tected values. By implication, one must realize that most doubts and most criticisms of the proposed standard will be, in effect, criticisms of a districted electoral system itself.
The obvious limitation is that not all residential patterns can be expected to lead to a normal distribution. Nevertheless, for the purpose of detailing the logic of the proposed standard, unencumbered by the details of how the normality assumption can be augmented, we assume the conditions leading to normality are approximated. Until the augmentation is detailed, one can keep in mind that, as a general guideline, the more the relevant groups are segregated beyond the "norm" (see, e.g., Gudgin and Taylor, 1979, pp. 40-47, esp. Tables 3.3 and 3.4), the more the power function equation will overstate the majority group's seat proportion.