concurring and dissenting:

The school district’s policy requiring all pregnant teachers to take an involuntary, unpaid leave of absence one month before their expected date of confinement violated Title VII. The majority correctly decides that deLaurier made out a prima facie case of discrimination, but the majority affirms the district court’s decision on the mandato *686 ry termination issue by saying that factual findings, which are not clearly erroneous, established a business necessity defense. That conclusion rests on a misreading of controlling principles of law. In my view, no remand is justified on the denial of sick leave issue because, as a matter of law, the school district cannot establish a defense of business necessity based upon the additional expense of including pregnancy in its sick leave coverage.

I

The school district’s effort to justify its mandatory termination date as a business necessity failed because the district did not prove the job-relatedness of the termination date as it specifically related to teachers over eight months pregnant, and the record showed that there were alternative methods, without a similarly undesirable discriminatory effect, whereby the district could have had ample notice of intended leave to fill its business purpose of acquiring substitute teachers.

The district court did not apply the correct job-relatedness test, and, therefore, its findings of fact are not directly relevant to the business necessity defense issues. To the extent that they are relevant, however, the findings are not supported by the evidence in the record, and they are, therefore, clearly erroneous.

To establish a prima facie violation of Title VII, 1 a plaintiff need only show that the employment practice in question operates “in a significantly discriminatory pattern.” (Dothard v. Rawlinson (1977) 433 U.S. 321, 329 , 97 S.Ct. 2720 , 53 L.Ed.2d 786 ; Albemarle Paper Co. v. Moody (1975) 422 U.S. 405, 425 , 95 S.Ct. 2362 , 45 L.Ed.2d 280 .) Once a discriminatory effect is shown, the employer has the burden to justify the practice by proving that it is “necessary to safe and efficient job performance. . . .” (Dothard v. Rawlinson, supra, 433 U.S. at 332 n.14, 97 S.Ct. at 2728 ; Nashville Gas Co. v. Satty (1977) 434 U.S. 136, 143 , 98 S.Ct. 347 , 54 L.Ed.2d 356 ; Griggs v. Duke Power Co. (1971) 401 U.S. 424, 431-32 , 91 S.Ct. 849 , 28 L.Ed.2d 158 .) Even if the employer establishes a business necessity for the practice, the plaintiff will nevertheless prevail by showing that alternative methods, without a similarly undesirable discriminatory effect, would serve the employer’s legitimate interests as well as the challenged practice. (Dothard v. Rawlinson, supra, 433 U.S. at 329 , 97 S.Ct. 2720 ; Albemarle Paper Co. v. Moody, supra, 422 U.S. at 425 , 95 S.Ct. 2362 .)

Employment practices that disadvantage pregnant women are a violation of Title VII if those practices have a “discriminatory effect,” in that their necessary operation disproportionately burdens women as a class. (Nashville Gas Co. v. Satty, supra, 434 U.S. 136 , 98 S.Ct. 347 , 54 L.Ed.2d 356 ; General Electric Co. v. Gilbert (1976) 429 U.S. 125 , 97 S.Ct. 401 , 50 L.Ed.2d 343 .) A gender-based discriminatory effect arises from distinctive treatment of pregnancy not only when employers’ practices are a pretext for discriminating against women, but also where the practices impose on women as a class “a substantial burden that men need not suffer” because of “their different role” in the reproductive process. (Nashville Gas Co. v. Satty, supra, 434 U.S. at 142 , 98 S.Ct. at 351 .)

“A prima facie case of discrimination can be made by showing disability due to pregnancy, action based on that disability which adversely affects employment opportunities and that others were not similarly treated when suffering temporary disabilities.” *687 (Roller v. City of San Mateo (9th Cir. 1977) 572 F.2d 1311, 1313 .)

deLaurier established a prima facie case under Title VII. The district’s mandatory maternity leave policy did not simply deny pregnant women a benefit that other employees did not receive. Rather, it imposed a special burden on pregnant women by placing restrictions on their employment opportunities which were inapplicable to any other class of teachers taking leaves of absence. All other leaves, whether health-related or not, were voluntary. All health leaves but those for pregnancy began on a date chosen individually by the teacher and lasted for a period determined by the individual teacher’s needs. Only in the case of pregnancy did the district either require a leave or enforce a uniform date for beginning the leave. The mandatory leave policy not only deprived pregnant teachers of their opportunity to earn a living during the leave period, but also imposed additional losses, such as the inability to acquire tenure, deprivation of health and welfare benefits (granted only for leaves with pay), and, at the time of deLaurier’s termination, the lack of any assurance of being returned to the same position upon expiration of the leave.

The mandatory leave policy burdened only women teachers. It imposed the same kinds of disproportionate burdens on the disadvantaged class (women teachers) as the practices that were invalidated in Nashville Gas Co. v. Satty, supra, 434 U.S. 136 , 98 S.Ct. 347 , 54 L.Ed.2d 356 (divesting only pregnant women of seniority); Dothard v. Rawlinson, supra, 433 U.S. 321 , 97 S.Ct. 2720 , 53 L.Ed.2d 786 (height and weight standards disproportionately screened out women); and Albemarle Paper Co. v. Moody, supra, 422 U.S. 405 , 95 S.Ct. 2362 , 45 L.Ed.2d 280 and Griggs v. Duke Power Co., supra, 401 U.S. 424 , 91 S.Ct. 849 , 28 L.Ed.2d 158 (tests barred primarily blacks from employment).

The burden thereupon shifted to the district to prove that business necessity justified its disadvantageous treatment of pregnant women. The business necessity defense is, of course, available regardless of the nature of the practice under review but it is not, as the majority seems to imply, a defense that is easy to establish. On the contrary, it is a highly restrictive and carefully limited defense. The employer must show that the practice is “necessary to safe and efficient job performance . . . .” (Dothard v. Rawlinson, supra, 433 U.S. at 332 n.14, 97 S.Ct. at 2728.) The defense cannot be established merely by a showing that it is administratively convenient to the employer, or even by a showing that other practices would be highly inconvenient. The employer must show that the practice in question is specifically required for the operation of the business. Evidence must be offered correlating the particular practice with factors relevant to the ability to perform the affected job. (Dothard v. Rawlinson, supra, 433 U.S. at 331 , 97 S.Ct. 2720, 2728 .) In Dothard , for example, the employer argued that height and weight requirements for prison guards were necessary because the job required physical strength. The Court held that the defense of business necessity was not thereby established because the employer did not produce evidence “correlating the height and weight requirements with the requisite amount of strength thought essential to good job performance.” The Court emphasized the need for evidence “in specific justification of the . standards.” (Ibid.) The business necessity defense is thus only available if it is shown that the requirement is sufficiently particularized that it “measure[s] the person for the job and not the person in the abstract.” 2 (Griggs v. Duke Power Co., supra, 401 U.S. at 436 , 91 S.Ct. at 856 .)

The evidence shows that the district applied a mandatory leave policy only to preg *688 nancy leaves. All other leaves of absence were voluntary, and advance notice was not required for any other health-related leave. The date for beginning all other kinds of leave was determined on an individualized basis, taking the needs of the particular employee into account. Therefore, the business necessity defense must be based on evidence that proves that the ninth month of pregnancy presents problems that are not presented in any of the earlier stages of pregnancy or in any other type of temporary disability, and that the problems thus presented are such that the safe and efficient operation of the school system requires that all teachers who are nine months pregnant be kept out of the schools. If the reasons advanced by the district for its policies are applicable equally to conditions to which the policy is not applied, then deLaurier has been treated disadvantageously as compared to others similarly situated.

The district tried to justify its policy as a business necessity on three grounds: (1) The administrative needs and educational goals of the district required that a date certain for beginning leave be known in advance to help the district in engaging substitute teachers; (2) a teacher entering the ninth month of her pregnancy becomes unable to perform her duties adequately; and (3) the health of the mother and child requires mandatory leave at the end of the eighth month. The district failed to offer evidence in specific justification of the job-relatedness of its employment cut-off date on any of these three grounds.

The contention that the school district needed ample time to find appropriate replacements for pregnant teachers is not a justification for compelling the teacher to leave work at the beginning of the ninth month of pregnancy. Of course, it is always convenient for an employer to know when any employee will be absent. Nothing in the evidence, however, suggested any reason why advance notification should be mandatory only for pregnancy and not, for example, for elective surgery. Even assuming that the district showed a business necessity for advance notice and the fixing of a date certain for the long-term substitute to begin work, this need could be fully and adequately served by an agreement between the teacher and the school district that her leave would begin on a particular date. In order to successfully rebut deLaurier’s prima facie case, the district had to prove a business necessity for a uniform termination date at the beginning of the ninth month. Apparently my brethren believe that the evidence concerning the difficulty of predicting the exact date of confinement satisfies this requirement. It does not.

The district court found that 80 percent of all births would occur within “plus or minus one week” of the predicted date, and that 98 percent would occur within three weeks of the predicted date. There was uncontradicted testimony that half of the predicted error would occur before and half after the predicted date, and, of course, some of the early deliveries would occur before the ninth month. The court also found that 110 teachers in the San Diego Unified School District commenced maternity leaves in 1973-74. We may infer from these facts that in 1973-74 only 11 teachers in the district would be likely to give birth before one week prior to their predicted delivery date. Since the district’s Director of Personnel Administration testified that fewer than six teachers that year had wanted to teach into the ninth month, the actual number of early deliveries would be considerably fewer — perhaps one or two. I cannot believe that this order of uncertainty justifies requiring all pregnant teachers to leave work four weeks before their expected delivery date. The evidence shows, at most, a business necessity for a termination date one week before the predicted date of confinement. Since unanticipated early onset of delivery was not shown to be in any respect different from other unanticipated temporary conditions which were handled routinely by the district without any notice whatever, the small number of cases in which birth occurred before a one-week cut *689 off could be handled through the normal short-term substitute process. 3

The district proved that advance notice and a date certain for commencing leave would be administratively convenient for the district. It did not show that the administrative needs and educational goals of the district required any form of mandatory leave, or any particular date for beginning the leave.

The district next argued that as pregnancy progresses, women become unable to do a good job of teaching because of irritability, lack of balance and agility, and fatigue. The district court found that “there is a declining ability of a teacher to perform the multifarious duties of a teacher as the date of delivery approaches.” That finding does not aid the district for two reasons. First, the finding was not specifically related to the district’s automatic eighth-month cutoff date, and thus failed to meet the Dothard standard. Second, the finding is not supported by the evidence in the record. No testimony of any kind was presented tending to prove that women in the ninth month of pregnancy were less competent as teachers than women in earlier stages of pregnancy, or less competent than other teachers who were not pregnant. Some anecdotal testimony was given by several school administrators and principals concerning their observations of pregnant teachers. Because the district’s policy of mandatory leave precluded these witnesses from observing any teachers more than eight months pregnant, their testimony was irrelevant to prove the necessity for the eighth month date. Some administrators expressed fear that due to fatigue or fear of collisions, a teacher in the ninth month of pregnancy would decline to carry her fair share of the duties of hall and playground monitoring. No factual foundation for these opinions was cited. Similarly, the medical testimony did not establish that teachers in the ninth month of pregnancy were physically unable to perform their duties.

The testimony regarding inability to perform satisfactorily was further defective because it was not correlated with factors related to the performance of any particular teaching jobs. An employer may not avail itself of the business necessity defense unless the necessity is shown as to the particular job to which the practice is applied. (Dothard v. Rawlinson, supra, 433 U.S. at 331 , 97 S.Ct. 2720 ; Albemarle Paper Co. v. Moody, supra, 422 U.S. 405 , 95 S.Ct. 2362 , 45 L.Ed.2d 280 ; Griggs v. Duke Power Co., supra, 401 U.S. at 431, 436 , 91 S.Ct. 849 .) The San Diego Unified School District’s policy was applied across the board to every teacher in the district. Specifically, it was applied to deLaurier, who was a junior high school language teacher with minimal responsibilities for supervising hallways and playgrounds. The testimony regarding increased girth and loss of agility involved difficulties experienced by elementary school teachers in bending over small desks and supervising strenuous outdoor recreation. These concerns were simply not relevant to deLaurier’s job.

Finally, the district claimed that continued teaching during the final month of pregnancy increased risks to the health of the mother and child. Complications unique to pregnancy can, of course, occur at any time during a pregnancy. Expert witnesses for both the district and deLaurier testified that the most hazardous period of pregnancy was the first trimester when the danger of miscarriage is most acute. Some complications of pregnancy, including toxemia, premature rupture of the membranes, false labor, and premature separation of the placenta, tend to occur in the later stages of pregnancy. None of the expert witnesses testified that these unusual complications of pregnancy were more likely to occur among pregnant teachers who continued teaching than among other pregnant women.

The district court found: “Although serious complications unique to the ninth *690 month of pregnancy occur in only three percent of all pregnancies, the very duties required of teachers entail a possibility of serious dangers to the health of a mother and her child during the ninth month of pregnancy.” This finding was not supported by the evidence, and is clearly erroneous. There was no testimony that any complication was unique to the ninth month of pregnancy. There was no testimony correlating teachers’ duties with increased dangers to pregnant women. 4 The district court also found: “There is greater likelihood of injury to a mother and her child during the ninth month of pregnancy than before the ninth month of pregnancy.” This finding is not supported by the evidence, and is irrelevant in the absence of any evidence that the risks are greater for pregnant women who are teaching than for those who are not.

Some kinds of jobs for some teachers may be hazardous during pregnancy, but the school district completely failed to prove that teachers in the school district generally, teachers in her particular school, or deLaurier herself held jobs that entailed any such hazards or dangers. 5

The district’s justifications based on administrative convenience and decreasing ability to teach are virtually identical to those made, and rejected by the Supreme Court, in Cleveland Bd. of Education v. LaFleur (1974) 414 U.S. 632 , 94 S.Ct. 791 , 39 L.Ed.2d 52 . The Court in that case invalidated a mandatory requirement that pregnant teachers take maternity leaves after the fifth month of pregnancy on the ground that the “irrebuttable presumption” that teachers were unfit to work after the fifth month was arbitrary. (See also id. at 651-57 , 94 S.Ct. 791 , Powell, J., concurring in result, applying an equal protection analysis.) The Court observed that the Board’s interest in administrative convenience and in having physically fit teachers could be adequately served by the less intrusive means of having teachers give ample advance notice of their pregnancy and a “date certain” for termination, together with an individualized' evaluation of their ability to perform their jobs. While the Court did not hold that uniform dates for maternity leaves were never permissible, it did hold that employers who chose such termination or leave practices would have to meet the heavy burden of proving that the particular method was justified by proof that pregnant women at the chosen stage of pregnancy could not adequately perform their jobs or that administrative considerations compelled use of the particular cut-off date chosen. ( 414 U.S. at 640-50 , 94 S.Ct. 791 .) 6 .

Although LaFleur rested on constitutional grounds, it is nevertheless instructive because a mandatory leave policy that cannot pass due process muster necessarily fails to meet the more stringent standards of Title VII. Here, as in LaFleur , the school dis *691 trict offered no reason why the choice of a firm date by the teacher in consultation with her physician would not have served the administrative convenience of the school district as well as the date imposed by the district. The school district did not even attempt to prove that either administrative necessity or medical considerations required placing all pregnant teachers, no matter what their teaching assignments, on mandatory leave at the end of the eighth month of pregnancy. 7

Even if the district court’s findings were unassailable, the business necessity conclusion cannot be sustained because the district court completely failed to consider the existence of less burdensome alternatives, such as individualized decision-making, that would serve the legitimate interests of the school district as well as the mandatory leave policy. The evidence presented by both parties, and the court’s own findings of fact, strongly suggest the availability of such alternatives, and Dothard requires that they be considered in establishing the business necessity defense.

Since the mandatory leave policy was in violation of Title VII, we have no reason to reach deLaurier’s constitutional claims based on the equal protection and due process clauses of the Fourteenth Amendment. Indeed, in the present posture of the case, it is improper to discuss the constitutional issue no matter what the determination of the statutory claim. The California legislature has enacted a statute requiring that henceforth pregnancy-related disabilities be treated the same as other temporary disabilities. Thus, only deLaurier’s claim for her own individual damages remains. The complaint stated no claim for relief in damages on a constitutional theory under 42 U.S.C. § 1983 . Therefore, the majority’s discussion of the Fourteenth Amendment is completely unnecessary and is dictum. Based on my understanding of the Supreme Court’s decision in LaFleur , I disagree with the majority’s constitutional discussion. I find it unnecessary to spell out my disagreements in detail because the issue is not properly before us.

II

Remand on the denial of sick pay issue serves no useful purpose. Contrary to the suggestion of the majority, the district cannot establish a defense of business necessity based upon the extra expense of including pregnancy in its sick leave coverage. Under legislation intervening after the decision by the district court, pregnancy-related disabilities must be included in sick leave programs of California school districts. Thus, even if it costs more to provide sick leave for pregnancy, that issue is irrelevant since the passage of the legislation. The only expense which is involved is deLaurier’s own sick pay. Under the sick pay scheme in existence during deLaurier’s pregnancy, full-salary sick leave benefits are fully vested when earned. Sick leave is earned at the rate of 10 full salary days per year. Unused sick leave accumulates from *692 year to year without limit, and any unused days are paid to the employee on retirement. The additional expense of including deLaurier’s sick leave pay involves such a minute expense to the district that it could not possibly satisfy the requirements of business necessity.

I would reverse with directions to declare the mandatory leave policy a violation of Title VII, and restoration to deLaurier of sick pay, together with reasonable attorney’s fees both for attorney’s services in the district court and on appeal.

. The Equal Employment Opportunity Act, Title VII, Section 703 of the Civil Rights Act of 1964 (42 U.S.C. § 2000e-2(a) provides in pertinent part:

. An employment test given as a condition of hire or transfer, for example, must be “demonstrably a reasonable measure of job performance” on a particular job. (Griggs v. Duke Power Co., supra, 401 U.S. at 436 , 91 S.Ct. at 856 .) The test must be validated against particular jobs; it may be used for a job that has not been specifically studied only if there is “no significant difference” between that job and one against which the test has been validated. (Albemarle Paper Co. v. Moody, supra, 422 U.S. at 432 , 95 S.Ct. 2362 .)

. See Cleveland Bd. of Education v. LaFleur (1974) 414 U.S. 632 , 642 n.10, 94 S.Ct. 791 , 39 L.Ed.2d 52 .

. The only evidence that was remotely relevant to this issue was testimony that premature separation of the placenta, an extremely rare phenomenon, might be caused by a sharp blow to the abdomen. But there was no evidence that teachers performing their regular duties, whether pregnant or not, were subjected even occasionally to sharp blows to the abdomen. The sole mention of any physical contact between a student and a teacher was the testimony of a high school principal that he had been “jostled” while monitoring hallways. Ms. deLaurier’s school has no hallways; all classrooms are in individual buildings.

. Teachers of physical education, mechanical and industrial arts, or other kinds of teaching duties that require physical agility, lifting, climbing, or other kinds of similar work may be classified differently if medical opinion exists that job performance will be lessened or risks unduly increased after a certain stage of pregnancy. (E. g., 7 FEP Cases 455 (1971) (no Title VII violation in requiring employee to leave at end of seventh month of pregnancy rather than the eighth, as the employee desired, where job involved filling orders for heavy materials, lifting, climbing ladders, riding forklifts, and other activities dangerous or difficult for a person in advanced stages of pregnancy).)

. In a footnote, the Court indicated that a fixed, uniform date during the last few weeks of pregnancy might be justified by considerations such as “widespread medical consensus about the ‘disabling’ effect of pregnancy on a teacher’s job performance during these latter days . . . (Id. at 647 n.13, 94 S.Ct. at 800 .) The school district here failed to produce evidence of any such medical consensus.

. The draftsmen of the 1972 amendments of Title VII which expanded its coverage to state and local government employees like deLaurier (see LaFleur, supra, 414 U.S. at 639 n.8, 94 S.Ct. 791 ) noted that these employees already had similar remedies under 42 U.S.C. § 1981 but that Title VII coverage would be co-extensive with and supplementary to those pre-existing rights. The draftsmen made the following observations:

(H.R.Rep.No.92-238, 92d Cong., 2d Sess. (1972) reproduced in 1972 U.S.Code Cong. & Admin. News at 2154; and see id. at 2152-55 .)

In LaFleur , the Court observed that Title VII provided for similar relief in that guidelines promulgated by the Equal Employment Opportunity Commission provided that a mandatory leave or termination policy for pregnant women presumptively violated Title VII. ( 414 U.S. at 639 n.8, citing 29 C.F.R. § 1604.10 , 37 Fed. Reg. 6837 .)