concurring.
I join in the opinion, but set out separately my rationale for the result.
(1) Section 2 claims under the Voting Rights Act must meet the three-part test of Thornburg v. Gingles, 478 U.S. 30, 50-51, 106 S.Ct. 2752, 2766-67, 92 L.Ed.2d 25, 46-47 (1986): (a) a minority population that “is sufficiently large and geographically compact to constitute a majority” in a district; (b) a “politically cohesive” minority group; and (c) a white majority that “votes sufficiently as a bloc to enable it ... usually to defeat the minority’s preferred candidate.”
(2) For Dade County, the second and third parts of the test are established by the overwhelming weight of the evidence in this case, subject to some variations among Hispanics from different counties of origin, and subject to the distinction that the Hispanic minorities are generally diametrically opposed to the voting preferences of the African-American minorities.
(3) Thus, the issue in both the challenge to the Florida Plan for the state Senate districts and for the state House districts in Dade County is a very narrow one: Have the plaintiffs established that the minority populations are sufficiently large and geographically compact to constitute a voting age majority in a district?
(4) Regarding the Senate challenge, the Hispanic plaintiffs established that they could have a fourth Hispanic-majority district that would be geographically compact, but only by either diluting the voting strength of *1583African-American majority districts or by preserving such strength in districts that are not geographically compact. The African-American plaintiffs established that they could create a third African-American majority district only by using a district that was not geographically compact and which would also greatly dilute the voting strength of Hispanics. Moreover, it is impossible to accommodate both four Hispanics and three African-American majority districts in any acceptable manner. Therefore, while the plaintiffs have established a Section 2 violation as to the four Hispanic districts, any remedy implementing the fourth district would violate Section 2 as to African-Americans. Thus, the existing Florida Plan for the Senate districts best accommodates the competing interests of both Hispanics and African-Americans.
(5) Regarding the House challenge, the plaintiffs have established in the De Grandy plan that eleven districts can be established in which Hispanics constitute over 64% of the voting age population and which are geographically compact. Further, these districts can be drawn without any dilution of African-American voting strength. Therefore, plaintiffs have 75 established all requisites of their Section 2 claim as to the Dade County districts for the House of Representatives.
(5)(a) The defendants have attempted to show that it is the citizen Hispanic voting age population that is determinative. That Hispanic citizenship data is not available, however. Despite the lack of available data, the defendants presented estimates of Hispanic citizenship and attempted to apply them to the individual districts. Those estimates are unreliable. For example, William De Grove’s analysis was based on a somewhat unorthodox regression analysis methodology that gave a range of non-citizenship rates of 9.5% for precincts with no Hispanics to 55% for precincts with 100% Hispanic voters. (Acknowledging, of course, that there is no such precinct.) He recognized that the error in that analysis was greatest at the extremes, i.e. at the 9.5% and 55% intercepts. Nevertheless, he used the 55% as the basis of all his citizenship calculations. Plainly, his estimated Hispanic citizenship ratios for the eleven districts must be viewed with low confidence and with a large range of error.
(5)(b) A better gauge of eligible voters within a district is the analysis of past election results. Those statistics established that Hispanics would be able to elect Hispanic candidates of choice in all eleven districts proposed by plaintiffs, and those statistics automatically account for voter citizenship, registration, and turn out. Thus, in the absence of any better data, the plaintiffs’ evidence is sufficient. Further, Dr, Allan Lichtman’s calculations for the plaintiffs, based in part on the defendants’ citizenship data estimates, reflected an Hispanic voting age citizenship of 50% or more in all of the questioned eleven districts. This is a more relevant calculation, for the supermajority percentages are merely designed to account for, inter alia, citizenship.
(5)(c) Equally important, it is unrefuted that Hispanics have been able to elect the Hispanic candidate of choice in every district in which Hispanics constitute 59% or more of the voting age population (without regard to citizenship). It is not necessary to target anything greater than that. Nor is it necessary, although it may be desirable, to create a supermajority of 65% Hispanics (or 55% African Americans) to accomplish Section 2’s purposes. Additionally, the growth trends in Dade County’s Hispanic population indicates that the percentage of Hispanics will continue to increase, but the plaintiffs’ Section 2 claims have been evaluated only on the basis of the 1990 census data. That data clearly establishes that the eleven districts meet the Gingles test.
(6) The remedial aspect of this court’s adoption of the modified De Grandy plan presented a major challenge. In recognition of the respect due to the state’s own policy and plan for redistricting, this court considered the defendants’ two submitted plans, neither of which was even close to meeting Section 2’s requirements and incorporating eleven Hispanic districts in Dade County in accordance with this court’s findings on liability. At plaintiffs’ suggestion, this court then requested that the defendants draw a *1584plan that incorporated the De Grandy plan for Dade County into the rest of the state. The defendants refused, however, and the court recessed to consider all of the alternatives available and make a decision. Because of the severe time constraints and the upcoming July 4th holiday weekend, it' was critical that a plan be adopted without further delay. It was also apparent that the defendants were intent on delaying the adoption of any plan that implemented a true eleven-Hispanic districts House plan for Dade County. Accordingly, this court adopted the modified De Grandy plan. It does, of course, have a “ripple effect,” but that effect is relatively minor and simply cannot be avoided. The defendants’ belated change of position and announcement that they would attempt to draw a plan incorporating the eleven Hispanic districts from the De Grandy plan into the rest of the state came too late.