigh in at a lower poundage than, say, public safety or national security considerations. Because of their subjective nature, aesthetic concerns are easily manipulated, and not generally susceptible of objective proof. The danger is not just, as the majority suggests, that government might adopt an aesthetic rationale as a pretext for an impermissible motive, but rather that so many forms of robust expression are by their very nature boisterous, untidy, unsightly, and downright unpleasant for unsympathetic viewers. Distaste for the vig- or with which a message is asserted can too easily be cast as an aesthetic interest in compelling others to be more moderate and decorous — and, in consequence, less effective — in conveying their message.

If, as I agree, aesthetics are nonetheless to be recognized as legitimate governmental objectives, we must face squarely the implications of applying them to first amendment eases. The government has justified its 20-yard picture window regulation 'by citing the aesthetic interests of the visiting public in being able to see and photograph the White House against a tranquil foreground. The majority upholds this regulation in part because the government has regulated in accordance with the public’s aesthetic views, not just its own. I am not entirely sure of the utility of this distinction in first amendment analysis, in light of that amendment’s traditional function of protecting unpopular minorities against majoritarian excesses. But in this case, the government has an obvious interest of its own: its natural ambivalence toward the existence of vociferous demonstrators at the very gates of the White House, attracting news coverage and often raising unwelcome complaints about administration policies. We must therefore examine the government’s asserted purpose and the efficacy of this regulation in satisfying that purpose with particular care.

The evidence showed that the main complaints from members of the public involved the proliferation of large, billboard-like signs left for extensive periods propped up against the White House fence. Under separate portions of the regulations upheld today, that complaint is assuaged; the maximum vertical dimension of signs is now three feet, and signs cannot be left unattended or leaned against the fence. 15 Photographs submitted by the government show that the view of the White House from the street is not seriously obscured by demonstrators with signs of these modest dimensions. See Defendants’ Exhibits AAAA(1)-(14), Joint Appendix at 332-45. There was also evidence that demonstrators are generally cooperative in moving out of the way if tourists want to take unobstructed photographs. But most perplexing, the twenty-yard zone is closed only *1552 to stationary demonstrators with signs, not moving demonstrators carrying the same signs, although the latter could surely block as much of the view as much of the time. 16 I agree with the district judge that in the aggregate, the evidence does not come near to demonstrating that a prohibition on stationary signs of the kind allowed under the new regulations will add discernibly to aesthetic enjoyment of the White House. To the extent that ordinary visitors carrying possessions who stop within this zone obstruct the view as much as demonstrators, the regulation moves perilously close to selectively penalizing those who visit the White House for the purpose of political demonstrations.

On the other hand, the record revealed that the regulation does impose a real burden on demonstrators. The majority brushes away their concerns with the declaration that there is no first amendment right to media attention, but I do not think that is the pertinent inquiry. When government bans stationary demonstrators from one section of a uniquely important public forum, it has obviously burdened their speech rights. If, as the district court found on ample evidence, the media are most likely to cover a demonstration in the center zone, then that fact is relevant in determining the extent of the burden and the need for its justification. It is not that protesters have an absolute right to the prime spot, but that the government must have an acceptable reason for excluding them from it or regulating the way they protest in it, and the means it chooses to implement its goal must be geared to achieve that end. The restriction on stationary as opposed to moving demonstrators, on the basis of aesthetics alone, does not, in my view, meet these requirements of the “narrowly tailored” test.

C

The parcels regulation provides that:

No parcel, container, package, bundle or other property shall be placed or stored on the White House sidewalk ... Provided, however, that such property, except structures, may be momentarily placed or set down in the immediate presence of the owner on those sidewalks.

36 C.F.R. § 50.19 (e)(10) (1983). While not deciding the issue, the majority expresses doubt that “the first amendment protects the conduct proscribed by the parcels regulation.” Maj. Op., text at n. 127. I believe it does.

As a general matter, carrying parcels is not, of course, “speech” within the meaning of the first amendment. However, the conduct this regulation prohibits not only arises in the immediate course of a demonstration, but according to the district court is necessary if some demonstrators are to convey their messages at all. See White House Vigil for the ERA Committee v. Clark, No. 83-1243, slip op. at 26 (D.D.C. Apr. 26, 1984). A regulation on facilitative conduct that cuts off or sharply restricts expression itself certainly burdens that expression.

I do not think we are required to ignore the fundamental proposition that it is people, with the basic needs of people, who exercise first amendment rights. Practically, old people, handicapped persons, mothers with children and children’s paraphernalia, and even young and unencumbered demonstrators can demonstrate or distribute literature only for limited periods if they are not permitted to put down their possessions more than “momentarily.” A demonstration is not some kind of ritualistic marathon dance, the prize dependent on how long a participant can stay on her feet and moving. In the trial below, an organizer for the White House Vigil for the ERA Committee testified that as a result of the parcel regulation, the Vigil had been forced *1553 to curtail the distribution of leaflets and petitions. According to the organizer, carrying literature, petitions, and clipboards on which people could sign the petitions, as well as banners, poles, and personal belongings on one’s person at all times, simply proved too burdensome. See Beall Tr., R. 170 at 121-22. Since organizers were not certain how many people would attend each demonstration, see id., it became difficult to gauge how much material to bring. We evidenced similar concerns about the effect of the regulation on distribution of leaflets in our opinion directing modification of the preliminary injunction. See White House Vigil for the ERA Committee v. Watt, 717 F.2d 568, 570 (D.C.Cir. 1983).

The first amendment looks to realities, not mere formalities. Justice Marshall, dissenting in Clark , said:

[FJacilitative conduct that is closely related to expressive activity is itself protected by First Amendment considerations____ [T]hat linkage, itself “suffices to require a genuine effort to balance the demonstrators’ interests against other concerns for which the government bears responsibility.”

Clark v. Community for Creative Non-Violence, — U.S.-, 104 S.Ct. 3065 , 3077 n. 7, 82 L.Ed.2d 221 (1984) (Marshall, J., dissenting) (quoting Community for Creative Non-Violence v. Watt, 703 F.2d 586, 607 (1983) (Ginsburg, J., concurring in the judgment)). 17 The government “cannot foreclose the exercise of constitutional rights by mere labels,” NAACP v. Button, 371 U.S. 415, 429 , 83 S.Ct. 328, 336 , 9 L.Ed.2d 405 (1963), and neither should it be able to lower the level of scrutiny of a law that directly and substantially abridges protected expression by calling it “facilitative.” If the first amendment does not permit the government to impose unreasonable restrictions on leafletting, surely it cannot suppress the same expression by over-regulating conduct that “facilitates” leafletting. Otherwise, government could more easily discourage the presence of people at a demonstration than regulate the signs they carry. A demonstrator forbidden to set down a receptacle in which leaflets are carried for more than a moment may lack the endurance to leaflet at all, or to do so for any appreciable time.

In other contexts, the Supreme Court has recognized that the first amendment requires attentive concern with protecting the conditions that are necessary for effective communication. Thus, the amendment offers protection against undue financial burdens on expression, see Minneapolis Star and Tribune Co. v. Minnesota Commissioner of Revenue, 460 U.S. 575 , 103 S.Ct. 1365 , 75 L.Ed.2d 295 (1983) (newspaper taxation); Citizens Against Rent Control v. Berkeley, 454 U.S. 290 , 102 S.Ct. 434 , 70 L.Ed.2d 492 (1981) (contributions to political committees concerned with referendum votes); Buckley v. Valeo, 424 U.S. 1 , 96 S.Ct. 612 , 46 L.Ed.2d 659 (1976) (campaign expenditures), as well as restrictions on access to certain highly newsworthy events, see Globe Newspaper Co. v. Superior Court for Norfolk, 457 U.S. 596 , 102 S.Ct. 2613 , 73 L.Ed.2d 248 (1982) (criminal trials); Richmond Newspapers, Inc. v. Virginia, 448 U.S. 555 , 100 S.Ct. 2814 , 65 L.Ed.2d 973 (1980) (same), and impermissible discrimination among ideas disseminated through public education, see Board of *1554 Education v. Pico, 457 U.S. 853 , 102 S.Ct. 2799 , 73 L.Ed.2d 435 (1982) (plurality opinion) (removal of books from school library). See generally Richmond Newspapers, 448 U.S. at 584 , 100 S.Ct. at 2831 (Brennan, J., concurring in the judgment). These cases, while obviously not controlling here, nevertheless suggest that the burdens of the parcel regulation should not be too casually dismissed in this case on the ground that they do not concern “speech.” “[I]n an area so closely touching our most precious freedoms,” NAACP v. Button, 371 U.S. at 438 , 83 S.Ct. at 340 , the definition of what impedes freedom of expression must be a pragmatic one. The imposition of a substantial, direct, and immediate burden on expression should be sufficient to invoke first amendment analysis. 18

So viewed, I do not believe the parcels regulation can be sustained. This regulation is clearly not a broad rule that, in only a small portion of its applications, has incidental effects on expression. Its application is limited to a public forum of unique national significance. Although its terms literally affect all visitors to the White House sidewalk, it will plainly have its major impact on political demonstrators who stay on the sidewalk for extended periods in order to demonstrate. Moreover, it was promulgated as part of a regulating plan targeted at expressive activities, and the burden it imposes on demonstrators must be considered not only individually but cumulatively in light of the other burdens that will be imposed by dint of these regulations on expression in the same forum.

As discussed above, the impact of this regulation will be substantial. The majority rationalizes that demonstrators can now keep their parcels across the street in Lafayette Park, see Maj.Op., text at n. 147, but even if this is feasible in group demonstrations, a bar against putting down personal or demonstration-related possessions for more than an instant at a time still represents a substantial burden, in some cases an impossible one, for individual demonstrators.

The government’s security experts testified that explosives can be concealed in abandoned parcels. See, e.g., Parr Tr., R. 162 at 117-19; Jones Dep., R. 193 at 62-64. The district court acknowledged the possibility but held this danger could be adequately dealt with by a less restrictive mandate that individuals remain within five feet of any parcels placed on the sidewalk. The government countered that in a large crowd law enforcement officers might not be able to identify the owners of parcels or even tell whether the owners were present within the five-foot zone.

Viewing the evidence offered by government security experts deferentially, it is still not apparent why a parcel on the ground is more dangerous than one held in the hand, provided that its owner is in immediate physical contact with it. Continuous physical contact is, after all, considered sufficiently safe for the care of signs under these regulations; it is unclear why it would not also suffice with parcels. The majority opinion may consider this “rewriting” the regulation, but I view it as in accord with a “narrowly tailored” O’Brien analysis that will not allow a regulation if it prohibits an identifiable, readily segregated class of conduct with no demonstrated relationship to the governmental purpose asserted. In my judgment, the parcels requirement, as written, violates that fundamental principle. It substantially im *1555 plicates a type of first amendment conduct that does not, on the evidence submitted, pose any genuine threat to the security of the White House.

Conclusion

I am concerned that the majority rationale defers too much to agency determinations and does not credit enough the district court’s time-honored function of reviewing the government’s evidence and making the primary balancing between its interests and first amendment protections. I agree, however, that many of these regulations meet the O’Brien “narrowly tailored” test and should be upheld. I respectfully dissent from the decision to uphold the materials regulation as applied to plywood signs, and the center zone and parcels regulations, as written.

. See also Pullman-Standard v. Swint, 456 U.S. 273, 285-88 , 102 S.Ct. 1781, 1788-89 , 72 L.Ed.2d 66 (1982) (reaffirming doctrine that factual findings, including findings on so-called "ultimate” facts, are reviewable on appeal only under "clearly erroneous” standard). Some constitutional questions, however, present special problems. Cf. Bose Corp. v. Consumers Union, — U.S. -, 104 S.Ct. 1949 , 80 L.Ed.2d 502 (1984).

. O’Brien concerned a restriction on what was arguably symbolic speech. See 391 U.S. at 376-77 , 88 S.Ct. at 1678-79 . However, the Supreme Court has held that the O'Brien standard governs time, place, and manner restrictions as well. See Clark v. Community jor Creative Non-Violence, — U.S.-, 104 S.Ct. 3065 , 3071 & n. 8, 82 L.Ed.2d 221 (1984); id. 104 S.Ct. at 3076 n. 6 (Marshall, J., dissenting).

. The handbilling cases are the classic examples: banning the public distribution of leaflets is an effective method of litter control, but its high cost to free speech and the existence of feasible alternatives, such as a prohibition on littering itself, make the tradeoff impermissible. See, e.g., Schneider v. State, 308 U.S. 147, 161-62 , 60 S.Ct. 146, 150-51 , 84 L.Ed. 155 (1939); see also Schad v. Mount Ephraim, 452 U.S. 61 , 67-71 & 70 n. 8, 101 S.Ct. 2176 , 2181-84 & 2183 n. 8, 68 L.Ed.2d 671 (1981) (discussing cases). These cases are also discussed in Ely, Flag Desecration: A Case Study in the Roles of Categorization and Balancing in First Amendment Analysis, 88 Harv.L.Rev. 1482, 1485-90 (1975).

. The majority thinks that Justice White’s footnote is unrelated to his rejection of less-restrictive-alternative analysis. See Maj.Op. at n. 93. I disagree.

In the footnote, Justice White was responding to Justice Brennan’s suggestion that "the particular negatives and plates used by Time would be of little assistance to counterfeiters,” Regan, 104 S.Ct. at 3272 n. 12 (opinion of White, J.) — an argument that was basically a variation of the less-restrictive-alternative analysis urged by Time. Justice White argued that exempting Time would require the government to exempt all those who wanted to engage in similar expression, including expression that might assist counterfeiters. But if exempting Time and all those similarly situated would not have impaired the governmental interest at stake, Justice White’s rationale would logically have required invalidation of the statute, as applied to that class of expression. For the statute would not then have "substantially serve[d] the Government’s legitimate ends," id. at 3272 , no matter how much unrelated, harmful conduct it prohibited.

. Justice Stevens, who added the decisive fifth vote to uphold the color and size provisions, argued even more emphatically that Time's publication of highly distorted images could assist counterfeiters. See Regan, 104 S.Ct. at 3294-96 (Stevens, J., concurring in the judgment in part and dissenting in part). Since his views limit the holding of the case, Regan clearly cannot be read to establish that a regulation that burdens protected expression more broadly than necessary is constitutional merely because some other part of the regulation is genuinely addressed to a legitimate objective.

. Regan suggests an important difference between the narrowly-tailored requirement of O’Brien , and less-restrictive-alternative analysis as applied to first amendment cases that do not involve time, place, and manner restrictions. Cf. Schad v. Mount Ephraim, 452 U.S. 61 , 69 n. 7, 101 S.Ct. 2176 , 2183 n. 7, 68 L.Ed.2d 671 (1981) (suggesting that lower demands apply to time, place, and manner restrictions). Ordinarily, a law that, at the margins, regulates some harmless protected expression may not constitutionally be enforced against that expression, even if this excessive scope is insignificant enough so that the law escapes facial invalidation as overbroad. See generally New York v. Ferber, 458 U.S. 747, 767-74 , 102 S.Ct. 3348, 3360-63 , 73 L.Ed.2d 1113 (1982). In the view of the Regan plurality, a time, place, and manner restriction that unavoidably and very marginally regulates harmless expression may be enforced against that expression.

. The challengers in Grace sought only to demonstrate on the sidewalk, and the Court there ióte had no occasion to rule on the constitutionality of the statute as it applied to other parts of the building and grounds. See 103 S.Ct. at 1706. However, the Court’s rationale centered on the public character of the sidewalk and the harmlessness of the prohibited activities when conducted there. See id. at 1708-10. Obviously, quite different considerations would apply to enforcement of the statute against, for example, demonstrations inside the building. Grace certainly did not foreclose enforcement of the statute in such a case, as it would have if courts were really empowered only "to uphold regulations [or statutes] which are constitutional and to strike down those which are not.” Maj.Op., text at n. 83. My point is that courts may — and traditionally do — find regulations valid as applied to some situations and invalid as applied to others, which is all the district court did in this case.

. The Assistant Director of Protective Research for the Secret Service testified that plywood could be used to scale the fence, and he stated that one person had used plywood to do so. *1548 Parr Tr., R. 162 at 53. However, later questioning by the court suggested that the single episode in question apparently involved a "structure,” see id. at 73, and not a simple plywood sheet. Cf. Defendant’s Exhibit MMM, Joint Appendix at 264 (Secret Service report on event referring to "structure"). The Assistant Director stated that about sixty people had gotten over the fence during the previous couple of years, and that no other person had used a sign or structure. Parr, R. 162 at 87. He agreed that the White House fence was constructed so that it was almost a ladder. Id. at 67.

The Assistant Director did testify that he might utilize a plywood sheet in attempting to scale the fence

"[i]f I was more athletic than I am right now, your honor. But I think I could use [the sheet]. Plywood is incredibly resilient, and it would probably bend before it would break, and all you would have to do is get one handhold on the top of that fence and be vaulted over, if you got a running jump, especially if you had a nice ramp to it____ Well, if you went back to Pennsylvania Avenue and ran toward this ramp effect here, it would probably bend in the middle, and you could get a handhold and be right over that fence.”

Parr Tr., R. 162 at 70-72. It was evidently this testimony that the judge disbelieved.

The plaintiffs introduced a plywood sheet in evidence, Plaintiffs’ Exhibit 133, which the judge was able to inspect. See R. 162 at 74 (recording receipt). The Assistant Director also told thé district judge that the Secret Service would know immediately if someone jumped the fence even if no officer was watching, Parr Tr., R. 162 at 83, that the Secret Service is perfectly able to respond if someone does get over the fence, id. at 84, and that a coricealed "counter-sniper” is always stationed so that he could shoot a fence-jumper if necessary, id. A government witness testified that plexiglass signs had caused problems in demonstrations on the sidewalk, see Lindsey Tr., R. 162 at 163-64, but no episodes involving plywood were described.

. The Park Service official responsible for park police operations in the District of Columbia expressed general concern about the use of rigid signs as weapons or shields, see Lindsey Tr., R. 162 at 163-64, Pegula Dep., R. 195 at 15-16. The single episode he described in which rigid signs were so used involved plexiglass signs. See Lindsey Tr., R. 162 at 163.

Appellee Picciotto testified that cardboard or cloth signs cannot stand up under the long hours of use her vigil on the sidewalk requires; particularly in bad weather, such signs are difficult to control and likely to be destroyed. See Picciotto Tr., R. 157A at 48-49; Picciotto Tr., R. 170 at 134-35, 161-62.

In essence, neither side disputes the evidence offered by the other; instead, dispute centers on whether the clear burden on speech is offset by the marginal increase in security on the sidewalk. In my view, it is not. This regulation differs from, for example, the sign support regulation in that (1) law enforcement officers testified to the actual use of thick wooden dowels as weapons, see Lindsey Tr., R. 162 at 165-66, but did not do so as to plywood signs; (2) a demonstrator wielding a two-by-four is a good deal more dangerous than one with a thin plywood sheet; and (3) banners of the maximum size allowed can be adequately supported with the permitted wooden supports, see infra note 11, but demonstrators for whom long-term presence on the sidewalk is a part of their message must use plywood signs or face the constant destruction of signs made from less permanent materials.

. See Lindsey Tr., R. 162 at 166 (recounting episode in which demonstrators “started dropping flares out of the hollow tubing, plus other materials that were used against [the] officers. They lit the flares and threw them at us, along with crimped nails, marbles, and a few other items.”). Appellees argue that these items could easily be concealed on the person of demonstrators, but common sense suggests that hollow tubing cduld enable demonstrators to carry a greater number of dangerous objects more easily-

. Plaintiffs’ expert witness on flags testified that a cloth banner of the maximum permissible size supported only by two persons might sag in the middle, and that demonstrators might break the supports in an effort to hold the banner aloft. See Christianson Tr., R. 179 at 12-14, 20-22. However, the same expert also agreed that a banner of that size, if supported in the middle as well as at the ends, could be effectively displayed. See id. at 8, 24.

. See Parr Tr., R. 162 at 61. The government's witness characterized the three-foot limit as "arbitrary,” id. at 60, but he followed that comment by observing that "[w]e could have four feet, we could have one foot.” Any linedrawing in this area necessarily has an arbitrary element.

. The district court invalidated the three-foot restriction at least in part based on the possibility of this less restrictive alternative. See White House Vigil for the ERA Comm. v. Clark, No. 83-1243, slip op. at 21, 24 (D.D.C. Apr. 26, 1984). However, security officials testified to their preference for regulations that do not require frequent interference with demonstrations, see Lindsey Tr., R. 162 at 167, and a demonstrator within three feet of the ledge who did not at one moment obscure any area under the ledge might quickly and entirely innocently change position or location so as to create a visual obstacle. Cf. Parr Tr., R. 162 at 61.

. See White House Vigil for the ERA Comm. v. Clark, No. 83-1243, slip op. at 24 (D.D.C. April 26, 1984) (categorically stating that “[t]he requirement that a demonstrator maintain contact with a sign is oppressive, and has little or nothing to do with security of any other government interest”).

. The majority says that even signs three feet high can be held "at different levels.” Maj.Op. at n. 118. This is true, of course, but it scarcely establishes that such signs so held will create a visual obstruction anywhere near as formidable as that which previously might have been possible. Actually, there is no evidence in the record to back up the majority's concern that large numbers of demonstrators frequently block the view in this way. The government does not, I believe, have as much leeway in regulating for the "worst case" scenario based on aesthetic concerns as it is granted when security is at stake.

. The majority observes that a total ban on sign-carrying demonstrators in the center zone would make it difficult for protesters to get from one side of the zone to the other. See Maj.Op. at n. 118. This may be so, but it is largely irrelevant. The question is whether the regulation as presently written significantly advances the purpose claimed for it.

. Bafflingly, the majority concludes that Justice Marshall would "apparently apply the first amendment to some restrictions of facilitative activity that have no effect on expression itself.” Maj.Op. at n. 144. I can find nothing in his opinion to support such an inference.

The Clark Court commented that the main value of sleeping in the park “would be facilitative.” 104 S.Ct. at 3070. It did so, however, in the course of explaining that the kind of demonstration sleep would facilitate — prolonged demonstrations that might damage the park — was exactly the kind of demonstration the Park Service had a legitimate interest in regulating. Despite the discussion in the dissent and in the opinions in the court of appeals, the Supreme Court conspicuously declined to make any broad statement about the application of the first amendment to regulations of facilitative conduct.

. The majority is uncertain “how substantial a burden on expression is necessary before the first amendment is implicated.” Maj.Op., text at n. 144 (emphasis in original) (footnote omitted). In my view, the parcels regulation imposes a burden well beyond any reasonable threshold, and I therefore need not decide precisely where the threshold is. Certainly a regulation of facilitative conduct need not absolutely bar expression in order to trigger first amendment concerns. In evaluating time, place, and manner restrictions, courts are traditionally called upon to weigh the substantiality of the burden a regulation imposes on speech. I would adapt the analysis developed in those cases in determining when the burden a facilitative regulation imposes on speech is substantial.