OPINION
This Court, upon remand from the Court of Appeals after an appeal by defendant Enterprise Association Steamfitters Local 638 of U.A. (“Local 638” or the “Union”), has reconsidered, as discussed herein, the percentage goals for non-white membership in the A Branch of Local 638 1 as originally pro*984vided by this Court’s Order and Judgment of June 21, 1973. 360 F.Supp. 979 (S.D.N.Y.1973).
This case is a consolidation of two actions for the purposes of trial: one of these actions was brought by four nonwhite workers alleging unlawful employment discrimination by Local 638, the Mechanical Contractors’ Association of New York, Inc. (“MCA”), and the Joint Steamfitting Apprenticeship Committee of the Steamfitters’ Industry (“JAC”), in violation of Title VII of the Civil Rights Act of 1964, 42 U.S.C. § 2000e et seq., of 42 U.S.C. §§ 1981 and 1983, and of the Fifth and Fourteenth Amendments (Rios v. Enterprise Association Steamfitters Local 638 of U.A., et al., 71 Civ. 847); and the other action is one of several suits brought by the Attorney General of the United States under Title VII of the Civil Rights Act of 1964, pursuant to the authority granted to him in that Act (42 U.S.C. § 2000e-6(a)) against several local unions in the building trades industry servicing metropolitan New York (Equal Employment Opportunity Commission v. Enterprise Association Steamfitters Local 638 of U. A., et al., 71 Civ. 2877). By Order and Judgment filed June 21, 1973, this Court permanently enjoined the defendants from engaging in any act or practice which had the purpose or effect of discriminating against any individual or class of individuals on the basis of race, color or national origin.
By the Order and Judgment of June 21, 1973, this Court directed the parties to submit to the Administrator appointed therein, Vincent D. McDonnell, Esq., an “affirmative action program” designed to secure admission of a sufficient number of non-whites to membership as journeymen in Local 638’s A Branch “to achieve a minimum goal of 30% non-white membership by July 1, 1977.” On March 29, 1974 this Court adopted an “Affirmative Action Plan” (the “Plan”) to implement the terms of the Order of June 21, 1973 and directed that the minimum goal of 30% nonwhite membership in the A Branch be achieved in four stages: 15% by July 15, 1974; 20% by July 15, 1975; 25% by July 15, 1976; and 30% by July 1, 1977.
Local 638 appealed from several aspects of this Order and Judgment. See Rios v. Enterprise Association of Steamfitters Local 638, 501 F.2d 622, 627 (2d Cir. 1974). The Court of Appeals affirmed this Court’s Order but remanded for a “reassessment of the percentage goal figure” for non-white membership in Local 638 in light of relevant statistical data. 501 F.2d at 633.
In reassessing the percentage goal, this Court notes the guiding principle, as stated by the Court of Appeals:
“[T]he objective of remedial quotas is a limited one. It seeks to place eligible minority members in the position which the minority would have enjoyed if it had not been the victim of discrimination.” 501 F.2d at 632.
The Court of Appeals recommended the use of “reliable statistics with respect to [members of the male] labor force” living within the jurisdiction of Local 638,2 stated that this Court should “be guided by the most precise standards and statistics available,” and noted that the Court “in prescribing non-white goals must at all times be limited to eradication of past discrimination.” 501 F.2d at 633. See also Patterson v. Newspaper and Mail Deliverers’ Union, 514 F.2d 767 (2d Cir. 1975).
At the hearing held on April 17 and 18, 1975, the Court heard the expert testimony of Dr. Marc Rosenblum offered by the plaintiffs and the government, and of Mrs. Shirley Gilbert offered by the defendants Local 638 and MCA. *985After considering these witnesses’ testimony and reviewing the Bureau of the Census data received as exhibits at the hearing, the Court concludes that the July 1, 1977 percentage goal for nonwhite membership in the A Branch of Local 638 should be 26%.
The Basic Calculations
A starting point for calculation is the fact that 19.79% of the male labor force in the jurisdiction of Local 638 is black and Puerto Rican.3 1970 Census of Population, General Social and Economic Characteristics, New York, PC(1)-C34 (hereinafter referred to as “General Characteristics”).
This figure initially must be refined to include only members of the male civilian labor force since men in the armed forces were not available for work as steamfitters. See generally, General Characteristics, supra, Tables 119-131; id., Appendix B, at 15.
In selecting the appropriate Census data for the percentage goal calculation, the Court finds that labor force figures in the 1970 Census are a good measure of the relevant labor force even though those data reflect persons who were age 16 in 1970. Only persons age 18 and older have ever been eligible for membership in Local 638, but an individual who was 16 years old in April, 1970 (when the Census data was collected) would have reached at least age 19 by June 21, 1973, the date of this Court’s Order, and thus would have met the Union’s age requirement.
The plaintiff class, as defined prior to trial, encompassed “Negro and Spanish Sur-named Americans” residing within the jurisdiction of Local 638 who had the skill or the capability and interest to learn the skill of a journeyman steamfitter. Rios v. Enterprise Association Steamfitters Local Union # 638, 54 F.R. D. 234, 237 (S.D.N.Y.1971). Data on Negro and Puerto Rican males are readily available (see General Characteristics, supra, Tables 126, 131), but there are no statistics by the Bureau of the Census on Spanish Sur-named individuals in the New York area. See id. Appendix B, at 7. Therefore, absolute precision in this regard is not possible. The Court concludes that statistics on non-Puerto Rican Spanish Sur-named males in the labor force can best be calculated from data compiled for the Spanish Language population.4 To determine the labor force statistics on Spanish Language males, it is necessary first to determine the ratio of the number of Spanish Language males in a given county to the number of Puerto Rican males in that county (each of which numbers are available in the Census reports),5 and then to multiply the number of Puerto Rican males in the labor *986force in that county by the corresponding ratio.6
Also relevant to the calculation of the percentage goal is the indication from the Census data that 11.6% of “Spanish Origin” males in the labor force within the jurisdiction of Local 638 are black.7 1970 Census of Population, Subject Reports, Persons of Spanish Origin, PC(2)-1C, Table 2. As a result, certain individuals have been “double-counted”, once as a black and once as a member of the Spanish Origin group, thereby artificially inflating any statistic derived by adding figures concerning blacks to figures covering Spanish Language individuals. Therefore, the percentage goal calculations described below include a reduction of 11.6%.
Taking into account the foregoing basic factors, the number of non-white males in the labor force in the jurisdiction of Local 638 may be obtained by taking the sum of the results of the following computations performed by county:
TOTAL NUMBER OF NON-WHITE MALES IN LABOR FORCE =
(Number of \ Black males 1 in labor I force /
(Number of Puerto Rican males in labor force
X
No. Span. Lang, males in popul’n No. Puerto Rican males in popul’n
X (1 - 11.6%)
After adding up the results for the seven counties, the percentage goal then may be derived by the following formula:
Total number of non-white males in labor force within Union’s jurisdiction
PERCENTAGE GOAL =--
Total number of males in labor force within Union’s jurisdiction
which yields the figure 22.7%. See General Characteristics, supra, Tables 121, 126.
Educational Factor
In seeking to determine a more accurate percentage goal, the Court finds that the educational achievement of steamfitters should be taken into account. The Census reports indicate that *987the median 8 educational level for pipe-fitters and plumbers (the Census category acknowledged by the parties as the appropriate indicator) is 11.9 years (1970 Census of Population, Subject Reports, Occupational Characteristics, PC(2)-7A, Table 5 (hereinafter referred to as “Occupational Characteristics”)) and the mean9 educational level is only 10.6 years (1970 Census of Population, Subject Reports, Occupation by Industry, PC(2)-7C, Table 3, at 130). Even more significant is the fact that as of 1970, 91.8% of the pipefitters and plumbers reported by the Bureau of the Census had completed only four years of high school or less. See Occupational Characteristics, supra, Table 5. Therefore, the realistic labor pool for 91.8% of the membership of Local 638 10 consisted of males in the labor force in the jurisdiction of the Union who, as of 1970, had completed no more than four years of high school.11 The Census data reveals that a smaller percentage of non-white males attended college than white males, leaving a disproportionately large percentage of non-whites in the “high school or less” educational category. See General Characteristics, supra, Tables 126, 129, 130; Occupational Characteristics, supra, Tables 5 & 6. The Court recognizes, on the other hand, that account should be taken of the relatively small proportion of non-whites in the 8.2% group of pipefitters and plumbers who have completed at least one year of post-high school education. Calculations analogous to those performed in the preceding section, but with adjustments for these educational factors, result in a percentage goal of 27.2%.
However, the Court takes notice of the Census data indicating that approximately 75% of the group of pipefitters and plumbers with four years of high school or less (the 91.8% group described above) had completed between one and four years of high school (see Occupational Characteristics, supra, Table 5), but that non-whites comprise a statistically significant larger proportion of the 25% group which have no high school education than they do of the 75% group. See Union’s Exhibit D (chart prepared for the hearing). The parties performed no calculations taking this aspect of the educational data into account, but the Court finds that its consideration is necessary and results in a lowering of the percentage goal.
The Undercount
Another factor in the statistical analysis necessary to establish a percentage goal which is as accurate as possible is the population “undercount”, defined as omission in coverage of the decennial Census. The undercount is computed by the Bureau of the Census and is used by the Bureau to update the decennial Census findings in the intermediate years. See 1970 Census of Population and Housing, Evaluation and Research Pro*988gram, Estimates of Coverage of Population by Sex, Race, and Age: Demographic Analysis, PHC(E)-4 (issued Feb. 1974). The Bureau of Census reports that the national12 undercount for Negro males was 9.9% while that for white males was only 2.4%. Id., Table 3, at 29. Inclusion of this factor in the computations raises the percentage goal to 28.8%.
Conclusion
Taking into account the basic factors noted above, as well as both considering the heavier concentration of non-whites in the group without any high school education and allowing for possible over-inclusiveness of the Spanish Language data as a substitute for statistics for Spanish Sur-named persons, the Court finds that the 1977 percentage goal should be 26%. Cf. Patterson v. Newspaper and Mail Deliverers’ Union, supra.
The Court rejects plaintiffs’ contentions that the percentage goal should reflect the location of work within the Union’s jurisdiction. In the context of the steamfitting industry, the assumption that workers would not travel outside their county of residence seems unwarranted.
The Court also finds that it is inappropriate to inflate the percentage goal because of the alleged high non-white representation in a group known as “discouraged workers”, persons who have ceased looking for work and therefore do not fall within the Census definition of “unemployed”, which thereby prevents their inclusion in the “labor force” category. See General Characteristics, supra, Appendix B, at 15. While this factor may well be a recognized phenomenon, empirical data seems too speculative to permit reliance by the Court, and no reliable statistical material has been submitted.
Accordingly, the Court finds that the percentage goal of 26% non-white membership in Local 638 by July 1, 1977 is reasonable and amply supported by reliable factual data.
Goals for the intermediate years shall be 18% by July 15, 1975 and 23% by July 15, 1976, and the Affirmative Action Plan will be modified accordingly in the Order to be entered herein.
The foregoing shall constitute the Court’s findings of fact and conclusions of law.
Settle Order on Notice.