Robert Valdespino; Brenda Rolon, Plaintiffs-appellants, v. Alamo Heights Independent School District; Ethyl Wayne; Inher Official Capacity As a Member of the Board of Trusteesof the Alamo Heights Independant School District, Sanantonio, Texas, Harry Orem, in His Official Capacity As Amember of the Board of Trustees of the Alamo Heightsindependent School District, San Antonio, Texas; Stephen P.allison; in His Official Capacity As a Member of the Boardof Trustees of the Alamo Heights Independent Schooldistrict, San Antonio, Texas; Anne Ballantyne, in Herofficial Capacity As a Member of the Board of Trustees Ofthe Alamo Heights Independant School District, San Antonio,texas; Thomas A. Kingman, Dr., in His Official Capacity Asa Member of the Board of Trustees of the Alamo Heightsindependent School District, San Antonio, Texas; Terrimusselman, in Her Official Capacity As a Member of the Boardof Trustees of the Alamo Heights Independant Schooldistrict, San Antonio, Texas; Vicki Summers, in Herofficial Capacity As a Member of the Board of Trustees Ofthe Alamo Heights Independant School District, San Antonio,texas, Defendants-appellees, 168 F.3d 848 (5th Cir. 1999)

The panel hereby withdraws its previous opinion and substitutes the following.

Plaintiffs-Appellants claim that the at-large, by-place, majority-vote elections for positions on the AHISD board of trustees dilute their votes as Hispanics in violation of Section 2 of the Voting Rights Act of 1965. See 42 U.S.C. § 1973 (1994) (as amended). The district court found that the Plaintiffs failed to make out a vote dilution claim because they could not prove, under the first Gingles threshold factor, that Hispanics are a "sufficiently large and geographically compact [group] to constitute a majority in a single-member district." Thornburg v. Gingles, 478 U.S. 30, 50, 106 S. Ct. 2752, 2766, 92 L. Ed. 2d 25 (1986). All the issues on appeal involve proof of the first Gingles factor. In particular, we reject the appellants' contention that a "majority" may be less than 50% of the citizen voting-age population. As appellants' other contentions fare no better, the judgment is affirmed.

The School District conceded at trial that the Plaintiffs' demonstration district¹  did comprise a majority of Hispanic voting-age citizens according to 1990 census data. The School District, however, presented evidence that demographic changes between the 1990 census and the 1997 trial had eliminated that majority. AHISD is a small district in which a few strategic land-use changes could and did significantly alter the district's population and neighborhood ethnic mix.

The School District's evidence was presented in expert testimony by Dr. Bill Rives, a demographer. Using the 1990 census data as a baseline, Rives investigated post-1990 changes in the school district's housing stock to determine how the population had changed in the Plaintiffs' demonstration district and in the school district at large. He testified that this methodology is "by far the most popular demographic estimation technique" and is especially appropriate for small areas.

Rives testified that two main trends combined to leave the Plaintiffs' demonstration single-member district "underpopulated" in 1997. Since 1990, the demonstration district had lost population (and the proportion of Hispanics in the demonstration district declined) because a large apartment complex had closed, been renovated, and reopened with a smaller number of residents. Simultaneously, the population of the school district at large had increased because of substantial new residential development in the Lincoln Heights area (formerly a quarry and cement plant), outside the demonstration district. As a result of these changes, the Plaintiffs' demonstration district no longer approached one-seventh of the school district's population, and thus could not be a proper single-member district.

To correct for the underpopulation, Rives added territory to the demonstration district. He added a contiguous area to the north that had been included in some of the appellants' prior proposed demonstration districts. That northern area ran clear to the edge of the school district and had just about the right number of people to make a proper district. Furthermore, if it were not added to the Plaintiffs' district, the northern area would have to be attached to a different district via a mile-long, narrow strip of unpopulated land. After the northern area was added to the demonstration district population, Hispanics made up only 47.9% of the voting-age citizen population of the revised demonstration district. Even if the demonstration district were then partially depopulated (by 8.1% of the ideal population), this number would be 48.3%.

The general thrust of Rives's testimony had been clear for some time before trial. For example, in a November 1996 affidavit attached to the Defendants' Response to Plaintiffs' Motion for Partial Summary Judgment, Rives had explained that no additions to the Plaintiffs' proposed demonstration district could sufficiently increase the Hispanic population because " [t]here are no Hispanic-majority blocks that are adjacent to the proposed district." As the trial grew nearer, Rives updated his analysis, redoing calculations based on more recent data from the apartment complexes about the ethnicity of their residents. Ultimately, the district court relied on Rives's April 1997 Report.

Much of the controversy in the case comes from Rives's August 1997 Report. In July 1997, updated rental rolls became available from The Reserve, the large apartment complex within the Plaintiffs' demonstration district that had been renovated and reopened since the 1990 census. Rives then recalculated the results in his April 1997 report after learning that The Reserve had a slightly higher proportion of Hispanic residents than he had earlier believed. But he also corrected a calculation error in his April Report that had understated population growth in the school district at large. Nevertheless, he concluded that the revised data still did not yield a majority of Hispanics among voting-age citizens within the Plaintiffs' demonstration district. The August Report was given to the Plaintiffs in early August 1997, not long before the original trial setting. On August 22, however, the district court reset the trial for Monday, September 15.

The Plaintiffs' proffered expert witness, George Korbel, claims that he was surprised by the conclusions in Rives's August Report. In response, Korbel scrambled the week before the September trial date to conduct a door-to-door survey of the residents in a small area to the south of the demonstration district. He thought he could find there a high proportion of Hispanic residents that could increase their demonstration district's population without diluting its Hispanic majority. At 4:21 P.M. on Friday, September 12, the Plaintiffs faxed to the School District's counsel a letter disclosing the existence of this new survey. At 4:13 P.M. on Saturday, September 13, the Plaintiffs faxed the data from the survey.

On the Monday morning set for trial, September 15, the School District filed a motion to strike the survey on grounds of unfair surprise. The Plaintiffs' lawyer told the district court that their case in chief would rest entirely on 1990 census data, but that if Rives testified for the School District that more current data changed the Hispanic majority, then the Plaintiffs might use the recent survey as rebuttal testimony. The district court postponed until rebuttal any ruling on the motion to strike and granted a motion in limine to prevent mention of the survey during the case in chief or cross-examination. During the Plaintiffs' rebuttal, the School District renewed its objections to the survey evidence, and the district court granted the motion to strike. The Plaintiffs filed an offer of proof as to what their expert witness would have testified about the survey.

In its findings of fact and conclusions of law, the district court reiterated that Korbel's survey constituted unfair surprise and was excluded under Local Rule CV-16(e). To accommodate the Plaintiffs' objections to the lateness of Rives's August Report, the district court decided to rely solely upon the April Report, which it found to be "thoroughly documented, [with] a high degree of accuracy," and "clear, cogent, and convincing enough to override the presumptive correctness of the prior decennial census." Relying on Rives's report, the district court found that the Plaintiffs had not proved a demonstration district with less than 10% population deviation that included more than 50% Hispanics among its voting-age citizens.

On appeal, the Plaintiffs present three arguments: that they were not required to meet a "bright line" test of 50% Hispanic voting-age citizens in their demonstration district; that the School District's evidence did not adequately overcome the presumed accuracy of the 1990 census data; and that the district court abused its discretion in excluding Korbel's proposed rebuttal testimony about the last-minute, door-to-door survey.

This court reviews district court "findings on the Gingles threshold requirements for clear error." League of United Latin Am. Citizens v. Roscoe Indep. Sch. Dist., 123 F.3d 843, 847 (5th Cir. 1997). See also Gingles, 478 U.S. at 77-79, 106 S. Ct. at 2780-81. A district court's refusal to allow an expert to testify as a rebuttal witness may be overturned only for abuse of discretion. See Tramonte v. Fibreboard Corp., 947 F.2d 762, 764 (5th Cir. 1991); Bradley v. United States, 866 F.2d 120, 124 (5th Cir. 1989).

The Supreme Court has established a three-part threshold inquiry when a racial or ethnic minority group asserts that its distinctive votes have been submerged by the racial majority in a multimember legislative district. The minority group must be able to (1) "demonstrate that it is sufficiently large and geographically compact to constitute a majority in a single-member district," (2) "show that it is politically cohesive," and (3) "demonstrate that the white majority votes sufficiently as a bloc to enable it--in the absence of special circumstances ...--usually to defeat the minority's preferred candidate." Gingles, 478 U.S. at 50-51, 106 S. Ct. at 2766-67.

The Plaintiffs here argue that the district court erred in applying the first Gingles factor as a "bright line" requirement. The Plaintiffs advert to the general purpose of the Gingles factors, which is to provide a framework for showing that there could be "a single-member district in which they could elect candidates of their choice." This is intended to support the proposition that the Plaintiffs need only show generally their electoral potential. The Plaintiffs further argue that the Supreme Court has disavowed "mechanical [ ]" application of the Gingles factors.²  And they complain that the district court did not evaluate evidence of vote dilution under the totality of the circumstances test.

All of these complaints are baseless. In reality, this court has interpreted the Gingles factors as a bright line test. Each factor must be proved before it is necessary to proceed to the totality of the circumstances test. We have repeatedly disposed of vote dilution cases on the principle that " [f]ailure to establish any one of these threshold requirements is fatal." Campos v. City of Houston, 113 F.3d 544, 547 (5th Cir. 1997); accord Rangel v. Morales, 8 F.3d 242, 249 (5th Cir. 1993); Overton v. City of Austin, 871 F.2d 529, 538 (5th Cir. 1989). See also Growe v. Emison, 507 U.S. 25, 40-41, 113 S. Ct. 1075, 1084, 122 L. Ed. 2d 388 (1993) ("Unless [the three Gingles factors] are established, there neither has been a wrong nor can be a remedy.").

Furthermore, contrary to the Plaintiffs' suggestion, this court has required vote dilution claimants to prove that their minority group exceeds 50% of the relevant population in the demonstration district. In Gingles, the Supreme Court required plaintiffs to demonstrate "a majority." 478 U.S. at 50, 106 S. Ct. at 2766. Both of the Fifth Circuit cases cited by the Plaintiffs assumed that 50% was the threshold for "majority" and simply addressed what evidence could be used to prove that the 50% threshold was met. In Brewer v. Ham, the court acknowledged that a super-majority of black residents could be used to prove that blacks constituted a majority of voting-age residents. See 876 F.2d 448, 452 (5th Cir. 1989) (citing cases with raw super-majorities of 65.9%, 71.5%, and higher). In Westwego II, this court repeated Brewer 's holding and expanded on it in a footnote. The footnote, much cited by the Plaintiffs, explained that those plaintiffs "unable to produce hard data" on voting-age population because of the way census data are collected and reported would be able to submit "other probative evidence" to prove voting-age population. See Westwego Citizens for Better Gov't v. City of Westwego, 906 F.2d 1042, 1045 n. 3 (5th Cir. 1990). In context, Westwego II 's statements did not alter what must be proved, only what can be used to prove it. The Plaintiffs still must meet their burden of proving that Hispanics constitute more than 50% of the relevant population in their demonstration district.

Finally, this court has already determined what factors limit the relevant population in the district: voting-age and citizenship. This was made clear in Campos v. City of Houston, 113 F.3d 544, 548 (5th Cir. 1997) (courts "must consider the citizen voting-age population" in evaluating the first Gingles factor). See also Perez v. Pasadena Indep. Sch. Dist., 165 F.3d 368 (5th Cir. 1999). Given that the Supreme Court has avoided the issue of what population to use for the first Gingles factor,³  and that other circuits have used the same approach as Campos,⁴  the district court used the correct legal test here.

In this case, evaluating the district court's application of that test involves two questions: whether the School Board's evidence was adequate to counter the Plaintiffs' census data, and whether the district court abused its discretion in excluding the Plaintiffs' proposed rebuttal evidence.

Except for a cavil, the parties and the district court essentially agree about what standard should be required to overcome census data.⁵  As the district court summarized it:

[C]ensus figures are presumed accurate until proven otherwise. Proof of changed figures must be thoroughly documented, have a high degree of accuracy, and be clear, cogent and convincing to override the presumptive correctness of the prior decennial census.

This standard appears to be an elaboration on one used by the Seventh Circuit. See McNeil v. Springfield Park Dist., 851 F.2d 937, 946 (7th Cir. 1988). Two Fifth Circuit cases are relevant.⁶  The first is Westwego II, which, as mentioned above, opened the door to the use of non-census data when census data are not sufficiently probative of the voting-age proportion of a population. See Westwego, 906 F.2d at 1045 n. 3. The second is Perez, in which this court affirmed a district court's decision that the plaintiffs' population projections were too unreliable to overcome 1990 census data. See Perez, 165 F.3d at 371. Based on Westwego and Perez, the district court properly acknowledged the persuasiveness of census data while admitting evidence that demonstrated its inaccuracy in this case. Because the district court found that the School Board's 1997 population data overcame the 1990 census figures, the question is whether that finding was clearly erroneous.

The Plaintiffs present a laundry list of purported problems concerning the methodology of Rives, the School Board's demographics expert. The School Board's responses as well as Rives's cross-examination at trial suffice to show that the Plaintiffs' challenges are generally misdirected, exaggerations of hypothetical problems, based upon criticisms of assumptions that played no role in Rives's methodology, or based on the analysis in Rives's superseded 1995 reports. The Plaintiffs' most emphatic argument--that Rives himself admitted his April Report was "wrong"--is overstated, because Rives did so only in the context of explaining how the August Report was based on more current data and corrected a calculation error. Rives's admissions did not affect the underlying finding of both the April and August Reports: the Plaintiffs' demonstration district did not contain a majority of Hispanic voting-age citizens.

The general description of Rives's methodology given above reveals that the Gingles I issues in this case do not involve any complicated statistical formulae or tests of significance that might bedazzle or bamboozle an unwary district court. Cf. Overton v. City of Austin, 871 F.2d 529, 544-45 (5th Cir. 1989) (Jones, J., concurring) (discussing some district courts' ill-founded assumptions about the levels at which correlation coefficients become statistically significant). The data here were relatively simple; their manipulation involved only rudimentary arithmetic.

Under these circumstances, the district court did not clearly err in deciding that Rives's report was clear and convincing in its demonstrated of sufficient post-census demographic changes to erode the Hispanic majority in the Plaintiffs' demonstration district. In doing so, we take special note of the School Board's responses to the Plaintiffs' three weightiest methodological criticisms, each of which the district court could have credited without committing clear error: (1) the housing stock methodology can be appropriate for calculating population changes in small areas, (2) it was appropriate to account for some apartment complexes by projecting their imminent populations at the end of ongoing lease-up periods, and (3) despite some lapses, the municipal power company's records of new electrical hook-ups were an accurate gauge of newly developed housing in the entire school district. Further, the School Board's methodology was much more sophisticated than the crude straight-line population projection that was rejected in Perez. See Perez v. Pasadena Indep. Sch. Dist., 958 F. Supp. 1196, 1212-13 (S.D. Tex. 1997), aff'd 165 F.3d 368 (5th Cir. 1999).

Even if the district court properly credited the School Board's post-census demographic evidence, the Plaintiffs argue that it abused its discretion by excluding their proposed rebuttal evidence about post-census populations.

The district court excluded any evidence from Korbel's last-minute survey "because it unfairly surprised the Defendants," citing W.D. Tex. R . CV-16(e), under which the district court may, "upon the showing of good cause," permit a party to supplement the written summary of an expert's proposed testimony.⁷ 

On appeal, the Plaintiffs offer two reasons why their evidence was not an unfair surprise: (1) Korbel had testified in his deposition that Hispanic population was available south of the demonstration district; and (2) the survey was done in response to "new methodologies and numbers" in Rives's August Report and was made available as soon as it was completed.

The Plaintiffs' first reason fails to account for how modern discovery handles expert witnesses. The Local Rule required a "written summary of [Korbel's] proposed testimony." It further required that summary to include "the basis of the opinions which purport to be the testimony of the witness" and "specific references to any exhibits that will be used by the witness in support of any opinions." W.D. Tex. R . CV-16(e) & note. It can scarcely be maintained that Korbel's cursory reference in a deposition could serve as adequate notice of his intent to rely on a door-to-door survey of a specific neighborhood. Nor could that deposition response have provided sufficient information for the School Board to prepare to cross-examine Korbel about the survey. Cf. Sierra Club v. Cedar Point Oil Co., 73 F.3d 546, 571 (5th Cir. 1996) (initial expert disclosures not sufficiently "complete and detailed" to meet discovery order).

The Plaintiffs' second reason takes no account of the fact that the district court relied only upon Rives's April Report, the admissibility of which the Plaintiffs never contested. If the survey was made necessary only by the novelty of the August Report, then apparently it could not have been detrimental to the Plaintiffs to exclude both.

In sum, the court did not abuse its discretion by excluding this evidence for unfair surprise when the proffering party failed to meet its duty to supplement its expert disclosures. See Alldread v. City of Grenada, 988 F.2d 1425, 1436 (5th Cir. 1993) (no error in excluding expert witness' testimony when information crucial to understanding it was not provided until two weeks prior to trial); Mills v. Beech Aircraft Corp., 886 F.2d 758, 764 (5th Cir. 1989) (proper use of discretion to exclude results from tests conducted by plaintiffs' expert the week before trial).⁸ 

Because it was not an abuse of discretion to exclude the survey results on the grounds of unfair surprise, we need not address whether Korbel's survey would have constituted proper rebuttal testimony.V.

For the foregoing reasons, the district court properly placed the burden on the Plaintiffs to prove a majority of Hispanics among voting-age citizens in their demonstration district; the district court did not clearly err in finding the School Board presented sufficient evidence to prove demographic changes since the census; and the district court did not abuse its discretion in excluding the Plaintiffs' proposed rebuttal testimony for unfair surprise.

The judgment of the district court is AFFIRMED.