THE PRINCIPLE AND POLITICS OF EQUAL VALUE

THE PRINCIPLE AND POLITICS OF EQUAL VALUE

An unfamiliar equality principle is gaining prominence in consti­tutional discourse. Equal value presumptively prohibits government from regulating protected activities while exempting other activities to which the government’s interest applies just as readily. Although the principle is being developed in the context of free exercise, it has implica­tions for other guarantees in constitutional law. This Article offers two arguments. First, a version of equal value holds real attraction, not only within religious freedom law but also in areas such as freedom of expres­sion, reproductive rights, and equal protection. Second, however, the rule is operating in a patterned manner, favoring traditional religions at a moment when their social status is facing contestation and extending to decisions concerning free exercise and free speech but not non­establishment, due process, or equal protection. That implementation promotes a problematic political program. If the account here is correct, then equal value promises not an antidote to excessive judicial deference, as some have hoped, but instead a controversial politics.

The full text of this Article can be found by clicking the PDF link to the left.

Introduction

An unfamiliar equality rule has become prominent in constitutional discourse. Promoted by scholars and judges in the context of the free exercise of religion, it has implications for other provisions that guarantee evenhandedness toward conduct.

Here is the essence of equal value. Suppose a government regulates protected activities while exempting other activities. 1 Although equal value only pertains to “protected” persons, there is an ambiguity about whether they are shielded only because of their activities or also by virtue of their identities. Interestingly, the central examples all seem to involve protection against discrim­ination with respect to conduct, rather than against discrimination on the basis of a vulner­able status as such. See infra text accompanying notes 70–76 (discrimination against Muslim practices), 213–230 (discrimination affecting press publication), 330–341 (discrimination affecting the right to terminate a pregnancy). For the full argument that equal value probably pertains to conduct, and not status independently, see infra note 253 and accompanying text. If its interest applies evenly to the regulated and unregulated categories, then it presumptively has devalued protected practices—it has treated them as less worthwhile than the exempted activities. The government can only carry its burden by showing that its regulation of the protected conduct is necessary to further a compelling interest.

Cases arising in the context of the coronavirus pandemic have pro­vided the most recent and vivid illustrations. During the crisis, religious organizations were subject to limitations on gatherings. If other entities were exempted from those same limitations, and if those “essential” oper­ations presented equivalent health risks, then government had devalued the religious reasons for gathering, on this theory.

In Tandon v. Newsom, the Supreme Court relied on equal value. 2 141 S. Ct. 1294, 1296 (2021) (per curiam). California had limited private gatherings to three households and it had required social distancing, limited duration, and mask wearing. 3 Tandon v. Newsom, 992 F.3d 916, 917–18 (9th Cir. 2021). Those restrictions were challenged by a pastor and a congregant who wished to hold Bible study and worship services at private residences with members of more than three households. 4 Id. at 919. After both lower courts turned away requests for interim relief, the Supreme Court issued an emergency injunction. 5 Tandon, 141 S. Ct. at 1296. It held that regulations are presumptively unconstitutional “whenever they treat any comparable secular activity more favorably than religious exercise,” and it explained that comparability “must be judged against the asserted government interest that justifies the regulation at issue.” 6 Id. Because California had permitted hair salons, retail stores, movie theaters, and restaurants to include more than three households at a time, it faced a presumption of invalidity that it could not rebut. 7 Id. at 1297. In reasoning this way, the Tandon Court applied what some are calling the “most favored nation” approach to religious discrimination. 8 Arguably the Court applied the approach in other coronavirus decisions as well, though less explicitly. See infra section I.D (discussing those other decisions); see also Roman Cath. Diocese of Brooklyn v. Cuomo, 141 S. Ct. 63, 73 (2020) (Kavanaugh, J., concurring); Calvary Chapel Dayton Valley v. Sisolak, 140 S. Ct. 2603, 2612 (2020) (Kavanaugh, J., dissenting from the denial of application for injunctive relief) (embracing what scholars have called “‘something analogous to most-favored nation status’” for religious organizations (quoting Douglas Laycock, The Remnants of Free Exercise, 1990 S. Ct. Rev. 1, 49–50)).
Moreover, the judiciary is not alone—two states recently adopted versions of the rule through legislation. Act of March 29, 2021, ch. 192, sec. 1, § 12(a)(2), 2021 N.D. Sess. Laws (codified as amended at N.D. Cent. Code § 23-01-05(12)(d)(2) (2021)) (providing in relevant part that disease control orders may not “treat religious conduct more restrictively than any secular conduct of reasonably comparable risk, unless the government demonstrates through clear and convincing scientific evidence that a particular religious activity poses an extraordinary health risk”); An Act to Provide Protections for the Exercise of Religious Freedom, ch. 3, sec. 1, 2021 S.D. Sess. Laws (codified at S.D. Codified Laws § 1-1A-4 (2021)) (“[N]o state agency . . . may . . . [t]reat religious conduct more restrictively than any secular conduct of reasonably comparable risk . . . .”).

Several characteristics distinguish the new equality. First, it does not require any showing of discriminatory purpose, object, or intent. California’s coronavirus regulations, for instance, were neutral in their purpose and still presumptively invalid as to private religious gatherings. 9 Tandon, 992 F.3d at 922 (noting that the religious challengers “have not asserted that the object of the gatherings restrictions is to restrict religious practices, and there is no indication that the restrictions were adopted for discriminatory purposes instead of addressing public health concerns”). Second, it does not require a facial classification. Mere inclusion of reli­gious actors in a regulation, while comparable nonreligious actors are not regulated, may be sufficient to trigger the presumption of invalidity. 10 See, e.g., Roman Cath. Diocese, 141 S. Ct. at 73 (Kavanaugh, J., concurring) (“[O]nce a State creates a favored class of businesses, as New York has done in this case, the State must justify why houses of worship are excluded from that favored class.”). California’s regulation applied to all private gatherings, regardless of religiosity, and it was enforced evenhandedly, for example. 11 Tandon, 992 F.3d at 922 (noting that “the gatherings restrictions apply equally to private religious and private secular gatherings” and that “[t]here is no indication that the State is applying the restrictions to in-home private religious gatherings any differently than to in-home private secular gatherings”); see also Gateway City Church v. Newsom, 141 S. Ct. 1460, 1460 (2021) (mem.) (granting injunctive relief without mentioning the absence of a religious classification). Although an earlier decision did involve a facial distinction, that turned out to be inessential. Roman Cath. Diocese, 141 S. Ct. at 80 (Sotomayor, J., dissenting) (“New York treats houses of worship far more favorably than their secular comparators.”). Therefore, the approach differs from the constitutional rule for racial equality, which works only against classifications. 12 Suspect classifications may be facial or purposive—but in the absence of either type, a law is presumptively constitutional. See Washington v. Davis, 426 U.S. 229, 242 (1976) (holding that equal protection does not protect against policies with no discriminatory object or facial classification); cf. Brief of Respondent Flores at 10, City of Boerne v. Flores, 521 U.S. 507 (1997) (No. 95-2074), 1997 WL 10293 (explaining, with Douglas Laycock as Counsel of Record, that “[a] law may discriminate against religion without a finding of bad motive even if the law does not mention religion, if, for example, the law provides exemp­tions for secular hardship and no exemptions for religious exercise”). Of course, a suspect facial classification triggers a presumption of invalidity under equal protection even absent a showing of invidious motive. See, e.g., Pers. Adm’r v. Feeney, 442 U.S. 256, 272 (1979) (“A racial classification, regardless of purported motivation, is presumptively invalid and can be upheld only upon an extraordinary justification.”). So the mere fact that a showing of impermissible purpose is not required does not distinguish equal value from the main equal protection rule. That an equal value claim does not require any classification at all, whether facial or purposive, does distinguish the rule. For more on equal value’s distinctiveness as compared to equal protection, see infra sections II.A, II.C.

Finally, equal value may differ from a guarantee against disparate impact, though the difference is debatable. On the one hand, equal value appears to be less protective insofar as it only applies in situations where at least one category is exempted. So religious actors claiming a violation of equal value cannot prevail against a uniform limitation, even if they are disproportionately burdened in some sense. 13 Notice, however, that proponents may apply equal value wherever a regulation does not extend to comparable activities, even if the failure to regulate is a limitation in scope rather than an “exemption” as such. The articulated rule in Tandon allows this interpretation. See Tandon, 992 F.3d at 922–23. On the other hand, equal value seems more protective than disparate impact doctrine in certain ways. Chiefly, religious actors may succeed regardless of whether they are disproportionately affected. In the coronavirus context, for instance, con­gregations might be closed in precise proportion to nonreligious organi­zations, yet they would still win a presumption of invalidity if some of the latter were allowed to open in the face of similar safety risks. For example, some lower courts ruled in favor of religious schools that were shuttered during the pandemic, even though secular schools likewise were closed, on the ground that other organizations remained open. 14 See Monclova Christian Acad. v. Toledo-Lucas Cnty. Health Dep’t, 984 F.3d 477, 482 (6th Cir. 2020) (ruling for religious schools); Danville Christian Acad., Inc. v. Beshear, 503 F. Supp. 3d 516, 531 (E.D. Ky. 2020) (ruling for religious schools), stayed pending appeal sub nom. Kentucky v. Beshear, 981 F.3d 505, 511 (6th Cir. 2020), stay denied sub nom. Danville Christian Acad., Inc. v. Beshear, 141 S. Ct. 527, 528 (2020). That is not straightforward disparate impact analysis. 15 It is possible to measure disparate impact in more than one way. If schools were disproportionately religious compared to retailers, or compared to organizations even more generally, then closing all schools could have an outsized impact on religious actors. But the key point here is that even if schools were proportionately religious, so that there was no disparate impact, equal value could still apply. It is worth noting, however, that the two approaches might overlap significantly and that little turns on whether equal value is completely distinct from disparate impact protection.

Regardless of its distinctiveness, should equal value be embraced? It does capture an intuition that the government can wrongly burden pro­tected actors through disregard or devaluing. An ideal constitutional sys­tem might well guarantee against that kind of disdain of protected conduct, at least presumptively. And unelected judges plausibly have the capacity and competence to administer such a rule. Elected representa­tives may not be structurally incentivized to safeguard powerless groups, and they may fall into carelessness when operating under the pressure of exigency. Courts therefore might be necessary. 16 That is not to say that legislatures cannot be attentive to religious minorities, or that courts are necessarily more protective. See infra note 268 and accompanying text.

Consequently, some prominent liberal theorists are attracted to some­thing like the principle of equal value. 17 See, e.g., Christopher L. Eisgruber & Lawrence G. Sager, Religious Freedom and the Constitution 126 (2007) [hereinafter Eisgruber & Sager, Religious Freedom] (defend­ing a theory of “equal regard” with important similarities to equal value); Cass R. Sunstein, Our Anti-Korematsu (Harvard Pub. L. Working Paper No. 21-21, 2020), https://ssrn.com/‌‌abstract=3756853 [https://perma.cc/VH3X-S62W] [hereinafter Sunstein, Our Anti-Korematsu]. For them, equality has been too weakly protected in too many constitutional contexts. The new approach also could help to correct the Supreme Court’s moments of excessive def­erence to elected officials during emergencies. 18 Sunstein, Our Anti-Korematsu, supra note 17, at 11 (“The claim here is that insofar as the Roman Catholic Diocese Court was willing to vindicate antidiscrimination principles under exceedingly unusual circumstances posing severe risks, and to do so with genuinely strict scrutiny, it reflected an approach that is directly antithetical to that in Korematsu.”). Korematsu stands as the classic example of a failure to check the government in the face of racial discrimination, and—in a disturbing juxtaposition—the travel ban deci­sion provides the most consequential recent illustration of deference to the government despite undoubted religious hostility. 19 See infra section V.A (citing the cases). The juxtaposition is disturbing because the Court chose the travel ban case to formally repudiate Korematsu.

Equal value holds another attraction: It helps to safeguard liberty of conscience, which is underprotected by current constitutional law. In Employment Division v. Smith, the Court announced that the Free Exercise Clause would no longer protect against laws that were “neutral” and “generally applicable,” however much they might burden religious observance incidentally. 20 494 U.S. 872, 878–79 (1990). Today, that is the main rule for free exercise. Like many on the right and some on the left, I believe that this Smith rule is regrettable because it fails to shield those with minority beliefs on matters of conscience who are vulnerable to being overlooked by powerful political actors. 21 Nelson Tebbe, Free Exercise and the Problem of Symmetry, 56 Hastings L.J. 699, 709–10 (2005) [hereinafter Tebbe, Free Exercise] (endorsing the argument that Smith underprotects religious minorities). A companion piece evaluates a replacement. 22 Nelson Tebbe, The Principle and Politics of Liberty of Conscience, 135 Harv. L. Rev. 267 (2021) [hereinafter Tebbe, Liberty of Conscience]. Here, I only note that a requirement of equal value would provide ersatz liberty protection, if only in situations where a comparable actor happened to fall outside the regulation at issue. 23 Cf. Fulton v. City of Philadelphia, 141 S. Ct. 1868, 1931 (2021) (Gorsuch, J., concurring in the judgment) (suggesting that equal value addresses the “symptoms” but not the “underlying ailment”). For a contention that equal value insufficiently protects religious freedom because it only engages laws that happen to exempt some unprotected actor, see the first argument presented infra section III.C.

So if there is a problem with equal value, it is not simply that free exercise doctrine is being overruled implicitly. 24 Insincerity is a legitimate concern. See James M. Oleske, Jr., Free Exercise (Dis)Honesty, 2019 Wis. L. Rev. 689, 693–94 [hereinafter Oleske, Free Exercise (Dis)Honesty] (arguing that manipulating the general applicability requirement to effec­tively overrule Smith is insincere); cf. Michael C. Dorf, Under-Reacting to SCOTUS Theocracy, Dorf on L. (Dec. 2, 2020), http://www.dorfonlaw.org/2020/12/under-reacting-to-scotus-theocracy.html [https://perma.cc/9EXD-RNAL] (“What’s really going on is that the Roman Catholic Diocese majority are disregarding the Smith rule while pretending to follow it . . . . If a court can use the narrow tailoring inquiry itself to ascertain whether a law discriminates against religion, then the court has effectively overruled Smith.”). Weakening Smith is not necessarily unwelcome, and it can be done in the service of a principle that is defensible. Nor is the difficulty just that the new rule itself gives too little deference to elected officials who may be acting to save lives under emer­gency conditions such as a pandemic or a threat to national security. Judges can also err by deferring too much. 25 See Sunstein, Our Anti-Korematsu, supra note 17, at 11 (embracing Roman Catholic Diocese as a corrective to the Court’s excessive deference in Korematsu); see also infra section V.A.

What does deserve close examination is the possibility that equal value is being administered in the service of a problematic political program. Administrators of the rule must make two key choices: whether to apply the new equality and how to apply it. Both require the exercise of judgment, but the latter entails a particularly contestable baseline deter­mination. Are the actors who are exempt from the regulation comparable to the religious actors who are bringing the claim, with respect to the gov­ernment’s interest? During the pandemic, for instance, state officials have maintained that businesses like grocery stores and gymnasiums are less dangerous to public health than congregations or schools because they are not designed as gathering places. 26 See, e.g., Roman Cath. Diocese of Brooklyn v. Cuomo, 141 S. Ct. 63, 79 (2020) (Sotomayor, J., dissenting) (noting “the conditions medical experts tell us facilitate the spread of COVID-19” and arguing that, “[u]nlike religious services . . . , bike repair shops and liquor stores generally do not feature customers gathering inside to sing and speak together for an hour or more at a time” (citations omitted)). How are judges evaluating those determinations?

To gain perspective on that question, it is helpful to step back and consider whether equal value is being deployed in a patterned way across cases and contexts. When that is done, a troubling hypothesis emerges, namely that the new equality is being applied, and comparators are being found, in cases concerning religious groups disproportionately—and not all religious groups in any consistent way. In the travel ban decision, 27 Trump v. Hawaii, 138 S. Ct. 2392 (2018). for example, the Court failed to consider equal value. And that was telling because that situation was strikingly similar to the coronavirus context. The Court confronted a religious freedom challenge to an executive action concerning a threat to public safety in the travel ban case. It could have asked why the regulation did not apply evenly—why, for instance, the government exempted certain individuals in banned countries. 28 For a detailed analysis, see infra section V.A. Yet the Court did not raise the question of whether religious travelers were being devalued. 29 The argument was put before the Court by a prominent religious freedom organi­zation, if only in the abstract and not applied to the facts. Brief Amicus Curiae of the Becket Fund for Religious Liberty in Support of Neither Party at 22–23, Trump v. Hawaii, 138 S. Ct. 377 (2018) (No. 16-1540), 2017 WL 3588206 [hereinafter Becket Fund]. That omission cannot be explained by deference to executive expertise in a situation of exigency, for judicial humility might reasonably be thought to pertain to executive efforts to manage a global health crisis.

Nor has the requirement of equal value been applied in the Court’s decisions concerning racial equality under the Equal Protection Clause. 30 See infra section V.C. And that is unlikely to change. Given its past decisions, the Supreme Court cannot be expected to temper, for example, its rejection of disparate impact liability under the Constitution by asking whether the important activities of racial groups have been evenly regulated.

Even within religious freedom law, the approach is asymmetric. 31 See infra section IV.A. Equal value may have  no  exact  conceptual  analogue  in  the  Establishment  Clause,  as  explained below, 32 See infra section IV.A; see also Calvary Chapel Dayton Valley v. Sisolak, 140 S. Ct. 2603, 2612 (2020) (Kavanaugh, J., dissenting from the denial of application for injunctive relief) (“[T]he Court’s precedents make clear that the legislature may place religious organizations in the favored or exempt category rather than in the disfavored or non-exempt category without causing an Establishment Clause problem.”). but to the degree that it does suggest a non­establishment approach, none is forthcoming—no Justice is proposing a constitutional presumption against laws that place religious groups in exempted categories while restricting comparable actors. 33 Justice Kavanaugh made the asymmetry explicit. First, he discussed laws “that sup­ply no criteria for government benefits or action, but rather divvy up organizations into a favored or exempt category and a disfavored or non-exempt category,” and he concluded that governments may place religious organizations in the favored category without raising Establishment Clause concerns. Calvary Chapel, 140 S. Ct. at 2611–12. In the next paragraph, he acknowledged that the free exercise rule is stronger:
The converse free-exercise or equal-treatment question is whether the legislature is required to place religious organizations in the favored or exempt category rather than in the disfavored or non-exempt category. The Court’s free-exercise and equal-treatment precedents also supply an answer to that question: Unless the State provides a sufficient justification otherwise, it must place religious organizations in the favored or exempt category.
Id. at 2612.
On the other hand, equal value will be welcomed into freedom of expression law. Some­thing like it has historically been applied in press cases and it can be expected to further influence speech rules on content discrimination. 34 See infra section IV.B. Lopsidedness like that does little to dispel the impression of partiality. Moreover, the disparity is a social and political circumstance, not a legal or moral inevitability. 35 Moreover, the Court’s use of the “shadow docket” to foreground equal value con­tributes to the impression that its members are motivated to reach particular results. See Stephen I. Vladeck, Opinion, The Supreme Court Is Making New Law in the Shadows, N.Y. Times (Apr. 15, 2021), https://www.nytimes.com/2021/04/15/opinion/supreme-court-religion-orders.html (on file with the Columbia Law Review) (noting that Tandon “came on the court’s ‘shadow docket,’ and in a context in which the Supreme Court’s own rules supposedly limit relief to cases in which the law is ‘indisputably clear’”).

This Article concludes that although equal value holds real attraction as a matter of ideal theory, its implementation under nonideal conditions is another matter. Now is the time to consider the concept carefully. Equal value must be understood and evaluated, not only within religious free­dom but across equality doctrines, and not only in principle but also in its practical implementation. Doing that suggests a disquieting possibility—that it is being deployed to support a program of religious preferentialism and laissez-faire constitutionalism that extends beyond free exercise and nonestablishment to free speech and equal protection. Religious groups, including the largest denominations, are winning cases, and private speak­ers are being protected against public regulation, while sexual and racial communities are left undefended  by  constitutional  law  against  a  naturalized  stratification  of  social power. 36 See infra section IV.D (on reproductive freedom for women), sections IV.C, V.C (on racial justice). In fact, the existing distribution itself is constitutionalized and thus insulated from democratization efforts. Equality law is being strained across doctrines to rationalize these results.

Part I accounts for the origins of the new equality. Its judicial visibility can be  traced  to  an  opinion  by  Justice  Samuel  Alito,  written  while  he  was  sitting  on  the  Third Circuit. 37 See infra section I.B (discussing Fraternal Ord. of Police Newark Lodge No. 12 v. City of Newark, 170 F.3d 359, 366 (3d Cir. 1999)). Among scholars, versions of equal value were proposed earlier in the 1990s, both by egalitarians 38 See, e.g., Eisgruber & Sager, Religious Freedom, supra note 17, at 90–91 (embracing the Fraternal Order ruling as invalidating a failure of “equal regard”). and by others. 39 Compare Douglas Laycock & Steven T. Collis, Generally Applicable Law and the Free Exercise of Religion, 95 Neb. L. Rev. 1, 21–22 (2016) (embracing Fraternal Order), with Douglas Laycock, The Broader Implications of Masterpiece Cakeshop, 2019 BYU L. Rev. 167, 182–83 (arguing that the Colorado baker was devalued because the state had crafted “an implicit secular exception” from its public accommodations law). Their arguments were taken up in the coronavirus cases and in dicta in Fulton v. City of Philadelphia. 40 141 S. Ct. 1868, 1877 (2021) (“A law also lacks general applicability if it prohibits religious conduct while permitting secular conduct that undermines the government’s asserted interests in a similar way.”).

Part II defends the view that the new equality differs from other forms of antidiscrimination found in existing constitutional law. Not only is it not reducible to the main free exercise rule, but it also diverges from the lead­ing alternative. Part II also distinguishes disparate impact protection for nonwhite people, albeit with caveats. It ends by identifying two sorts of cases that could be seen as precursors, namely certain decisions in the fundamental interest branch of equal protection law and a line of speech opinions concerning press freedoms. 41 See infra sections II.D, II.E.

Part III explores two justifications for the rule of equal value. First, the model polices a form of unfairness. Not every situation where government regulates protected actors while exempting comparable others is unfair, but the heuristic does capture many laws that subordinate members of the political community or frustrate the exercise of basic liberties. Second, equal value provides vicarious political protection for powerless groups. You could think of this second point as an argument from positive political theory or constitutional political economy. Part III concludes by considering critiques of the principle, some of which are persuasive.

Part IV suggests ways in which the approach could improve antidis­crimination doctrines in other areas of constitutional law. These include the rule against favoring religious activities under the Establishment Clause, the prohibition of content discrimination under the Speech Clause, the protection for conduct closely associated with racial identity, and the due process right to reproductive freedom. 42 For reasons of readability and manageability, this Article otherwise puts to one side discrimination on the basis of sex, gender, and sexual orientation. It also brackets the question of whether equal value applies to funding programs.

Part V turns to equal value’s actual administration, which has a per­ceptible shape. First, it has been deployed in some constitutional domains and not others. Cases concerning free exercise and freedom of speech are growth areas for the new equality, whereas those concerning nonestablish­ment and racial justice are not. Second, where equal value is applied, it requires comparators to be identified. The way that has been done in the Tandon line of cases contributes to the impression that the rule has been administered selectively. Inferring from such outcomes, the Part hypothe­sizes that the Roberts Court is pursuing a combination of preferentialism in religion cases and constitutionalization of existing power distributions in free speech and equal protection. And it is doing so at a moment in history when traditional religious actors are facing status contestation.

Part V closes by drawing out implications for judicial review. 43 See infra section V.E. Alt­hough the Court’s willingness to enforce constitutional rights during the coronavirus crisis has been received as a welcome corrective to its past def­erence to executive action, 44 See supra note 18. we might learn something additional by com­paring the coronavirus cases to the Court’s decisions during the early decades of the twentieth century. On one understanding of that period, the Court engaged in baseline manipulation in order to naturalize the existing distribution of wealth and entitlements and to impose a particular economic program on democratic politics. Today, too, the determinative factor may not be substantive law, nor the institutional design of courts, but rather the development and deployment of a politics that should then be evaluated on its own terms.

A brief conclusion acknowledges that similar pressures could affect any ideal approach to free exercise, if not in the same way. 45 See Tebbe, Liberty of Conscience, supra note 22, at 296–319; cf. Brief Amicus Curiae of Professor Eugene Volokh in Support of Neither Party at 1, Fulton v. City of Philadelphia, 141 S. Ct. 1868 (2021) (No. 19-123), 2020 WL 3078491 [hereinafter Volokh, Brief Amicus Curiae] (“[O]verruling Smith would revive all the flaws of a broad substantive due process regime.”). Still, it is imperative to identify the particular interaction of political power and legal discourse here and now, so that we can understand what equal value is likely to entail under existing historical conditions.