88 
William E. Sanderson 
F 
R 
E 
Q 
U 
E 
N 
C 
Y 
10 
8 
6 
4 
2 
p = 0.18 
rn r 
±1 
-1 -2 
-3 -4 -5 
DISCRIMINANT SCORE 
Fig. 3. Histograms of discriminant scores for male (A) and female (B) 
Nerodia cyclopion calibration sample (open bars), N. cyclopion holdout sample 
(shaded bars), N. floridana calibration sample (stippled bars), and N. floridana 
holdout sample (crosshatched bars). 
extremely high level of confidence (P < 0.0001) (Fig. 3). Internasal 
length, prefrontal length, muzzle length, and frontal extension length 
did not contribute significantly to the discrimination in either sex 
and were dropped from the analysis. The remaining variables (loreal 
dorsal length, frontal length, muzzle width, posterior genial length, 
loreal ventral length, anterior genial length, and parietal length) and 
their standardized discriminant function coefficients are given in Table 
2. For both taxa, 100% of the specimens were classified correctly 
by the discriminant analysis, meaning that all specimens, which had 
been previously categorized as either cyclopion or floridana based 
Table 2. Variables incorporated by the stepwise discriminant analysis, listed in 
order of entrance, with standardized discriminant function coefficients. See Figure 
1 for description of variables. 
Males Females 
Variable 
Function 
Variable 
Function 
LD 
0.581 
FL 
2.335 
FL 
-1.561 
MW 
-2.736 
MW 
2.666 
PG 
1.150 
PG 
-0.922 
LD 
-0.587 
LV 
-0.797 
LV 
0.916 
AG 
0.972 
AG 
-0.985 
PL 
-0.843 
