warrant recognition of subspecies (Fig. 6).

Table 5. Means of "Diagnostic" Characteristics of T. g. gracilis Baird and Girard and T. g. hallowelli Cope.

Fig. 6. Frequency polygons presenting ratios of tail length to total length for males of T. g. gracilis and T. g. hallowelli from Kirn, et al. (1949). The dotted vertical lines represent the means given by Kirn, et al. (1949) for gracilis (left) and hallowelli (right); the solid vertical lines represent the recalculated means for gracilis and hallowelli, respectively, using the data of Kirn, et al. The hatched bars represent the range included by Kirn, et al. (1949) in their diagnoses of gracilis (left) and hallowelli (right).

The data presented by Kirn et al. do not demonstrate intergradation between two populations. Moreover, the diagnostic ranges of the number of subcaudals in males of hallowelli and the ratio of tail length to total length in males of gracilis do not include the recalculated means for those characteristics. Furthermore, the means for the following characteristics are dangerously close to being excluded from their respective "diagnostic" ranges: in hallowelli, the number of ventrals in males; in gracilis, the number of subcaudals in females, and the ratio of tail length to total length in females. It is incongruous that Kirn et al. state (p. 243) that "the coefficient of geographic divergence is .6 per degree latitude" for ventrals, and on the same page they state that "The average number of ventrals gradually increases toward the north in Tantilla gracilis except in southern Oklahoma and central Arkansas (the area of intergradation between T. g. gracilis and T. g. hallowelli) where the change seems to be more abrupt." The data presented in Kirn et al. (1949) do not demonstrate an abrupt change.

The present sample of T. gracilis from Kansas is clearly within the geographic range of T. g. hallowelli as mapped by Kirn et al. (1949:241). How

ever, for the sample from Kansas, the mean number of subcaudals of males is well within the diagnostic range for T. g. gracilis (Table 6), the mean number of subcaudals of females is closer to the mean of gracilis than it is to the mean of hallowelli (Table 5), and the mean of the ratio of tail length to total length of both sexes is within or very close to the diagnostic range for gracilis (Table 6).

Table 6. Comparison of the Means of Some Characteristics of the Sample of Tantilla gracilis from Kansas (Our Data) with Some of the Diagnostic Characteristics (Kirn et al., 1949:240) of Tantilla gracilis gracilis.