announced the judgment of the Court and delivered an opinion in which The Chief Justice and Mr. Justice Powell joined.

We are asked once again to police the constitutional boundary between church and state. Maryland, this time, is the alleged trespasser. It has enacted a statute which, as amended, provides for annual noncategorical grants to private colleges, among them religiously affiliated institutions, subject only to the restrictions that the funds not be used for “sectarian purposes.” A three-judge District Court, by a divided vote, refused to enjoin the operation of the statute, 387 F. Supp. 1282 (Md. 1974), and a direct appeal has been taken to this Court pursuant to 28 U. S. C. § 1253.

*740I

The challenged grant program was instituted by Laws of 1971, c. 626, and is now embodied in Md. Ann. Code, Art. 77A, §§ 65-69 (1975). It provides funding for “any private institution of higher learning within the State of Maryland,” provided the institution is accredited by the State Department of Education, was established in Maryland prior to July 1, 1970, maintains one or more “associate of arts or baccalaureate degree” programs, and refrains from awarding “only seminarian or theological degrees.” §§ 65-66.1 The aid is in the form of an annual fiscal year subsidy to qualifying colleges and universities. The formula by which each institution’s entitlement is computed has been changed several times and is not independently at issue here. It now provides for a qualifying institution to receive, for each full-time student (excluding students enrolled in seminarian or theological academic programs), an amount equal to 15% of the State’s per-full-time-pupil appropriation for a student in the state college system. § 67. As first enacted, the grants were completely unrestricted. They remain noncategorical in nature, and a recipient institution may put them to whatever use it prefers, with but one exception. In 1972, following this Court’s decisions in Lemon v. Kurtzman, 403 U. S. 602 (1971) (Lemon I), and Tilton v. Richardson, 403 U. S. 672 (1971), § 68A was added to the statute by Laws of 1972, c. 534. It provides:

“None of the moneys payable under this subtitle *741shall be utilized by the institutions for sectarian purposes.”

The administration of the grant program is entrusted to the State’s Board of Public Works “assisted by the Maryland Council for Higher Education.” These bodies are to adopt “criteria and procedures ... for the implementation and administration of the aid program.” They are specifically authorized to adopt “criteria and procedures” governing the method of application for grants and of their disbursement, the verification of degrees conferred, and the “submission of reports or data concerning the utilization of these moneys by [the aided] institutions.” § 68.2 Primary responsibility for the program rests with the Council for Higher Education, an appointed commission which antedates the aid program, which has numerous other responsibilities in the educational field, and which has derived from these a “considerable expertise as to the character and functions of the various private colleges and universities in the State.” 387 F. Supp., at 1285.

The Council performs what the District Court described as a “two-step screening process” to insure compliance with the statutory restrictions on the grants. First, it determines whether an institution applying for aid is eligible at all, or is one “awarding primarily theo*742logical or seminary degrees.” 3 Several applicants have been disqualified at this stage of the process. Id., at 1289, 1296. Second, the Council requires that those institutions that are eligible for funds not put them to any sectarian use. An application must be accompanied by an affidavit of the institution's chief executive officer stating that the funds will not be used for sectarian purposes, and by a description of the specific nonsectarian uses that are planned.4 These may be changed only after written notice to the Council. By the end of the fiscal year the institution must file a “Utilization of Funds Report” describing and itemizing the use of the funds. The chief executive officer must certify the report and also file his own “Post-expenditure Affidavit,” stating that the funds have not been put to sectarian uses. The recipient institution is further required to segregate state funds in a “special revenue account” and to identify aided nonsectarian expenditures separately in its budget. It must retain “sufficient documentation of the State funds expended to permit verification by the Council that funds were not spent for sectarian purposes." Any question of sectarian *743use that may arise is to be resolved by the Council, if possible, on the basis of information submitted to it by the institution and without actual examination of its books. Failing that, a “verification or audit” may be undertaken. The District Court found that the audit would be “quick and non-judgmental,” taking one day or less. Id., at 1296.5

In 1971, $1.7 million was disbursed to 17 private institutions in Maryland. The disbursements were under the statute as originally enacted, and were therefore not subject to § 68A’s specific prohibition on sectarian use. Of the 17 institutions, five were church related; and these received $520,000 of the $1.7 million. A total of $1.8 million was to be awarded to 18 institutions in 1972, the second year of the grant program; of this amount, $603,000 was to go to church-related institutions. Before disbursement, however, this suit, challenging the grants as in violation of the Establishment Clause of the First Amendment, was filed.6 The $603,-000 was placed in escrow and was so held until after the entry of the District Court’s judgment on October 21, 1974.7 These and subsequent awards, therefore, are *744subject to § 68A and to the Council's procedures for insuring compliance therewith.

Plaintiffs in this suit, appellants here, are four individual Maryland citizens and taxpayers.8 Their complaint sought a declaration of the statute's invalidity, an order enjoining payments under it to church-affiliated institutions, and a declaration that the State was entitled to recover from such institutions any amounts already disbursed. App. 10. In addition to the responsible state officials,9 plaintiff-appellants joined as defendants the five institutions they claimed were constitutionally ineligible for this form of aid: Western Maryland College, College of Notre Dame, Mount Saint Mary’s College, Saint Joseph College, and Loyola College. Of these, the last four are affiliated with the Roman Catholic Church; Western Maryland, was a Methodist affiliate. The District Court ruled with respect to all five. Western Maryland, however, has since been dismissed as a defendant-appellee. We are concerned, therefore, only with the four Roman Catholic affiliates.10

After carefully assessing the role that the Catholic Church plays in the fives of these institutions, a matter to which we return in greater detail below, and applying *745the three-part requirement of Lemon I, 403 U. S., at 612-613, that state aid such as this have a secular purpose, a primary effect other than the advancement of religion, and no tendency to entangle the State excessively in church affairs, the District Court ruled that the amended statute was constitutional and was not to be enjoined. The court considered the original, unamended statute to have been unconstitutional under Lemon I, but it refused to order a refund of amounts theretofore paid out, reasoning that any refund was barred by the decision in Lemon v. Kurtzman, 411 U. S. 192 (1973) (Lemon II).11 The District Court therefore denied all relief. This appeal followed. We noted probable jurisdiction. 420 U. S. 922 (1975).

II

A system of government that makes itself felt as pervasively as ours could hardly be expected never to cross paths with the church. In fact, our State and Federal Governments impose certain burdens upon, and impart certain benefits to, virtually all our activities, and religious activity is not an exception. The Court fyas enforced a scrupulous neutrality by the State, as *746among religions, and also as between religious and other activities,12 but a hermetic separation of the two is an impossibility it has never required. It long has been established, for example, that the State may send a cleric, indeed even a clerical order, to perform a wholly secular task. In Bradfield v. Roberts, 175 U. S. 291 (1899), the Court upheld the extension of public aid to a corporation which, although composed entirely of members of a Roman Catholic sisterhood acting “under the auspices of said church,’ id., at 297, was limited by its corporate charter to the secular purpose of operating a charitable hospital.

And religious institutions need not be quarantined from public benefits that are neutrally available to all. The Court has permitted the State to supply transportation for children to and from church-related as well as public schools. Everson v. Board of Education, 330 U. S. 1 (1947). It has done the same with respect to secular textbooks loaned by the State on equal terms to students attending both public and church-related elementary schools. Board of Education v. Allen, 392 U. S. 236 (1968). Since it had not been shown in Allen that the secular textbooks would be put to other than secular purposes, the Court concluded that, as in Everson, the State was merely “extending the benefits of state laws to all citizens.” Id., at 242. Just as Bradfield dispels any notion that a religious person can never be in the State’s pay for a secular purpose,13 *747Everson and Allen put to rest any argument that the State may never act in such a way that has the incidental effect of facilitating religious activity. The Cburt has not been blind to the fact that in aiding a religious institution to perform a secular task, the State frees the institution’s resources to be put to sectarian ends.14 If this were impermissible, however, a church could not be protected by the police and fire departments, or have its public sidewalk kept in repair. The Court never has held that religious activities must be discriminated against in this way.

Neutrality is what is required. The State must confine itself to secular objectives, and neither advance nor impede religious activity. Of course, that principle is more easily stated than applied. The Court has taken the view that a secular purpose and a facial neutrality may not be enough, if in fact the State is lending direct support to a religious activity. The State may not, for example, pay for what is actually a religious education, even though it purports to be paying for a secular one, and even though it makes its aid available to secular and religious institutions alike. The Court also has taken the view that the State’s efforts to perform a secular task, and at the same time avoid aiding in the performance of a religious one, may.not lead it into such an intimate relationship with religious authority that it appears either tb be sponsoring or to be exces*748sively interfering with that authority.15 In Lemon I as noted above, the Court distilled these concerns into a three-prong test, resting in part on prior case law, for the constitutionality of statutes affording state aid to church-related schools:

“First, the statute must have a secular legislative purpose; second, its principal or primary effect must be one that neither advances nor inhibits religion . . . ; finally, the statute must not foster ‘an excessive government entanglement with religion.’ ” 403 U. S., at 612-613.

At issue in Lemon I were two state-aid plans, a Rhode Island program to grant a 15% supplement to the salaries of private, church-related school teachers teaching secular courses, and a Pennsylvania program to reimburse private church-related schools for the entire cost of secular courses also offered in public schools. Both failed the third part of the test, that of “excessive government entanglement.” This part the Court held in turn required a consideration of three factors: (1) the character and purposes of the benefited institutions, (2) the nature of the aid provided, and (3) the resulting relationship between the State and the religious authority. Id., at 615. As to the first of these, in reviewing the Rhode Island program, the Court found that the aided schools, elementary and secondary, were characterized by “substantial religious activity and purpose.” Id., at 616. They were located near parish churches. Religious instruction was considered “part of the total *749educational process.” Id., at 615. Religious symbols and religious activities abounded. Two-thirds of the teachers were nuns, and their operation of the schools was regarded as an “ 'integral part of the religious mission of the Catholic Church.' ” Id., at 616. The schooling came at an impressionable age. The form of aid also cut against the programs. Unlike the textbooks in Allen and the bus transportation in Everson, the services of the state-supported teachers could not be counted on to be purely secular. They were bound to mix religious teachings with secular ones, not by conscious design, perhaps, but because the mixture was inevitable when teachers (themselves usually Catholics) were “employed by a religious organization, subject to the direction and discipline of religious authorities, and work[ed] in a system dedicated to rearing children in a particular faith.” Id., at 618. The State's efforts to supervise and control the teaching of religion in supposedly secular classes would therefore inevitably entangle it excessively in religious affairs. The Pennsylvania program similarly foundered.

The Court also pointed to another kind of church-state entanglement threatened by the Rhode Island and Pennsylvania programs, namely, their “divisive political potential.” Id., at 622. They represented “successive and very likely permanent annual appropriations that benefit relatively few religious groups.” Id., at 623. Political factions, supporting and opposing the programs, were bound to divide along religious lines. This was “one of the principal evils against which the First Amendment was intended to protect.” Id., at 622. It was stressed that the political divisiveness of the programs was “aggravated ... by the need for continuing annual appropriations.” Id., at 623.16

*750In Tilton v. Richardson, 403 U. S. 672 (1971), a companion case to Lemon I, the Court reached the contrary-result. The aid challenged in Tilton was in the form of federal grants for the construction of academic facilities at private colleges, some of them church related, with the restriction that the facilities not be used for any sectarian purpose.17 Applying Lemon I’s three-part test, the Court found the purpose of the federal aid program there under consideration to be secular. Its primary effect was not the advancement of religion, for sectarian use of the facilities was prohibited. Enforcement of this prohibition was made possible by the fact that religion did not so permeate the defendant colleges that their religious and secular functions were inseparable. On the contrary, there was no evidence that religious activities took place in the funded facilities. Courses at the colleges were “taught according to the academic requirements intrinsic to the subject matter,” and “an atmosphere of academic freedom rather than religious indoctrination” was maintained. 403 U. S., at 680-682 (plurality opinion).

Turning to the problem of excessive entanglement, the Court first stressed the character of the aided institutions. It pointed to several general differences between college and precollege education: College students are less susceptible to religious indoctrination; college courses tend to entail an internal discipline that inherently limits the opportunities for sectarian influence; and a high degree of academic freedom tends to prevail at the college level. It found no evidence that the col*751leges in Tilton varied from this pattern. Though controlled and largely populated by Roman Catholics, the colleges were hot restricted to adherents of that faith. No religious services were required to be attended. Theology courses were mandatory, but they were taught in an academic fashion, and with treatment of beliefs other than Roman Catholicism. There wére no attempts to proselytize among students, and principles of academic freedom prevailed. With colleges of this character, there was little risk that religion would seep into the teaching of secular subjects, and the state surveillance necessary to separate the two, therefore, was diminished. The Court next looked to the type of aid provided, and found it to be neutral or nonideological in nature. Like the textbooks and bus transportation in Allen and Everson, but unlike the teachers’ services in Lemon I, physical facilities were capable of being restricted to secular purposes. Moreover, the construction grant was a one-shot affair, not involving annual audits and appropriations.

As for political divisiveness, no “continuing religious aggravation” over the program had been shown, and the Court reasoned that this might be because of the lack of continuity in the church-state relationship, the character and diversity of the colleges, and the fact that they served a dispersed student constituency rather than a local one. “ [C] umulatively,” all these considerations persuaded the Court that church-state entanglement was not excessive. 403 U. S., at 684-689.

In Hunt v. McNair, 413 U. S. 734 (1973), the challenged aid was also for the construction of secular college facilities, the state plan being one to finance the construction by revenue bonds issued through the medium of a state authority. In effect, the college serviced and repaid the bonds, but at the lower cost resulting from the tax-free status of the interest payments. The Court upheld the program on reasoning analogous to that in *752Tilton. In applying the second of the Lemon 7’s three-part test, that concerning “primary effect,” the following refinement was added:

“Aid normally may be thought to have a primary effect of advancing religion when it flows to an institution in which religion is so pervasive that a substantial portion of its functions are subsumed in the religious mission or when it funds a specifically religious activity in an otherwise substantially secular setting.” 413 U. S., at 743.

Although the college which Hunt concerned was subject to substantial control by its sponsoring Baptist Church, it was found to be similar to the colleges in Tilton and not “pervasively sectarian.” As in Tilton, state aid weht to secular facilities only, and thus not to any “specifically religious activity.” 413 U. S., at 743-745.

Committee for Public Education v. Nyquist, 413 U. S. 756 (1973), followed in Lemon Vs wake much as Hunt followed in Tilton’s. The aid in Nyquist was to elementary and secondary schools which, the District Court found, generally conformed to a “profile” of a sectarian or substantially religious school.18 The state aid took three forms: direct subsidies for the maintenance and repair of buildings; reimbursement of parents for a percentage of tuition paid; and certain tax benefits for parents. All three forms of aid were found to have an impermissible primary effect. The mainte*753nance and repair subsidies, being unrestricted, could be used for the upkeep of a chapel or classrooms used for religious instruction. The reimbursements and tax benefits to parents could likewise be used to support wholly religious activities.

In Levitt v. Committee for Public Education, 413 U. S. 472 (1973), the Court also invalidated a program for public aid to church-affiliated schools. The grants, which were to elementary and secondary schools in New York, were in the form of reimbursements for the schools’ testing and recordkeeping expenses. The schools met the same sectarian profile as did those in Nyquist, at least in some cases. There was therefore “substantial risk” that the state-funded tests would be “drafted with an eye, unconsciously or otherwise, to inculcate students in the religious precepts of the sponsoring church.” 413 U. S., at 480.

Last Term, in Meek v. Pittenger, 421 U. S. 349 (1975), the Court ruled yet again on a state-aid program for church-related elementary and secondary schools. On the authority of Allen, it upheld a Pennsylvania program for lending textbooks to private school students. It found, however, that Lemon I required the invalidation of two other forms of aid to the private schools. The first was the loan of instructional materials and equipment. Like the textbooks, these were secular and nonideological in nature. Unlike the textbooks, however, they were loaned directly to the schools. The schools, similar to those in Lemon I, were ones in which “the teaching process is, to a large extent, devoted to the inculcation of religious values and belief.” 421 U. S., at 366. Aid flowing directly to such “religion-pervasive institutions,” ibid., had the primary effect of advancing religion. See Hunt v. McNair, supra. The other form of aid was the provision of “auxiliary” educational serv*754ices: remedial instruction, counseling and testing, and speech and hearing therapy. These also were intended to be neutral and nonideological, and in fact were to be provided by public school teachers. Still, there was danger that the teachers, in such a sectarian setting, would allow religion to seep into their instruction. To attempt to prevent this from happening would excessively entangle the State in church affairs. The Court referred again to the danger of political divisiveness, heightened, as it had been in Lemon I and Nyquist, by the necessity of annual legislative reconsideration of the aid appropriation. 421 U. S., at 372.

So the slate we write on is anything but clean. Instead, there is little room for further refinement of the principles governing public aid to church-affiliated private schools. Our purpose is not to unsettle those principles, so recently reaffirmed, see Meek v. Pittenger, supra, or to expand upon them substantially, but merely to insure that they are faithfully applied in this case.

Ill

The first part of Lemon J’s three-part test is not in issue; appellants do not challenge the District Court’s finding that the purpose of Maryland’s aid program is the secular one of supporting private higher education generally, as an economic alternative to a wholly public system.19 The focus of the debate is on the second and third parts, those concerning the primary effect of ad*755vancing religion, and excessive church-state entanglement. We consider them in the same order.

A

While entanglement is essentially a procedural problem, the primary-effect question is the substantive one of what private educational activities, by whatever procedure, may be supported by state funds. Hunt requires (1) that no state aid at all go to institutions that are so “pervasively sectarian” that secular activities cannot be separated from sectarian ones, and (2) that if secular activities can be separated out, they alone may be funded.

(1) The District Court’s finding in this case was that the appellee colleges are not “pervasively sectarian.” 387 F. Supp., at 1293. This conclusion it supported with a number of subsidiary findings concerning the role of religion on these campuses:

(a) Despite their formal affiliation with the Roman Catholic Church, the colleges are “characterized by a high degree of institutional autonomy.” Id., at 1287 n. 7. None of the four receives funds from, or makes reports to, the Catholic Church. The Church is represented on their governing boards, but, as with Mount Saint Mary’s, “no instance of entry of Church considerations into college decisions was shown.” Id., at 1295.

(b) The colleges employ Roman Catholic chaplains and hold Roman Catholic religious exercises on campus. Attendance at such is not required; the encouragement of spiritual development is only “one secondary objective” of each college; and “at none of these institutions does this encouragement go beyond providing the opportunities or occasions for- religious experience.” Ibid. It was the District Court’s general finding that “religious indoctrination is not a substantial purpose or activity of any of these defendants.” Id., at 1296.

*756(c) Mandatory religion or theology courses are taught at each of the colleges, primarily by Roman Catholic clerics, but these only supplement a curriculum covering “the spectrum of a liberal arts program.” Nontheology courses are taught in an “atmosphere of intellectual freedom” and without “religious pressures.” 20 Each college subscribes to, and abides by, the 1940 Statement of Principles on Academic Freedom of the American Association of University Professors. Id., at 1288, 1293, and n. 3, 1295.

(d) Some classes are begun with prayer. The percentage of classes in which this is done varies with the college, from a “minuscule” percentage at Loyola and Mount Saint Mary’s, to a majority at Saint Joseph. Id., at 1293. There is no “actual college policy” of encouraging the practice. “It is treated as a facet of the instructor’s academic freedom.” Ibid. Classroom prayers were therefore regarded by the District Court as “peripheral to the subject of religious permeation,” as were the facts that some instructors wear clerical garb and some classrooms have religious symbols. Ibid. The court concluded:

“None of these facts impairs the clear and con*757vincing evidence that courses at each defendant are taught 'according to the academic requirements intrinsic to the subject matter and the individual teacher’s concept of professional standards.’ [citing Tilton v. Richardson, 403 U. S., at 681].” Id., at 1293-1294.

In support of this finding the court relied on the fact that a Maryland education department group had monitored the teacher education program at Saint Joseph College, where classroom prayer is most prevalent, and had seen “no evidence of religion entering into any elements of that program.” Id., at 1293.

(e) The District Court found that, apart from the theology departments, see n. 20, supra, faculty hiring decisions are not made on a religious basis. At two of the colleges, Notre Dame and Mount Saint Mary’s, no inquiry at all is made into an applicant’s religion. Religious preference is to be noted on Loyola’s application form, but the purpose is to allow full appreciation of the applicant’s background. Loyola also attempts to employ each year two members of a particular religious order which once staffed a college recently merged into Loyola. Budgetary considerations lead the colleges generally to favor members of religious orders, who often receive less than full salary. Still, the District Court found that “academic quality” was the principal hiring criterion, and that any “hiring bias,” or “effort by any defendant to stack its faculty with members of a particular religious group,” would have been noticed by other faculty members, who had never been heard to complain. Id., at 1294.

(f) The great majority of students at each of the colleges are Roman Catholic, but the District Court concluded from a “thorough analysis of the student ad*758mission and recruiting criteria” that the student bodies “are chosen without regard to religion.” Id., at 1295.

Wé cannot say that the foregoing findings as to the role of religion in particular aspects of the colleges are clearly erroneous. Appellants ask us to set those findings aside in certain respects. Not surprisingly, they have gleaned from this record of thousands of pages, compiled during several weeks of trial, occasional evidence of a more sectarian character than the District Court ascribes to the colleges. It is not our placo, however, to reappraise the evidence, unless it plainly fails to support the findings of the trier of facts. That is certainly not the case here, and it would make no difference even if we were to second-guess the District Court in certain particulars. To answer the question whether an institution is so “pervasively sectarian” that it may receive no direct state aid of any kind, it is necessary to paint a general picture of the institution, composed of many elements. The general picture that the District Court has painted of the appellee institutions is similar in almost all respects to that of the church-affiliated colleges considered in Tilton and Hunt.21 We *759find no constitutionally significant distinction between them, at least for purposes of the “pervasiye sectarianism” test.

(2) Having found that the appellee institutions are not "so permeated by religion that the secular side cannot be separated from the sectarian,” 387 F. Supp., at 1293, the District Court proceeded to the next question posed by Hunt: whether aid in fact was extended only to “the secular side.” This requirement the court regarded as satisfied by the statutory prohibition against sectarian use, and by the administrative enforcement of that prohibition through the Council for Higher Education. We agree. Hunt requires only that slate funds not be used to support "specifically religious activity.” It is clear that fund uses exist that meet this require*760ment. See Tilton v. Richardson, supra; Hunt v. McNair, supra. We have no occasion to elaborate further on what is and is not a “specifically religious activity/' for no particular use of the state funds is set out in this statute. Funds are put to the use of the college's choice, provided it is not a sectarian use, of which the college must satisfy the Council. If the question is whether the statute sought to be enjoined authorizes state funds for “specifically religious activity,” that question fairly answers itself. The statute in terms forbids the use of funds for “sectarian purposes,” and this prohibition appears to be at least as broad as Hunt’s prohibition of the public funding of “specifically religious activity.” We must assume that the colleges, and the Council, will exercise their delegated control over use of the funds in compliance with the statutory, and therefore the constitutional, mandate. It is to be expected that they will give a wide berth to “specifically religious activity,” and thus minimize constitutional questions.22 *761Should such questions arise, the courts will consider them. It has not been the Court’s practice, in considering facial challenges to statutes of this kind, to strike them down in anticipation that particular applications may result in unconstitutional use of funds. See, e. g., Hunt v. McNair, 413 U. S., at 744; Tilton v. Richardson, 403 U. S., at 682 (plurality opinion).

B

If the foregoing answer to the “primary effect” ques*762tion seems easy, it serves to make the “excessive entanglement” problem more difficult. The statute itself clearly denies the use of public funds for “sectarian purposes.” It seeks to avert such use, however, through a process of annual interchange — proposal and approval, expenditure and review — between the colleges and the Council. In answering the question whether this will be an “excessively entangling” relationship, we must consider the several relevant factors identified in prior decisions:

(1) First is the character of the aided institutions. This has been fully described above. As the District Court found, the colleges perform “essentially secular educational functions,” 387 F. Supp., at 1288, that are distinct and separable from religious activity. This finding, which is a prerequisite under the “pervasive sectarianism” test to any state aid at all, is also important for purposes of the entanglement test because it means that secular activities, for the most part, can be taken at face value. There is no danger, or at least only a substantially reduced danger, that an ostensibly secular activity — the study of biology, the learning of a foreign language, an athletic event — will actually be infused with religious content or significance. The need for close surveillance of purportedly secular activities is correspondingly reduced. Thus the District Court found that in this case “there is no necessity for state officials to investigate the conduct of particular classes of educational , programs to determine whether a school is attempting to indoctrinate its students under the guise of secular education.” Id., at 1289. We cannot say the District Court erred in this judgment or gave it undue significance. The Court took precisely the same view with respect to the aid extended to the very similar institutions in Tilton. 403 U. S., at 687 (plurality opinion), See also Hunt v. McNair, supra, at 746.

*763(2) As for the form of aid, we have already noted that no particular use of state funds is before us in this case. The process by which aid is disbursed, and a use for it chosen, is before us. We address this as a matter of the “resulting relationship” of secular and religious authority.

(3) As noted, the funding process is an annual one. The subsidies are paid out each year, and they can be put to annually varying uses. The colleges propose particular uses for the Council’s approval, and, following expenditure, they report to the Council on the use to which the funds have been put.

The District Court’s view was that in light of the character of the aided institutions, and the resulting absence of any need “to investigate the conduct of particular classes,” 387 F. Supp., at 1289, the annual nature of the subsidy was not fatal. In fact, an annual, ongoing relationship had existed in Tilton, where the Government retained the right to inspect subsidized buildings for sectarian use, and the ongoing church-state involvement had been even greater in Hunt, where the State was actually the lessor of the subsidized facilities, retaining extensive powers to regulate their use. See 387 F. Supp., at 1290.

We agree with the District Court that “excessive entanglement” does not necessarily result from the fact that the subsidy is an annual one. It is true that the Court favored the “one-time, single-purpose” construction grants in Tilton because they entailed “no continuing financial relationships or dependencies, no annual audits, and no government analysis of an institution’s expenditures,” 403 U. S., at 688 (plurality opinion). The present aid program cannot claim these aspects. But if the question is whether this case is more like Lemon I or more like Tilton — and surely that is the *764fundamental question before us — the answer must be that it is more like Tilton.

Tilton is distinguishable only by the form of aid. We cannot discount the distinction entirely, but neither can we regard it as decisive. As the District Court pointed out, ongoing, annual supervision of college facilities was explicitly foreseen in Tilton, 403 U. S., at 675; see also Lemon I, 403 U. S., at 669 (opinion of White, J.), and even more so in Hunt, 413 U. S., at 739-740, 745-749. Tilton and Hunt would be totally indistinguishable, at least in terms of annual supervision, if funds were used under the present statute to build or maintain physical facilities devoted to secular use. The present statute contemplates annual decisions by the Council as to what is a “sectarian purpose/’ but, as we have noted, the secular and sectarian activities of the colleges are easily separated. Occasional audits are possible here, but we must accept the District Court’s finding that they would be “quick and non-judgmental.” 387 P. Supp., at 1296. They and the other contacts between the Council and the colleges are not likely to be any more entangling than the inspections and audits incident to the normal process of the colleges’ accreditations by the State.

While the form-of-aid distinctions of Tilton are thus of questionable importance, the character-of-institution distinctions of Lemon I are most impressive., To reiterate a few of the relevant points: The elementary and secondary schooling in Lemon I came at an impressionable age; the aided schools were “under the general supervision” of the Roman Catholic diocese; each school had a local Catholic parish that assumed “ultimate financial responsibility” for it; the principals of the schools were usually appointed by church authorities; religion “pervade[d] the school system”; teachers were specifically instructed by the “Handbook of School Regula*765tions” that “ ' [r] eligious formation is not confined to formal courses; nor is it restricted to a single subject area.’ ” 403 U. S., at 617-618. These things made impossible what is crucial to a nonentangling aid program: the ability of the State to identify and subsidize separate secular functions carried out at the school, without on-the-site inspections being necessary to prevent diversion of the funds to sectarian purposes. The District Court gave primary importance to this consideration, and we cannot say it erred.

(4) As for political divisiveness, the District Court recognized that the annual nature of the subsidy, along with its promise of an increasing demand for state funds as the colleges’ dependency grew, aggravated the danger of “[pjolitical fragmentation ... on religious lines.” Lemon I, 403 U. S., at 623. Nonetheless, the District Court found that the program “does not create a substantial danger of political entanglement.” 387 F. Supp., at 1291. Several reasons were given. As was stated in Tilton, the danger of political divisiveness is “substantially less” when the aided institution is not an elementary or secondary school, but a college, “whose student constituency is not local but diverse and widely dispersed.” 403 U. S., at 688-689. Furthermore, political divisiveness is diminished by the fact that the aid is extended to private colleges generally, more than two-thirds of which have no religious affiliation; this is in sharp contrast to Nyquist, for example, where 95% of the aided schools were Roman Catholic parochial schools. Finally, the substantial autonomy of the colleges was thought to mitigate political divisiveness, in that controversies surrounding the aid program are not likely to involve the Catholic Church itself, or even the religious character of the schools, but only their “fiscal responsi*766bility and educational requirements.” 387 F. Supp., at 1290-1291.

The District Court’s reasoning seems to us entirely sound. Once again, appellants urge that this case is controlled by previous cases in which the form of aid was similar (Lemon I, Nyquist, Levitt), rather than those in which the character of the aided institution was the same (Tilton, Hunt). We disagree. Though indisputably relevant, see Lemon I, 403 U. S., at 623-624, the annual nature of the aid cannot be dispositive. On the one hand, the Court has struck down a “permanent,” nonannua! tax exemption, reasoning that “the pressure for frequent enlargement of the relief is predictable,” as it always is. Committee for Public Education v. Nyquist, 413 U. S., at 797. On the other hand, in Tilton it has upheld a program for “one-time, single-purpose” construction grants, despite the fact that such grants would, in fact, be “annual,” at least insofar as new grants would be annually applied for. 403 U. S., at 688. See Lemon I, 403 U. S., at 669 (opinion of White, J.). Our holdings are better reconciled in terms of the character of the aided institutions, found to be so dissimilar as between those considered in Tilton and Hunt, on the-one hand, and those considered in Lemon I, Nyquist, and Levitt, on the other.

There is no exact science in gauging the entanglement of church and state. The wording of the test, which speaks of “excessive entanglement,” itself makes that clear. The relevant factors we have identified are to be considered “cumulatively” in judging the degree of entanglement. Tilton v. Richardson, 403 U. S., at 688. They may cut different ways, as certainly they do here. In reaching the conclusion that it did, the District Court gave dominant importance to the character of the aided institutions and to its finding that they are capable of separating secular and religious functions. For the rea*767sons stated above, we cannot say that the emphasis was misplaced or the finding erroneous.23

The judgment of the District Court is affirmed,

It is so ordered.

with whom Mr. Justice Rehnquist joins, concurring in the judgment.

While I join in the judgment of the Court, I am unable to concur in the plurality opinion substantially for the reasons set forth in my opinions in Lemon v. *768Kurtzman, 403 U. S. 602 (1971) (Lemon I), and Committee for Public Education v. Nyquist, 413 U. S. 756 (1973). I am no more reconciled now to Lemon I than I was when it was decided. See Nyquist, supra, at 820 (White, J., dissenting). The threefold test of Lemon I imposes unnecessary, and, as I believe today’s plurality opinion demonstrates, superfluous tests for establishing “when the State’s involvement with religion passes the peril point” for First Amendment purposes. Id., at 822.

“It is enough for me that the [State is] financing a separable secular function of overriding importance in order to sustain the legislation here challenged.” Lemon I, supra, at 664 (opinion of White, J.). As long as there is a secular legislative purpose, and as long as the primary effect of the legislation is neither to advance nor inhibit religion, I see no reason — particularly in light of the “sparse language of the Establishment Clause,” Committee for Public Education v. Nyquist, supra, at 820—to take the constitutional inquiry further. See Lemon I, supra, at 661 (opinion of White, J.); Nyquist, supra, at 813 (White, J., dissenting). However, since 1970, the Court has added a third element to the inquiry: whether there is “an excessive government entanglement with religion.” Walz v. Tax Comm’n, 397 U. S. 664, 674 (1970). I have never understood the constitutional foundation for this added element; it is at once both insolubly paradoxical, see Lemon I, supra, at *769668, and — as the Court has conceded from the outset — a “blurred, indistinct, and variable barrier.” Lemon I, supra, at 614. It is not clear that the “weight and contours of entanglement as a separate constitutional criterion,” Nyquist, supra, at 822, are any more settled now than when they first surfaced. Today's plurality opinion leaves the impression that the criterion really may not be “separate” at all. In affirming the District Court’s conclusion that the legislation here does not create an “excessive entanglement” of church and state, the plurality emphasizes with approval that “the District Court gave dominant importance to the character of the aided institutions and to its finding that they are capable of separating secular and religious functions.” Ante, at 766. Yet these are the same factors upon which the plurality focuses in concluding that the Maryland legislation satisfies the second part of the Lemon I test: that on the record the “appellee colleges are not ‘pervasively sectarian,’ ” ante, at 755, and that the aid at issue was capable of, and is in fact, extended only to “ ‘the secular side’ ” of the appellee colleges’ operations. Ante, at 759. It is unclear to me how the first and third parts of the Lemon I test are substantially different.* The “excessive entanglement” test appears no less “curious and mystifying” than when it was first announced. Lemon I, supra, at 666.

I see no reason to indulge in the redundant exercise of evaluating the same facts and findings under a different label. No one in this case challenges the District *770Court’s finding that the purpose of the legislation here is secular. Ante, at 754. And I do not disagree with the plurality that the primary effect of the aid program is not advancement of religion. That is enough in my view to sustain the aid programs against constitutional challenge, and I would say no more.

with whom Mr. Justice Marshall joins, dissenting.

I agree with Judge Bryan, dissenting from the judgment under review, that the Maryland Act “in these instances does in truth offend the Constitution by its provisions of funds, in that it exposes State money for use in advancing religion, no matter the vigilance to avoid it.” 387 F. Supp. 1282, 1298 (1974) (emphasis in original). Each of the institutions is a church-affiliated or church-related body. The subsidiary findings concerning the role of religion on each of the campuses, summarized by the plurality opinion, ante, at 755-758, conclusively establish that fact. In that circumstance, I agree with Judge Bryan that “[o]f telling decisiveness here is the payment of the grants directly to the colleges unmarked in purpose. . , . Presently the Act is simply a blunderbuss discharge of public funds to a church-affiliated or church-related college.” 387 F. Supp., at 1298-1299. In other words, the Act provides for payment of general subsidies to religious institutions from public funds and I have heretofore expressed my view that “ [g] eneral subsidies of religious activities would, of course, constitute impermissible state involvement with religion.” Walz v. Tax Comm’n, 397 U. S. 664, 690 (1970) (concurring opinion). This is because general subsidies “tend to promote that type of-interdependence between religion and state which the First Amendment was designed to prevent.” Abington School Dist. v. *771Schempp, 374 U. S. 203, 236 (1963) (Brennan, J., concurring). “What the Framers meant to foreclose, and what our decisions under the Establishment Clause have forbidden, are those involvements of religious with secular institutions which . . . serve the essentially religious activities of religious institutions.” Id., at 294-295.

The history of the bitter controversies over public subsidy of sectarian educational institutions that began soon after the Nation was formed is recited in my separate opinion in Lemon v. Kurtzman, 403 U. S. 602, 642 (1971) (Lemon I). My reasons for concluding in Lemon I that all three statutes there before us impermissibly provided a direct subsidy from public funds for activities carried on by sectarian educational institutions also support my agreement with Judge Bryan in this case that “an injunction should issue as prayed in the complaint, stopping future payments under the Maryland Act to the [appellee] colleges.” 387 F. Supp., at 1300. I said in Lemon I, supra, at 659-660:

“I believe that the Establishment Clause forbids . . . Government to provide funds to sectarian universities in which the propagation and advancement of a particular religion are a function or purpose of the institution. . . .

“I reach this conclusion for [these] reasons . . . : the necessarily deep involvement of government in the religious activities of such an institution through the policing of restrictions, and the fact that subsidies of tax- monies directly to a sectarian institution necessarily aid the proselytizing function of the institution. . . .

“. . . I do not believe that [direct] grants to such a sectarian institution are permissible. The reason is not that religion 'permeates’ the secular education that is provided. Rather, it is that the secular edu*772cation is provided within the environment of religion; the institution is dedicated to two goals, secular education and religious instruction. When aid flows directly to the institution, both functions benefit.” (Emphasis in original.)

The discrete interests of government and religion are mutually best served when each avoids too close a proximity to the other. “It is not only the nonbeliever who fears the injection of sectarian doctrines and controversies into the civil polity, but in as high degree it is the devout believer who fears the secularization of a creed which becomes too deeply involved with and dependent upon the government.” Abington School Dist. v. Schempp, supra, at 259 (Brennan, J., concurring). The Maryland Act requires “too close a proximity” of government to the subsidized sectarian institutions and in my view creates real dangers of the “secularization of a creed.” Ibid.; Lemon I, supra, at 649 (opinion of Brennan, J.).

Unlike Judge Bryan, 387 F. Supp., at 1300, I would also reverse the District Court’s denial of appellants’ motion that the appellee institutions be required to refund all payments made to them. I adhere to the views expressed in Mr. Justice Douglas’ dissent, which I joined, in Lemon v. Kurtzman, 411 U. S. 192, 209 (1973) (Lemon II):

“There is as much a violation of the Establishment Clause of the First Amendment whether the payment from public funds to sectarian schools involves last year, the current year, or next year. . . .

“Whether the grant is for . . . last year or at the present time, taxpayers are forced to contribute to sectarian schools a part of their tax dollars.”

I would reverse the judgment of the District Court and remand with directions to enter a new judgment per*773manently enjoining the Board of Public Works of the State of Maryland from implementing the Maryland Act, and requiring the appellee institutions to refund all payments made to them pursuant to the Act.

dissenting.

In my view, the decisive differences between this case and Tilton v. Richardson, 403 U. S. 672, lie in the nature of the theology courses that are a compulsory part of the curriculum at each of the appellee institutions and the type of governmental assistance provided to these church-affiliated colleges. In Tilton the Court emphasized that the theology courses were taught as academic subjects.

“Although all four schools require their students to take theology courses, the parties stipulated that these courses are taught according to the academic requirements of the subject matter and the teacher’s concept of professional standards. The parties also stipulated that the courses covered a range of human religious experiences and are not limited to courses about the Roman Catholic religion. The schools introduced evidence that they made no attempt to indoctrinate students or to proselytize. Indeed, some of the required theology courses at Albertus Magnus and Sacred Heart are taught by rabbis.” Id., at 686-687.

Here, by contrast, the District Court was unable to find that the compulsory religion courses were taught as an academic discipline.

“[T]he hiring patterns for religion or theology departments are a special case and present a unique problem. All five defendants staff their religion or theology departments chiefly with clerics of the affiliated church. At two defendants, Western *774Maryland and Mt. St. Mary’s, all members of the religion or theology faculty are clerics. The problem presented by the make-up of these departments is obvious. Recognition of the academic freedom of these instructors does not necessarily lead to a conclusion that courses in the religion or theology departments at the five defendants have no overtones of indoctrination.

"The theology and religion courses of each de* fendant must be viewed in the light of that shared objective [of encouraging spiritual development of the students]. While most of the defendants do not offer majors in religion or theology, each maintains a vigorous religion or theology department. The primary concern of these departments, either admittedly or by the obvious thrust of the courses, is Christianity. As already noted, the departments are staffed almost entirely with clergy of the affiliated church. At each of the defendants, certain of these courses are required.

“. . . [A] department staffed mainly by clerics of the affiliated church and geared toward a limited array of the possible theology or religion courses affords a congenial means of furthering the secondary objective of fostering religious experience.” 387 F. Supp. 1282, 1294-1296 (emphasis in original).

In light of these findings, I cannot agree with the plurality’s assertion that there is “no constitutionally significant distinction” between the colleges in Tilton and those in the present case. Ante, at 759. The findings in Tilton clearly established that the federal building-construction grants benefited academic institutions that made no attempt to inculcate the religious beliefs of the affiliated church. In the present case, by contrast, *775the compulsory theology courses may be “devoted to deepening religious experiences in the particular faith rather than to teaching theology as an academic discipline.” 387 F. Supp., at 1288. In view of this salient characteristic of the appellee institutions and the non-categorical grants provided to them by the State of Maryland, I agree with the conclusion of the dissenting member of the three-judge court that the challenged Act “in these instances does in truth offend the Constitution by its provisions of funds, in that it exposes State money for use in advancing religion, no matter the vigilance to avoid it.” Id., at 1298- (emphasis in original).

For the reasons stated, and those expressed by Mr. Justice Brennan and Mr. Justice Stevens, I dissent from the judgment. of the Court and the plurality’s opinion.

dissenting.

My views are substantially those expressed by Mr. Justice Brennan. However, I would add emphasis to the pernicious tendency of a state subsidy to tempt religious schools to compromise their religious mission without wholly abandoning it. The disease of entanglement may infect a law discouraging wholesome religious activity as well as a law encouraging the propagation of a given faith.