mental is in contact with only the left anterior chin-shield in 15 specimens, only the right anterior chin-shield in 2

specimens, and the mental does not contact either anterior chin-shield in 39 specimens. Separation of the mental and an anterior chin-shield always results from the posterior elongation of a first infralabial.

Table 3. Variation in Number of Supralabials Entering the Eye in 246 Specimens of Tantilla gracilis.

Dorsal scale rows

Of the 245 specimens examined for this characteristic, 228 (92.7%) have scale counts of 15-15-15-15-15. The number of scale rows around the neck and one head length behind the head are 15 except in a single specimen that has 14 scale rows at these three places. The number of scale rows around midbody is 15 in 244 of 246 specimens (99.2%); two specimens have 14 scale rows around midbody. The number of preanal scale rows is more variable; counts of 14 through 17 were recorded. Fourteen specimens have 14 preanal scale rows (5.7%), 229 have 15 preanal rows (93.1%), two have 16 preanal rows (0.8%), and one has 17 preanal rows (0.4%). Known causes for atypical numbers of preanal scale rows are listed in Table 4. A preanal count at one head length anterior to the anus would have omitted most of the above variation in numbers of preanal scale rows.

Table 4. Atypical Numbers of Preanal Scale Rows in Specimens of Tantilla gracilis.

The complete dorsal scale formula for KU 83620, a male, is as follows:

[Textual representation of dorsal scale formula.]

Specimen 83620 is the only one for which we have determined the complete dorsal scale formula. Obviously the formula is complex, and possibly the specimen is atypical in this regard. Nevertheless, addition and reduction of scale rows is apparent on other specimens and can be easily observed by examining specimens under a dissection microscope.

Anal plate

Of the 246 specimens examined, 232 (94.3%) have the anal plate divided. Of the 14 specimens having undivided anal plates, four have an incomplete groove anteriorly in the plate.

Ventrals

Females have from 122 to 137 (mean, 130.8 ± 0.26; N = 107) ventrals; males have from 117 to 131 (mean, 122.3 ± 0.21; N = 139; Fig. 2). Of the 246 specimens examined, 29 (11.8%; 11 females and 18 males) have an added

half-ventral immediately anterior to the anal plate. Of these specimens, 27 have the added half-ventral on the left side and two have it on the right. Five specimens have the last ventral divided similar to the normal division of the anal plate. One specimen has the last ventral incompletely divided transversely. Excluding the last ventral, seven specimens have divided ventrals (half-ventrals). Four of these specimens have added half-ventrals on the right side (two specimens with one half-ventral each, one specimen with two half-ventrals, and one specimen with three half-ventrals), and four have added half-ventrals on the left side (one half-ventral on each specimen).

Fig. 2. Number of ventrals in females and males of T. gracilis.

Subcaudals

Females have from 37 to 47 (mean, 42.0 ± 0.20; N = 79) subcaudals; males have from 43 to 55 (mean, 48.4 ± 0.23; N = 109; Fig. 3). Of the 187 specimens examined, 13 (7.0%) have the last pair of subcaudals fused. One specimen has left subcaudal 28 fused with both adjacent subcaudals (also fused) on the right. While reading proof we found two omissions in Fig. 3; there should be illustrated two males with 43 subcaudals and three males with 44 subcaudals.