Justice BERCH,
opinion of the Court.
¶ 1 We granted review to determine whether A.R.S. § 15-1441(1), which adds two at-large members to the governing board of any community college district located in a county with a population of at least three million people, is a special law prohibited by article 4, part 2, section 19 of the Arizona Constitution. We hold that the amended section is not an unconstitutional special law.
I. BACKGROUND
¶ 2 Community college districts are authorized by statute. A.R.S. §§ 15-1401 to - 1410, -1441 to -1453. Ten such districts exist in Arizona, each comprising five precincts.
Id.
§ 15-1441 (A). Before the amendment at issue here, each district was governed by a local community college district board consisting of five members, one elected from each precinct for a six-year term.
Id.
§ 15-1441(C).
¶ 3 In April 2010, the Arizona Legislature amended A.R.S. § 15-1441(1) to require the election of two at-large members to community college boards in very populous counties:
[I]n addition to the governing board members who are elected from each of the five precincts in a community college district, a
county with a population of at least three million persons shall elect two additional governing members from the district at large.
2010 Ariz. Sess. Laws, eh. 48, § 1 (2d Reg. Sess.). The amendment also reduced board member terms from six years to four years, for each “county with a population of at least three million persons.”
Id.
(amending AR.S. § 15-1441(C), (I)).
¶4 Because only Maricopa County has a population of more than three million, it is the only county to which the amendment now applies. Arizona’s next most populous counties, Pima and Pinal, will likely not have three million people before the end of the century. Maricopa County is set to elect its first two additional at-large governing board members at the November 2014 general election.
¶ 5 Plaintiffs, six registered voters from Maricopa County, filed a complaint in superi- or court seeking a declaration that the legislation is unconstitutional and an injunction to prevent the election of the at-large board members. Plaintiffs contended, among other claims, that, as amended, A.R.S. § 15-1441(1) violates Arizona’s constitutional prohibition against special laws.
See
Ariz. Const, art. 4, pt. 2, § 19.
¶ 6 The superior court concluded that § 15-1441(1) does not violate the special law prohibition. The court of appeals reversed, holding that this statute is an unconstitutional special law.
Gallardo v. State of Arizona (“Gallardo
7”), 236 Ariz. 1, 335 P.3d 523 (2014). It found that the opportunity to enter the class, and thus receive the benefit of the law, must “be not only possible, but reasonably probable,” and must be able to occur “within a reasonable time.”
Id.
at 5 ¶ 12, 335 P.3d 523 (citations and internal quotation marks omitted). Because no other county will qualify to enter the class for more than eighty years, the court concluded that the class was not practically or reasonably elastic or expandable.
Id.
at 5, 6 ¶¶ 13, 16-17, 335 P.3d 523.
¶ 7 We granted review to clarify the test for determining when a law is an unconstitutional special law. The issue has statewide importance, as at least thirty-five Arizona statutes rely on population classifications.
II. DISCUSSION
A. Standard of Review
¶ 8 At the outset, Defendants argue that in determining whether a statute is a special law, we must apply a strong presumption in favor of its constitutionality, and Plaintiffs must prove its unconstitutionality beyond a reasonable doubt. Although prior cases have used similar language, it incorrectly states the standard. Determining constitutionality is a question of law, which we review de novo.
League of Ariz. Cities & Towns v. Brewer,
213 Ariz. 557, 559 ¶ 7, 146 P.3d 58, 60 (2006). Assessing the constitutionality of a law fundamentally differs from determining the existence of historical facts, the determination of which is subject to deference.
State v. Gonzalez-Gutierrez,
187 Ariz. 116, 118, 927 P.2d 776, 778 (1996). We therefore disapprove the use of the “beyond a reasonable doubt” standard for making constitutionality determinations.
¶ 9 We do, however, presume that “the legislature acts constitutionally.”
Ariz. Minority Coal. for Fair Redistricting v. Ariz. Indep. Redistricting Comm’n,
220 Ariz. 587, 595 ¶ 21, 208 P.3d 676, 684 (2009) (quoting
State v. Murphy,
117 Ariz. 57, 61, 570 P.2d 1070,1074 (1977)). But if a law burdens fundamental rights, such as free speech or freedom of religion, any presumption in its favor falls away.
See id.
at 595 ¶ 20 n. 7, 208 P.3d at 684 n. 7 (observing that “content-based restrictions on speech are ‘presumptively invalid,’ ” so “the burden shifts to the government to demonstrate that a legislative enactment is constitutional”) (quoting
R.A.V. v. City of St. Paul,
505 U.S. 377, 382, 112 S.Ct. 2538, 120 L.Ed.2d 305 (1992));
see also Cave Creek Unified Sch. Dist. v. Ducey,
233 Ariz. 1, 5 ¶ 11, 308 P.3d 1152, 1156 (2013) (observing that “[w]hen the statute in question involves no fundamental constitutional rights or distinctions based on suspect classifications, we presume the statute is constitutional and will uphold it unless it clearly is not”). In this case, the law in question touches only peripherally on the right to
vote.
See Ariz. Minority Coal.,
220 Ariz. at 595 ¶20 n. 7, 208 P.3d at 684 n. 7 (finding that “redistricting alone” does not affect the “essence of the fundamental right” to vote (citations and internal quotation marks omitted)). Thus, we will accord our traditional presumption of constitutionality to the legislative enactment.
Id.
B. Special Laws
¶ 10 Special laws favor one person or group and disfavor others.
Republic Inv. Fund I v. Town of Surprise,
166 Ariz. 143, 148-49, 800 P.2d 1251, 1256-57 (1990). Article 4, part 2, section 19(11) of the Arizona Constitution prohibits special laws that affect “[t]he conduct of elections.” The Constitution also prohibits special laws if “a general law can be made applicable.” Ariz. Const, art. 4, pt. 2, § 19(20).
¶ 11 This Court set forth the test for determining when a law is a special law nearly twenty-five years ago in
Republic Investment.
To survive scrutiny, (1) the law must have “a rational relationship to a legitimate legislative objective,” (2) the classification the law makes must be legitimate, encompassing all members that are similarly situated, and (3) the classification must be elastic, allowing “other individuals or entities to come within” and move out of the class.
Republic Inv.,
166 Ariz. at 149, 800 P.2d at 1257 (quoting
Petitioners for Deannexation v. City of Goodyear,
160 Ariz. 467, 472, 773 P.2d 1026, 1031 (App.1989)). We suggested that the analysis proceed in the following order:
[T]he court must first ascertain whether the law has a rational relationship to a legitimate legislative objective. If it does not, of course, our inquiry is over. But if it does, we must further decide if the act legitimately classifies by population, geography, or time limitations. If we find a legitimate classification, we must then determine if the act permits other individuals or entities to come within the class, and thus within operation of the law, within a reasonable time, or if at all.
Id.
(quoting
Petitioners for Deannexation,
160 Ariz. at 472, 773 P.2d at 1031).
1.
Rational Basis
¶ 12 The first prong of the special laws test requires that the law bear a rational relationship to a legitimate legislative objective.
Id.
Such a relationship is identical to that required for equal protection analysis.
See State Comp. Fund v. Symington,
174 Ariz. 188, 193, 848 P.2d 273, 278 (1993);
Republic Inv.,
166 Ariz. at 149, 800 P.2d at 1257;
Ariz. Downs v. Ariz. Horsemen’s Found.,
130 Ariz. 550, 557-58, 637 P.2d 1053, 1060-61 (1981). A law “may withstand equal protection review, yet still be found unconstitutional under the special/local law provision” if it fails the additional assessments required under the inelusiveness and elasticity prongs of the special laws analysis, not because a more rigorous or exacting type of rational basis analysis applies.
Republic Inv.,
166 Ariz. at 149, 800 P.2d at 1257.
¶ 13 Seizing on language in
Republic Investment
stating that, “in addition to whether the
classification
has a reasonable basis,” a challenged law must also be inclusive and elastic,
id.
(emphasis added), Plaintiffs argue that the classification itself, and not the law as a whole, must have a rational basis. We disagree. Consistent with our rational basis analysis of equal protection and equal privileges and immunities issues, we conclude that it is the law itself, and not merely the classification, that must have a rational basis.
Id.; Petitioners for Deannexation,
160 Ariz. at 472, 773 P.2d at 1031. We consider the reasonableness of the classification when analyzing the inclusiveness prong of the
Republic Investment
test.
¶ 14 The legislature did not state its purpose in amending AR.S. § 15-1441(1). Defendants posit that in very populous counties, community college districts are apt to be large and so may require a larger board to facilitate governance, ameliorate under-representation, and promote diversity. Plaintiffs counter that county population and boundaries bear little relationship to community college district size, noting that the state has only ten community college districts,
even though Arizona has fifteen counties.
¶ 15 The legislature could have drawn community college district lines more precisely based on a district’s student population or number of campuses. But a law’s imprecision does not preclude a rational basis finding.
Big D Constr. Corp. v. Court of Appeals,
163 Ariz. 560, 566, 789 P.2d 1061, 1068 (1990) (noting that “[a] perfect fit is not required”).
¶ 16 The legislature apparently used county population as a proxy for the district’s student population or number of campuses. Using population as a shorthand for other, more specific measures is not irrational.
Cf. Long v. Napolitano,
203 Ariz. 247, 255-56 ¶ 25, 53 P.3d 172, 180-81 (App.2002) (approving use of county population to identify counties having a large airport and a sufficient number of hotels to support tourism related to a stadium). Moreover, the structure of the community college districts shows the relationship between the districts and counties.
See
AR.S. §§ 15-1401 to -1410, -1441 to - 1453. Each district is named for the county or counties that it encompasses. And representatives on community college district boards are generally elected from districts that mirror the county supervisorial district boundaries.
See
AR.S. § 15-1402 (setting requirements for community college districts);
id.
§ 15-1441(A) (setting up precincts in districts to have the same boundaries as election precincts).
¶ 17 Citing
Long,
203 Ariz. at 254 ¶20, 53 P.3d at 179, Plaintiffs urge that every identified purpose for a law must have a rational basis.
Long
differs from this case, however, in that the statute there dealt with four distinct subjects: a football stadium, tourism, spring training, and youth sports.
Id.
at 254 ¶ 18, 53 P.3d at 179. Here, we deal with only one subject, for which many rationales have been advanced.
In such a situation, a court need not find that every rationale is furthered. As long as the law conceivably furthers a legitimate governmental purpose, a rational basis exists.
Ariz. Downs,
130 Ariz. at 555, 637 P.2d at 1058.
¶ 18 Reasons such as facilitating community college district governance and increasing representation and perhaps diversity on governing boards rationally relate to legitimate governmental interests “common ... to the
whole
state.”
Petitioners for Deannexation,
160 Ariz. at 472, 773 P.2d at 1031 (quoting
Abrams v. State,
534 P.2d 91, 94 (Alaska 1975)). In sum, the legislature could have rationally concluded that adding two at-large seats to the governing boards of very populous districts would promote better governance and representation.
2.
Inclusiveness
¶ 19 The
Republic Investment
test also requires that the classification be legitimate and that it encompass all members of the relevant class. 166 Ariz. at 149, 800 P.2d at 1257. A law may be general even if it applies only to one entity as long as that entity is the only member of a legitimate class.
Id.
at 150 & n. 4, 800 P.2d at 1258 & n. 4.
¶ 20 At the trial court, Defendants presented statistics showing that based on the county’s population, each Maricopa County Community College District (“MCCCD”) Board member serves more than 760,000 citizens, nearly four times the number in the next most populous county, Pima County. The citizen-to-board-member ratio for all other districts is much lower.
¶21 The parties do not contest that the MCCCD is one of the largest community college districts in the nation, having more than 250,000 students attending classes through ten campuses. Given the district’s size, the legislature could rationally have concluded that the district is so large that its board should have additional members to better serve the students and public, and that only when the number of citizens represented per board member reaches a certain threshold do the problems of under-repre
sentation, insufficient capacity, and need for increased governance arise.
¶ 22 Plaintiffs attempt to show that the law does not cover all affected members of the class by asserting that the ratio of representation between Maricopa County and Pima County is 3.9 to 1, while the disparity between Pima County and other counties is much larger; this, they maintain, demonstrates that the law is underinclusive. For example, they allege that Pima County Community College District Board members represent approximately 196,000 people, or 7.5 times the number represented by board members in Cochise County. This disparity (7.5 to 1), being larger than the disparity between Maricopa and Pima counties (3.9 to 1), they claim, indicates that the Pima County Community College District Board also needs additional members, even though Pima County does not have more than three million persons.
¶ 23 But the problem to be solved by this statute is not simply to equalize across community college boards the ratio between the district’s population and the number of board members. It is entirely possible that a five-member community college district board can adequately govern in a county like Pinal, with a population of approximately 375,000 and a ratio of approximately 75,000 citizens per board member, while a district like the Pima County Community College District, with a population just under a million persons and a ratio of approximately 196,000 persons per board member, can
still
be adequately governed by a five-member board. Stated differently, Plaintiffs presume that the optimum number of board members is linearly related to district population, but the legislature could reasonably have concluded that it is not.
See
Jordan Ellenberg,
How Not to Be Wrong: The Power of Mathematical Thinking
21-30 (2014).
¶24 Plaintiffs have not shown that Pima County suffers from under-representation or insufficient capacity, or that it needs increased governance. Nor have they established that any other district faces the problems confronting the MCCCD. The legislature has implicitly determined that only when a higher threshold is reached do these problems manifest to such a degree that they must be remedied. We defer to the legislature’s assessment that there is a problem to be solved and its policy choice as to how to resolve it.
See Ariz. Minority Coal.,
220 Ariz. at 595 ¶¶ 20-21, 208 P.3d at 684 (noting that “when there is a reasonable, even though debatable, basis for the enactment of a statute, we will uphold the act unless it is clearly unconstitutional” (quoting
Murphy,
117 Ariz. at 61, 570 P.2d at 1074)).
¶ 25 This lack of evidence distinguishes this case from other cases finding underinclu-siveness, in which evidence or concessions by the legislature established that other entities not included in the classification suffered from the problem the legislation endeavored to solve and so should have been included in the class.
See, e.g., State v. Levy’s,
119 Ariz. 191, 192-93, 580 P.2d 329, 330-31 (1978) (relying on “agreed facts” showing that all business suffered from the problem that the tax exemption offered only to border businesses was intended to solve);
Town of Gilbert v.
Maricopa
Cnty.,
213 Ariz. 241, 246 ¶¶ 18-19, 141 P.3d 416, 421 (App.2006) (concluding that county islands not included in the class also lacked the fire and emergency services that the legislation provided to a select group of county islands);
In re Marxus B.,
199 Ariz. 11, 13-14 ¶¶ 9-12, 13 P.3d 290, 292-93 (App. 2000) (relying on legislative findings that the problem was statewide, while the statute applied only to two counties);
In re Cesar R.,
197 Ariz. 437, 439 ¶ 6, 4 P.3d 980, 982 (App. 1999) (same). The case before us more closely resembles
Long,
in which the court concluded that counties with fewer than two million persons were not similarly situated to more populous counties for the purposes of remedying the problem at hand.
Long,
203 Ariz. at 256 ¶¶ 28-29, 53 P.3d at 181.
¶ 26 Plaintiffs also argue that the class here is underinclusive because community college district governing boards in counties with fewer than three million people might also benefit by having additional board members. But this alone does not render a classification underinclusive.
See id.
at 256 ¶ 27, 53 P.3d at 181 (“The legislature is not constrained from enacting class-based legislation
merely because non-members of the class would also derive some benefit from the legislation.”). The legislature need not tailor the law with absolute precision.
Plaintiffs have not shown that the population line that defines the classification here is unreasonable or creates an underinclusive class. Thus, we conclude that the class is legitimate and reasonably includes, at present, the lone member of the affected class.
3.
Elasticity
¶ 27 The third prong of the test, elasticity, is the sole ground on which the court of appeals struck down the challenged law. We agree with that court that the classification here — any county with a population of at least three million people — is facially elastic.
Gallardo I,
335 P.3d 523, 236 Ariz. at 5, ¶ 13, 2014 WL 3671571, at *3. It appears “open” in that it permits “entry of additional persons, places, or things attaining the requisite characteristics,” and it allows “others to exit the statute’s coverage when they no longer have those characteristics.”
Republic Inv.,
166 Ariz. at 150, 800 P.2d at 1258. Counties may achieve the population threshold over time and, if Maricopa County loses population, it would fall out of the classification.
¶ 28 We disagree, however, with the court of appeals’ conclusion that the elasticity prong requires that it be reasonably probable that potential class members will enter the class within a particular time. The court derived this temporal requirement from eases that used language suggesting such a precondition.
See Gallardo I,
236 Ariz. at 5 ¶ 12, 335 P.3d 523.
Republic Investment,
for example, stated that the law must “permit[ ] other individuals or entities to come within the class ... within a reasonable time.” 166 Ariz. at 149, 800 P.2d at 1257 (quoting
Petitioners far Deannexation,
160 Ariz. at 472, 773 P.2d at 1031). But in both
Republic Investment
and
Petitioners for Deannexation,
the class under consideration was already closed when the legislature passed the act in question.
Republic Inv.,
166 Ariz. at 147, 151, 800 P.2d at 1255, 1259 (limiting class to “territory from a city or town having a population of less than ten thousand persons according to the 1980 United States decennial census” (emphasis omitted));
Petitioners for Deannexation,
160 Ariz. at 469, 773 P.2d at 1028 (analyzing same law). The cases that
Petitioners for Deannexation
cited as the origin of this “reasonable time” phrase similarly offer no support for implying a temporal requirement.
See
160 Ariz. at 472-73, 773 P.2d at 1031-32.
Barbee v. Holbrook,
91 Ariz. 263, 371 P.2d 886 (1962), involved a statute that restricted the class to Santa Cruz County; thus the class was closed by definition.
Arizona Downs
and
Fund Manager v. Corbin
did not discuss a temporal limitation and focused only on whether the classifications at issue could have a “broader application in the future.”
Ariz. Downs,
130 Ariz. at 558, 637 P.2d at 1061;
Fund Manager, Pub. Safety Pers. Ret. Sys. v. Corbin,
161 Ariz. 348, 359-60, 778 P.2d 1244, 1255-56 (App.1988).
¶29 The court of appeals’ assertion that “the conditions that permit entry into the class must ‘be not only possible, but reasonably probable, of attainment,’ ”
Gallardo I,
335 P.3d 523, 691 Ariz. Adv. Rep. at 39 ¶ 12, 2014 WL 3671571, at *3 (quoting
Republic Inv.,
166 Ariz. at 150, 800 P.2d at 1258), also comes from
Petitioners for Deannexation,
which in turn took the phrase from
Bravin v. Mayor of Tombstone,
4 Ariz. 83, 90, 33 P. 589, 590 (1893). Like the classes in
Republic Investment
and
Petitioners for Deannexation,
though, the class in
Bravin
was closed
when the law was passed, as it was restricted to cities in which fewer than six hundred people voted in an election already held.
Id.
at 88, 33 P. at 590 (limiting class to “all cities ... in which the total vote cast at the general election held therein on the fourth day of November, 1890, was less than six hundred”). Thus future entry into the class was not only unlikely, it was impossible: Cities either had the requisite population when the law was passed or they did not. Thus any suggestion that entry of others into the class must be “reasonably probable” of attainment was dictum.
¶ 30
Republic Investment’s
statements that “[t]o decide whether a statute legitimately classifies, [courts must] consider the
actual probability
that others will come under the act’s operation when the population changes,” and if “the prospect is
only theoretical, and not probable,
we will find the act special or local in nature,” 166 Ariz. at 151, 800 P.2d at 1259, can be also traced to
Bra-vin,
4 Ariz. at 90, 33 P. at 590, which cited Sutherland’s Treatise on Statutory Construction. 2 Sutherland Statutory Construction § 40:9 (“Some opinions suggest that courts look to the
actual probability
of [population increases] and will hold an act invalid where such a prospect is
only theoretical and not practical.”
(emphasis added)). But because the class in
Bravin,
like the class in
Republic Investment,
was closed, any suggestion that this Court adopted an “actual probability” requirement was unnecessary to the holding, and therefore dictum. This requirement should focus on the probability that a potential class member will “come under the act’s operation when the population changes,” not on the likelihood that the population will change.
Republic Inv.,
166 Ariz. at 151, 800 P.2d at 1259. Here, petitioners have not questioned that when any county reaches the population threshold, it will come within the class.
¶31 The question in cases like
Republic Investment, Petitioners for Deannexation,
and
Bravin,
was whether each class would remain open long enough to allow a realistic opportunity for others to join the class. In each case, because the eligibility window had already closed, no such opportunity existed. It was against this background that the “probability of entry” statements were made. These courts were not creating a rule that future entry must occur within a particular time. Nonetheless, we understand how courts could have relied on such statements to imply an additional requirement in order to meet elasticity.
¶ 32 Since
Republic Investment,
the court of appeals has cited that decision’s dictum to find classifications inelastic based on the unlikelihood that entities would enter a class within a reasonable time.
See Town of Gilbert,
213 Ariz. at 247 ¶¶ 22-23, 141 P.3d at 422 (finding that the remote possibility that any county island might join the class in nineteen years does not satisfy the elasticity requirement);
In re Cesar R.,
197 Ariz. at 440 ¶¶ 10-11, 4 P.3d at 983 (finding that ten years does not satisfy the elasticity requirement);
In re Marxus B.,
199 Ariz. at 14 ¶ 14, 13 P.3d at 293 (interpreting the same statute as in
In re Cesar R.
and approving of its finding on elasticity). The findings regarding elasticity, however, were unnecessary in each case because the panel had already found that the statute at issue failed the inelusiveness prong of the
Republic Investment
test.
Town of Gilbert,
213 Ariz. at 246 ¶¶ 17-19, 141 P.3d at 421;
In re Cesar R.,
197 Ariz. at 440 ¶ 11, 4 P.3d at 983;
In re Marxus B.,
199 Ariz. at 14 ¶ 11, 13 P.3d at 293. In such a situation, the court need not have ruled on elasticity at all.
See Republic Inv.,
166 Ariz. at 149, 800 P.2d at 1257 (suggesting that failure to meet one part dooms the statute under special laws analysis).
¶ 33 As the court of appeals’ concurring opinion in this case noted, courts have been unable to incorporate a probability or temporal element into the elasticity requirement in any discernibly consistent way.
See Gallardo I,
236 Ariz. at 7 ¶ 22, 335 P.3d 523 (Howe, J., specially concurring) (comparing
Town of Gilbert,
213 Ariz. at 247 ¶¶ 22-23, 141 P.3d at 422 (finding nineteen years too long to satisfy elasticity requirement),
and In re Cesar R.,
197 Ariz. at 440 ¶¶ 9-11, 4 P.3d at 983 (finding ten years too long),
with Long,
203 Ariz. at 258 ¶ 38, 53 P.3d at 183 (upholding law with a population threshold
of two million without considering how long it might take any county to achieve that population)). We agree and therefore now disavow any suggestion that entry into the class within a particular time is a requirement of the elasticity prong of the special-laws analysis. Such a temporal component would preclude using population classifications to protect uniquely large or small jurisdictions unless the legislature expands the class to cover or allow immediate entry of other members that may not need the law’s protections.
¶34 While the imposition of a temporal reference may have arisen to preclude pre-textual classifications that mask an underlying discriminatory intent — a valid concern— adding a temporal requirement does not allay that concern. For if, instead of setting a county population threshold at three million, the legislature set the cutoff at 1.5 million, such a limit would likely satisfy the reasonable time analysis, even though the legislature could in the future raise the threshold. The concern about the legitimacy of the classification is better addressed in the analysis of the rational basis and inelusiveness prongs of the
Republic Investment
test than in the elasticity prong.
¶ 35 We reaffirm our holding in
Republic Investment
that the elasticity requirement is met when the statute looks to broader application in the future, no matter how imminent the application might be, and allows “persons, places, or things attaining the requisite characteristics” to enter and those that “no longer have those characteristics” to leave the class. 166 Ariz. at 150, 800 P.2d at 1258;
see also Luhrs v. City of Phoenix,
52 Ariz. 438, 451, 83 P.2d 283, 289 (1938) (elasticity met by statute that “contemplated that cities in the future that attain the stipulated requisite conditions automatically come within the terms of the law”);
Long,
203 Ariz. at 258 ¶ 38, 53 P.3d at 183 (“[A]ny county may seemingly enter the class upon achieving the requisite population and may exit upon falling below that level.”).
¶36 We conclude that the class here is facially and functionally elastic. Any county that attains the stated size will join the class, even though none likely will do so in the near future, and any county whose population falls below three million will leave the class. Rather than focusing on the imminence or likelihood of accomplishment, courts should focus on whether, once a county attains the population threshold, it will share the characteristics of the classification and fall within the class.
III. CONCLUSION
¶ 37 Having found the three parts of the
Republic Investment
test satisfied here, we hold that A.R.S. § 15-1441(1) does not violate the special laws provision of article 4, part 2, section 19 of the Arizona Constitution. We vacate the opinion of the court of appeals and affirm the judgment of the superior court.