The Equal Employment Opportunity Commission (EEOC) and Calvin Price, a former lieutenant in the Missouri State Highway Patrol (Patrol), sued the Patrol, the Missouri State Highway Commission, and various state agency officials, alleging violations of the Age Discrimination in Employment Act, 29 U.S.C. §§ 621-634 [hereinafter ADEA].1 The District Court ruled that the Patrol’s mandatory retirement age of sixty for all uniformed members violates the ADEA, that the maximum hiring age of thirty-two for radio operators violates the ADEA, and that the maximum hiring age of thirty-two for patrolmen does not violate the ADEA. EEOC v. Missouri State Highway Patrol, 555 F.Supp. 97 (W.D.Mo. 1982). The Patrol appeals from the holdings which invalidated the mandatory retirement age and the maximum hiring age for radio operators. In its cross-appeal, the EEOC challenges that portion of the decision which upholds the maximum hiring age for patrolmen. We reverse the Dis- ' trict Court’s invalidation of the mandatory retirement age and the maximum hiring age for radio operators. We affirm its holding that the maximum hiring age for Patrolmen is valid.
On December 7, 1983, the Patrol also filed in this Court a petition for a writ of *449prohibition, seeking to prevent the District Court from reopening the issue of a back pay award for Sergeant James R. Hightower, a patrol member who was mandatorily retired. The parties were instructed that no ruling would be made on the petition at that time, that the EEOC should file a written response, and that the parties should be prepared to discuss the writ application on December 13, 1983, when the merits of the case were to be argued. On December 22, 1983, we granted the petition pending further order of this Court. Because our resolution of the merits of this case rules out further consideration of relief for Hightower, we now direct the District Court to dismiss any claim that may still be pending regarding back pay for Hightower.
I.
Enacted in 1967, the ADEA originally applied only where the covered employer “engaged in an industry affecting commerce.” Pub.L. No. 90-202, § 11(b), 81 Stat. 602, 605 (codified at 29 U.S.C. § 630(b)). An amendment passed in 1974, however, makes state and local governments subject to the statute. Pub.L. No. 93-259, § 28(a)(2), 88 Stat. 55, 74 (codified at 29 U.S.C. § 630(b)).
The Patrol has not challenged the constitutionality of the extension of the ADEA to the .states. Instead, the Patrol argues that its policies survive scrutiny under the statute. Accordingly, we need not and do not inquire into the constitutional issues.2
II.
The ADEA forbids employment restrictions based on age unless age is “a bona fide occupational qualification reasonably necessary to the normal operation of the particular business ...” (BFOQ). 29 U.S.C. § 623(f)(1). Thus, the ADEA recognizes that in some instances age is inherently related to ability and may be a valid occupational qualification.
Our Court has set forth the standards in age discrimination cases in Hoefelman v. Conservation Commission of the Missouri Department of Conservation, 718 F.2d 281 (8th Cir.1983); EEOC v. City of St. Paul, 671 F.2d 1162 (8th Cir.1982); and Houghton v. McDonnell Douglas Corp., 553 F.2d 561 (8th Cir.), cert. denied, 434 U.S. 966, 98 S.Ct. 506, 54 L.Ed.2d 451 (1977). Under these cases, the employer has the burden of proving that its actions are within the scope of the BFOQ exemption, and the BFOQ test is that set forth in Usery v. Tamiami Trail Tours, Inc., 531 F.2d 224, 234-36 (5th Cir.1976).
The teaching of these precedents is that, in order to satisfy the BFOQ standard, the Patrol has the burden of establishing (1) a correlation between the age limitations in question and the safe and efficient performance of the Patrol’s functions, see Tamiami, 531 F.2d at 235 & n. 27; EEOC v. University of Texas Health Science Center, 710 F.2d 1091, 1094 (5th Cir.1983), and (2) that it has a factual basis for believing either that substantially all older uniformed Patrol members are unable to perform their duties safely and efficiently, or that some older Patrol members possess traits which preclude safe and efficient job performance and which cannot practically be ascertained other than through knowledge of a Patrol member’s age, see Tamiami, 531 F.2d at 236-37.
We note that the age restrictions under which the Patrol operates are mandated by state law.3 The Missouri General *450Assembly thus has made a legislative judgment that the age restrictions here in issue are in the best interest of the Patrol and the people the Patrol serves throughout the state. This fact does not relieve the Patrol from the burden of showing that each of these restrictions is a BFOQ. See Hoefel-man v. Conservation Commission, 718 F.2d 281 (8th Cir.1983). At the same time, in applying the BFOQ exemption, we should be guided by sound principles of federalism and should accord some deference to the state legislative declaration. See EEOC v. City of St. Paul, 671 F.2d at 1167 (legislative determination not entitled to presumption of correctness, but “[tjhis is not to say that a legislative declaration is not entitled to considerable deference.”) Cf. Massachusetts Board of Retirement v. Murgia, 427 U.S. 307, 96 S.Ct. 2562, 49 L.Ed.2d 520 (1976) (holding that Massachusetts statute making it mandatory for a uniformed state police officer to retire at age fifty does not deny equal protection of the laws in violation of the Fourteenth Amendment).
Guided by the standards and considerations outlined above, we have carefully reviewed the record to determine whether the District Court’s findings of fact are clearly erroneous and whether the District Court correctly applied the law. See EEOC v. City of St. Paul, 671 F.2d at 1166. We conclude that the District Court committed error in invalidating the mandatory retirement age for Patrol members and the maximum hiring age for radio operators, and that it correctly upheld the maximum hiring age for Patrol members.
III.
We turn initially to the question of the mandatory retirement age for Patrol members. Although the District Court made several findings concerning the duties and qualifications of Patrol members, it did not directly discuss the first part of the Tamiami test set forth above.4 We discuss it here in order to clarify the demands that are placed on all Patrol members regardless of age or rank. The central concern of this part of the BFOQ test is illuminated by language from Tamiami, in which the court, in sustaining as a BFOQ a bus company’s maximum hiring age of forty for intercity bus drivers, suggested that “[tjhe greater the safety factor, measured by the likelihood of harm and the probable severity of that harm in case of an accident, the more stringent may be the job qualifications designed to insure safe driving.” 531 F.2d at 236. Employers are entitled to substantial discretion in judging the reasonableness of safety-related job qualifications. Id. at 236 n. 30.
The Patrol’s main functions are to enforce the traffic laws and to maintain safety on the highway. The Patrol also does criminal work including investigations of burglaries, murders, and other crimes. Patrolmen and corporals spend the most time working the roads and conducting investigations. Higher ranking Patrol members often travel on the road as part of their normal duties or coming to and from work.5 The Patrol members’ obligation to the public extends to their off-duty hours. At all times, they are expected to enforce the traffic laws and to act when a felony is committed in their presence, when someone’s life is in danger, when property is threatened, and when they receive information requiring Patrol attention. All ranks, including those which are primarily administrative, participate in special assignments. For example, the Patrol steps up its policing of the highways on holiday weekends, and the participation of senior officers is required. Other special assignments in-*451elude riot control, civil disturbances, natural disasters, manhunts for escaped convicts, football game detail, and covering important meetings.
A member of the Patrol engages in physically and emotionally strenuous activities. His duties may involve handling verbal and physical resistance from suspects such as violent drunks or fleeing suspects. Apprehending a speeder or a drunk driver may necessitate a chase, during which the Patrol member must drive at high speeds, weave through highway traffic, and call the radio dispatcher. For instance, when Lieutenant Mills, a desk officer, was returning home from work one day, he chased a speeding motorist. After the motorist fled from his car, Mills chased him on foot. When the motorist turned on Mills and pointed a revolver at him, Mills shot and killed the motorist.
If the Patrol member is called to an accident scene, he determines what assistance is needed, such as the immediate removal of victims from cars, applies first aid to victims, and secures the scene from further damage. Patrol members also have the sad duty of informing relatives of the seriously injured or dead. Members of all ranks recounted instances in which they assisted motorists whose vehicles were mechanically disabled or stuck in the snow. To prevent further accidents, Patrol members push cars off the road and remove hazardous objects, such as dead animals, hay bales, tires, and accident debris from the road.
The Patrol’s obligations to the public mean that Patrol members of all ranks work the road, handle abusive and violent motorists, engage in high speed chases, assist motorists, and participate in other strenuous field work. Although a Patrol member may call for assistance, emergency situations do not always accord him the luxury of waiting for help. In light of these obligations and demands upon Patrol members, the Patrol has clearly established that physical ability and ability to withstand stress are job qualifications which are reasonably necessary to the performance of its functions, and that in these terms there is a correlation between the mandatory retirement age of sixty and the safe and efficient performance of the Patrol’s functions. Cf. Massachusetts Board of Retirement v. Murgia, 427 U.S. 307, 96 S.Ct. 2562, 49 L.Ed.2d 520 (1976).
The Patrol also presented persuasive evidence on both alternatives of the second part of the BFOQ test: (1) a factual basis for believing that substantially all members of the Patrol over sixty are unable to perform their duties safely and efficiently; (2) that there is no reliable alternative to distinguish qualified from unqualified Patrol members other than age. Dr. Alexander Lind, an expert physiologist, testified on the effect of age on a person’s aerobic capacity. Dr. Albert Antlitz, an expert in cardiology with extensive experience with the law enforcement field, explained the susceptibility of older persons to heart disease. In addition, several older Patrol members, including retired members, and Dr. David Westhoff, the physician responsible for administering physical exams to Patrol applicants, testified.
Dr. Lind stated that aerobic capacity becomes important whenever a Patrol member engages in physical activity such as chasing a suspect or pushing a car. Aerobic capacity, which refers to an individual’s ability to utilize oxygen, depends on the efficiency of the individual’s lungs and cardiovascular system. A reduction in aerobic capacity means that the oxygen supply is insufficient, which in turn reduces the ability of the muscles to work without fatigue. When the oxygen supply is insufficient an individual’s muscles utilize an emergency reserve, a special metabolism by which the muscles generate their own energy. Using this emergency reserve, however, generates only limited muscular function, rapidly induces fatigue, and requires more time to recover from the exertion.
Dr. Lind explained that sedentary persons, such as the Patrol’s members, have a peak aerobic capacity at age twenty when they utilize about 3.5 liters of oxygen per minute. A twenty-year-old will not suffer *452fatigue while working at an aerobic level of three liters per minute, which is the desired level for someone in a safety organization. However, three liters is the maximum capacity of a forty-five or fifty-year-old and this individual will suffer some fatigue working at this level. Moreover, fifty percent of the sedentary population at ages forty-five to fifty cannot reach the three liter level without relying on the emergency reserve. By the time Patrol members are age sixty, only ten percent will be able to reach the desired level, including those relying on the emergency reserve. According to Dr. Lind, the data indicates the capacity to do physical work drops off sharply after age sixty.
Generally, an individual loses about ten percent of his aerobic capacity each decade after age twenty; thus, by age sixty, he will have lost approximately thirty-five to forty percent. Although sustained exercise for twenty to thirty minutes three times per week at sixty percent of maximum aerobic capacity can increase an individual’s capacity, the maximum increase is ten percent. Moreover, interruptions in training such as those caused by illness result in a loss of whatever benefit was gained by exercise.
Thus, as individuals age, physiological changes affect their ability to perform. The body’s ability to make efficient use of oxygen declines, with a resulting decline in physical stamina. Dr. Lind also stated that muscle strength, an isometric function, which is involved in pushing, pulling, and carrying, declines as a person ages. In addition, the aging process affects an individual’s reaction time, hearing, and vision. This manifests itself, for example, in an individual’s inability to adapt to sudden changes in lighting, such as the glare from headlights.
Several retired Patrol members candidly admitted that they retired because they could not perform all the duties of the Patrol. Major Closson retired when he could no longer perform routine patrol duty. By retirement age, Captain Schmitt was also finding it more difficult to travel and to work long hours. Captain Tinnin, who retired at sixty, stated that he could not cope with working the road, chasing escapees, and handling drunks. Lieutenant Burnett retired six months before his sixtieth birthday because he could not face working the long days directing traffic and crowds at football games or inclement winter weather with its attendant accidents and hazardous roads. Lieutenant Wright retired early because his driving ability, reflexes, and night vision had deteriorated considerably. The testimony of each of these individuals supported Dr. Lind’s conclusions. Compare Hoefelman, 718 F.2d at 285-86 (plaintiff Hoefelman conceded that pilots over the age of sixty should not be allowed to fly low-level missions).
Coronary heart disease, according to Dr. Antiitz, offers another potential problem for older Patrol members.6 Although age does not cause coronary heart disease, the data reveals an increase in the number of initial discoveries of coronary heart disease as age increases. The average number of discoveries for the general population for persons thirty to thirty-nine is four per thousand. It is ten per thousand by age fifty and twenty per thousand by age sixty — a rate five times higher than that for persons thirty to thirty-nine.7 Latent coronary heart disease presents the danger that, in times of stress, such as sudden *453exertion in cold weather or during a confrontation, the Patrol member’s adrenline response may push him into total body collapse from fatigue or may cause a heart attack.
There is no single test or battery of tests that safely measures a Patrol member’s ability to perform his duties and that identifies, with reasonable accuracy, the presence of coronary artery disease. Dr. Antlitz testified that testing an asymptomatic population such as the Patrol will yield as many false positives of cardiac disease as false negatives.8 Laboratory tests cannot recreate safely the stressful, sometimes life-threatening, situations in which the' Patrol members work. The Patrol member will know that a test cannot be truly life-threatening. Some tests that might be used, such as pushing a car on a very cold day, carry unacceptably high risks of physical harm to the officers being tested. Preparation of the individual to be examined can also affect the reliability of testing. For example, persons who resume taking their blood pressure medication shortly before an examination may show a satisfactory blood pressure level on the day of the examination even though they have not been taking the medication regularly.
The treadmill or stress test often is used to detect the presence of coronary artery disease. Although it is a helpful diagnostic tool, it cannot detect coronary artery disease with complete accuracy. The test results suffer from problems of physician interpretation, which requires subjective assessment of the individual’s performance. For example, if a person reaches only seventy percent of his predicted maximum heart rate, the result may be considered borderline normal. Either coronary artery disease or some other disease may have affected the result. The test, as Dr. Westhoff noted, deals only with an average predictable response. Persons with coronary artery disease may have normal stress test results. Dr. Westhoff, for example, described one of his patients who wanted to participate in a physical fitness program. Although the patient went through screening studies, including a stress test, and came out with normal scores, the patient suffered a heart attack while doing calisthenics on the second day of his physical fitness program.
The Patrol generally disputes the efficacy of tests in measuring the physical qualities needed by Patrol members such as endurance, strength, agility, balance, flexibility, and speed. Similar problems of patient preparation, physician interpretation, and duplication of field demands affect these test results. Duplication of field demands involves more than the Patrol member’s mechanical ability to handle a suspect that may be out to kill him. It involves tensions arising from fear and from distractions such as repeated glare and loss of light. Testing also presents problems of arbitrariness in determining what particular result is a “passing” one which determines that the Patrol member can do the job.9
Although the District Court discounted the Patrol’s expert evidence as “not persuasive” and “often contradictory,” 555 F.Supp. at 104, we find this conclusion wholly unsupported by the record. For example, the District Court concluded that “[ajccording to [Dr. Lind], an average 45-year-old adult male who is sedentary does not have sufficient aerobic capacity to perform the duties of the Patrol.” Id. In *454fact, Dr. Lind testified that at least fifty percent of the sedentary population can perform at age forty-five or fifty at the desired maximum aerobic capacity. The other half must utilize the emergency muscular function which induces rapid fatigue.10 Dr. Lind further testified that by age sixty, ninety percent, or substantially all, Patrol members are unable to perform all their duties safely and efficiently.
The District Court criticized the Patrol for not requiring its members to undergo regular physical examinations and for not attempting “to test its members in accordance with the advice of its own experts.” Id. at 105. In so doing the District Court ignored-the testimony of the Patrol’s expert witnesses as to the inadequacy of physical examinations and testing as methods of distinguishing between qualified and unqualified individuals over the age of sixty.
The District Court also erroneously emphasized the fact that the Patrol does not require its members to participate in a regular physical fitness program. Id. at 106. The Patrol showed that it would be impractical- to force its members to participate in a regular program at a central location. It is inconvenient for members who travel or who are stationed in rural areas to come to a central location. Such a program also leaves some areas without Patrol protection. Moreover, the relevance of a physical fitness program to the issues in this case is questionable. The District Court’s digression into this subject reflects concern with the subjective motivations for the Patrol’s policies. The Tamiami test, however, focuses on objective evidence, particularly objective, medical evidence, in support of the BFOQ. The test also does not evaluate the age restrictions in terms of an ideal employee. It simply requires objective evidence on whether the older employees can perform their duties or on whether there is a practical way to differentiate among employees other than age. As the Patrol suggests, “[t]he district court ... appears to base its decision invalidating the bfoq, on what it considers a 60 year old patrolman could be, if only ‘remade’ as the District Court sees fit.” Reply Brief for Patrol at 2. The ADEA does not require so draconian an approach.
The District Court dismissed the Patrol’s evidence on the greater likelihood of coronary artery disease in older persons as “minimal.” 555 F.Supp. at 104. But the District Court appears to have misunderstood the testimony. For example, the District Court found that “[ojnly .4% of the population develops symptoms of coronary artery disease by age 30, only 1% develops symptoms by age 50, and only 2% by age 60.” Id. at 101.11 The Patrol’s experts actually testified that these statistics represented new discoveries of coronary artery disease in the general population and not the overall incidence of coronary artery disease. Transcript at 613-14. The significance of this data is that it reveals that, regardless of low, average, or high risk categorization, there is an increased inci*455dence of coronary artery disease in the older population. Moreover, according to Dr. Antlitz, cardiovascular disease causes fifty percent of deaths nationwide. Even the plaintiffs’ expert witness conceded that age is a factor in assessing the probability that a given person has coronary artery disease. Id. at 887-88, 937-38.
The District Court relied heavily on the opinion of the plaintiffs’ expert witness, Dr. Stanley Mohler, in concluding that “[w]hen the [treadmill test] is joined with a complete risk factor analysis, physicians can identify nearly every person who is likely to have a heart attack in the following few years.” 555 F.Supp. at 101. Dr. Mohler testified that “if [an] individual is clear on all of these factors ...,” including the family history of heart disease, the individual’s medical status such as blood pressure and pulse rate, the treadmill test, and blood studies, then “you can determine with a probability of .017 that in the next five years ... that individual will not have a cardiovascular heart event . . . .” Transcript at 890, 892. This probability, according to Dr. Mohler, “is far less than one in a hundred.” Id. at 890.12
The District Court failed to consider any of the serious deficiencies in Dr. Mohler’s testimony. Dr. Mohler’s sole explanation as to what he means by “clear on all these factors” is the equally vague description “essentially a normal person.” Id. at 892. His opinion oversimplifies the complex problem of distinguishing among individuals. He says nothing about whether every variation between individuals can be ranked in order to draw lines fairly among individuals to determine who is at risk by continuing to do Patrol work. For those individuals who are “normal,” Dr. Mohler offers his five-year guarantee. But Dr. Mohler says nothing about the factors which undermine the reliability of testing such as patient preparation, physician interpretation, and the realities of the stressful situations that Patrol members encounter in performing their jobs. The medical profession can approach the problem of the likelihood of coronary artery disease by categorizing individuals according to risk factors. But, as the Patrol’s experts pointed out, it still cannot tell us which individuals actually will contract or have contracted the disease and which individuals will suffer sudden cardiac attack. Dr. Mohler’s guarantee reflects only probabilities at best and is not a practical means of determining the individual susceptibility of Patrol members to coronary artery disease.13
In summary, the Patrol has established that there is a factual basis for believing that substantially all Patrol members over the age of sixty lack sufficient aerobic capacity to perform their duties safely and efficiently. The Patrol also has established the inefficacy of testing, as an alternative to age, as a means of distinguishing among individuals over sixty. Chronological age may be a poor predictor of physical ability and of the existence of coronary artery disease and other manifestations of arteriosclerosis. At this time, however, the medical profession, as is evidenced by the disagreements among qualified, respected experts, cannot practically or accurately distinguish between older individuals who can and older individuals who cannot perform safely and efficiently the hazardous and stressful duties that members of the Patrol must perform. In light of the fact that public safety and lives depend upon the capabilities of Patrol members, and in light of the overwhelming evidence of declining physical ability and increasing coronary risk among persons who have reached the age of sixty, we hold that the Patrol clearly has justified its mandatory retirement age.
We also find it persuasive that Congress, which has chosen to extend the ADEA to states and their political subdivisions, au*456thorizes a general mandatory retirement age of fifty-five for federal law enforcement personnel such as FBI agents, CIA agents, United States Marshals, and Secret Service agents. See 5 U.S.C. § 8335(b). In addition, Congress has provided exceptions to this rule including individual exemptions to age sixty and voluntary retirement at age fifty. Id. at § 8335(b), § 8336(c). These Congressionally authorized limits are based upon “the determination that [law enforcement officers] should be composed, insofar as possible, of young men and women physically capable of meeting the vigorous demands of ... occupations calling for the strength and stamina of the young rather than the middle aged.” S.Rep. No. 948, 93rd Cong., 2nd Sess., reprinted in 1974 U.S.Code Cong. & Ad. News 3698, 3699. See generally Stewart v. Smith, 673 F.2d 485 (D.C.Cir.1982) (analysis of relationship between maximum hiring ages for federal law enforcement personnel and ADEA provisions covering federal employees). We believe that it is no less valid for the State of Missouri to be concerned with the youth and vigor of its law enforcement personnel.
We note that our view of the mandatory retirement age issue finds support in recent decisions of the First and Fourth Circuits. See Mahoney v. Trabucco, 738 F.2d 35 (1st Cir.1984) (upholding as a BFOQ a mandatory retirement age of fifty for Massachusetts state police); Johnson v. Mayor of Baltimore, 731 F.2d 209 (4th Cir.1984) (upholding as a BFOQ a mandatory retirement age of fifty-five for City of Baltimore firefighters; mandatory retirement age for federal firefighters a reasonable federal standard against which to compare mandatory retirement age of city firefighters).
IV.
We agree with the District Court that the maximum hiring age of thirty-two for patrolmen is valid under the ADEA. The Patrol established that all ranks are expected to perform duties which are physically and emotionally strenuous. Documentary evidence showed that the average age of patrolmen is 27.5, of corporals 39.6, and of sergeants 45.7. Patrolmen, corporals, and sergeants spend the most time working the road and make the greatest number of arrests. Based on the evidence in the record, the District Court correctly found that it takes about five years of experience before patrolmen become proficient road officers.
The maximum entry age insures that the Patrol can take advantage of the physical skills and abilities of younger persons and also provide those persons with enough experience while they are relatively young to compensate for the inevitable reduction in their physical skills and abilities that comes with aging. The Patrol’s experts testified on the effects of aging, on the statistical correlation between age and coronary artery disease, and on the inadequacy of testing as a means of distinguishing among individuals. A lifting of the maximum hiring age would result in more older members in the lower ranks, which, as indicated above, spend a large proportion of their time on the road and make the greatest number of arrests. The safety of both Patrol members and the public would be in greater jeopardy with less experienced, less physically capable, older members in the lower and middle ranks. In addition, the Patrol explained that it has a policy of systematic promotion through the ranks which fosters good morale by allowing younger members to work toward promotion and to rise through the ranks in a steady progression. As the District Court noted, it takes approximately eleven years for a trooper to gain sufficient experience to serve the Patrol in an administrative capacity. All of these considerations support the conclusion that the maximum hiring age for Patrol members is a BFOQ. See Murnane v. American Airlines, Inc., 667 F.2d 98 (D.C.Cir.1981), cert. denied, 456 U.S. 915, 102 S.Ct. 1770, 72 L.Ed.2d 174 (1982); cf. Stewart v. Smith, 673 F.2d 485.
V.
The District Court based its ruling that the Patrol had failed to establish that its *457maximum hiring age of thirty-two is a BFOQ for radio operators on a finding that the Patrol had not shown “that substantially all persons over age 40 are unable to safely and efficiently perform the duties of a radio operator or that some persons over the age of 40 possess traits precluding safe and efficient performance unascertainable other than through knowledge of that person’s age.” 555 F.Supp. at 106. The District Court reasoned that the Patrol could rely upon testing to screen out applicants lacking the traits a radio operator must possess. Id.
The District Court ignored, however, the Patrol’s strong showing, similar to the showing made regarding the maximum hiring age of thirty-two for troopers, that years of training are required to bring a radio operator in the Patrol to full proficiency, that radio communications are essential to the proper functioning of the Patrol, that in emergency situations lives may depend upon the ability and experience of the radio operator on duty, that the safest and most efficient older radio operator is one with extensive experience in working under the unique conditions of the Patrol and serving its particular needs, and that qualification for promotion to supervisory and administrative levels requires many years of service in the Patrol. The District Court also overlooked the Patrol’s showing that although hearing and vision can be tested, it is impossible for any test to recreate the noisy, hectic, and extremely stressful conditions under which its radio operators daily perform their critical tasks.
We find the Patrol’s evidence highly persuasive. Further, from our review of the record we are unfavorably impressed by the plaintiffs’ efforts, which at best can be described as minimal, to rebut the Patrol’s convincing showing that the age thirty-two hiring cap is a BFOQ. In addition, we are mindful that the Patrol is a para-military organization whose pre-eminent concern is to provide for public safety, that we are dealing with a policy decision made by the legislature of a sovereign state, and that the legislative determination is entitled to some deference. Indeed, in our judgment the state must be accorded considerable discretion in determining the manner in which its law enforcement agencies may perform their vital functions most safely and efficiently. See EEOC v. City of St. Paul, 671 F.2d 1162, 1167. Cf. Mur-nane v. American Airlines, Inc., 667 F.2d 98, 101; Tamiami, 531 F.2d 224, 236 n. 30. Accordingly, we reverse as clearly erroneous the District Court’s conclusion that the maximum hiring age of thirty-two for radio operators is invalid under the ADEA.
VI.
For the reasons set forth above, we reverse the District Court’s rulings that (1) the mandatory retirement age for members of the Patrol violates the ADEA and (2) the maximum hiring age for radio operators violates the ADEA. We affirm the District Court’s ruling that the maximum hiring age for troopers does not violate the ADEA. If any claim is still pending regarding back pay for Sergeant Hightower, we direct the District Court to dismiss it.
. The EEOC challenged the Patrol’s mandatory retirement age for uniformed members and the maximum hiring ages for patrolmen and radio operators. Price challenged only the mandatory retirement age. These actions were consolidated.
. We note that in EEOC v. Wyoming, 460 U.S. 226, 103 S.Ct. 1054, 75 L.Ed.2d 18 (1983), a divided court upheld the extension of the ADEA to the states as an exercise of Congressional power under the commerce clause.
. See Mo.Ann.Stat. §§ 43.060, 104.080, 104.-010(23), (25). Prior to June, 1982, the mandatory retirement statute vested the Patrol Superintendent with discretion to extend the employment of members between ages 55 and 65. The Missouri legislature subsequently lowered the upper age limit to 60. 1982 Mo.Laws 273, § 1. Prior to this amendment, the Patrol routinely granted extensions to persons until age 60. The Patrol’s policy has been to retire all members at age 60 regardless of the member’s health, job responsibilities, or job performance. The only exceptions have been very short extensions *450where legislation pending in the Missouri legislature would favorably affect retirement benefits. Transcript at 21.
. The District Court may have assumed that the Patrol met this aspect of the BFOQ test.
. Patrol members’ entering rank is that of patrolman or trooper. The upward line of progression in rank is patrolman (or trooper), corporal, sergeant, lieutenant, captain, major, lieutenant colonel, and colonel. Colonel Hoffman, the Superintendent of the Patrol, testified that there are Patrol members over the age of fifty-five in every rank.
. In discussing medical problems concerning the heart, Dr. Antiitz distinguished among several different terms. A myocardial infarction involves the death of an area of the heart muscle. Coronary artery disease refers to the narrowing of a blood vessel in the heart area. Two symptoms or related problems of coronary artery disease are angina pectoris and coronary insufficiency. Angina pectoris refers to pain resulting from the lack of blood flow to an area of the heart because of some sort of stress; it does not result in the death of tissue. Coronary insufficiency also involves insufficient blood flow to the heart but is not accompanied by pain. It can result in arrhythmias such as irregular heartbeats and quivering of the heart muscle and, in times of stress, in sudden death.
. The District Court’s incorrect interpretations of this evidence are discussed infra at notes 10-11 and accompanying text.
. Testing a symptomatic population such as those complaining of angina or chest pain will detect seventy to eighty-five percent of those who actually have coronary heart disease.
. - One of the endurance tests advocated by a witness for the plaintiffs requires the subject to sit in a "V” with his feet off the floor for 60 seconds. The Patrol argues that such a test in no way duplicates the endurance needed for the demands of the Patrol. "[T]he Patrol ... ‘tests’ its member’s endurance by sending them out on manhunts for 17-18 hour days, under miserable climatic conditions, without adequate preparation or rest.’.’ Brief for Patrol at 40. Moreover, as the Patrol asserts, the "V” test exemplifies the arbitrariness of such tests and the problems of interpreting what constitutes a "passing” grade: Shall the Patrol immediately fire members who can sit in a "V” for only 58 seconds? 45 seconds? Id. at 40-41.
. The District Court criticized Dr. Lind’s testimony as "overly generalized” and "not factually based on the duties of patrolmen.” 555 F.Supp. at 104. The record does not support the District Court’s view of this witness's testimony. . Dr. Lind carefully explained that aerobic capacity was important to Patrol members’ ability to act efficiently, that is, .without severe fatigue. He also explained that they fall within the sedentary classification because, as a group, they do not engage in regular, vigorous, and sustained exercise. Even if they did, the benefits would be minimal. See supra p. 452. Based on his knowledge of the Patrol members’ activities, and his own studies of the comparable duties in firefighting, Dr. Lind testified on the desired aerobic capacity for Patrol members.
We also find the District Court’s criticism of Dr. Westhoff’s testimony as “self serving,” see 555 F.Supp. at 105 n. 1, an unjustifiable conclusion -on credibility. The experts on both sides had conceded that they were advocates of their respective positions. Dr. Westhoff's testimony was based on his personal experiences as a treating physician. It also was consistent with the testimony of Dr. Lind and Dr. Antlitz.
. Elsewhere-in its opinion, the court, misinterpreting the same evidence, concluded that "[t]he increase in risk is minimal and is present for members of the Patrol whose ages are in their 30’s. The risk peaks for those who are in their 50’s at a rate of one percent.” 555 F.Supp. at 104. But see supra p. 452.
. In fact, a probability of .017 means that 1.7 "normal” individuals per one hundred would have a "cardiovascular heart event.” See generally W. Curtis, Statistical Concepts for Attorneys, 45-57 (1983).
. Cf. supra note 8 and accompanying text.